UAV–Ground Vehicle Collaborative Delivery in Emergency Response: A Review of Key Technologies and Future Trends

Abstract

UAV delivery and ground transfer scheduling in emergency scenarios represent critical technological systems for enhancing disaster response capabilities and safeguarding lives and property. This study systematically reviews recent advances across eight core research domains: UAV emergency delivery systems, ground–air integrated transportation coordination, emergency logistics optimization, UAV path planning and scheduling algorithms, collaborative optimization between ground vehicles and UAVs, emergency response decision support systems, low-altitude economy and urban air traffic management, and intelligent transportation system integration. Research findings indicate that UAV delivery technologies in emergency contexts have evolved from single-aircraft applications to intelligent multi-modal collaborative systems, demonstrating significant advantages in medical supply distribution, disaster relief, and search-and-rescue operations. Current technological development exhibits four major trends: hybrid optimization algorithms, multi-UAV cooperation, artificial intelligence enhancement, and real-time adaptation capabilities. However, critical challenges persist, including regulatory framework integration, adverse weather adaptability, cybersecurity protection, human–machine interface design, cost–benefit assessment, and standardization deficiencies. Future research should prioritize distributed decision architectures, robustness optimization, cross-domain collaboration mechanisms, emerging technology integration, and practical application validation. This comprehensive review provides systematic theoretical foundations and practical guidance for emergency management agencies in formulating technology development strategies, enterprises in investment planning, and research institutions in determining research priorities.

1. Introduction

1.1. Research Background and Significance

The global community currently faces unprecedented emergency management challenges. Since 2020, the frequent occurrence of compound disasters—including earthquakes, the COVID-19 pandemic, extreme climate events, and geopolitical conflicts—has increasingly highlighted the limitations of traditional emergency response systems [1,2]. As explicitly noted by Yucesoy et al. [3] in their comprehensive review of Unmanned Aerial Vehicle (UAV) roles in disaster response, conventional ground-based rescue approaches become virtually ineffective when confronted with extreme conditions such as damaged roads, collapsed bridges, or flood isolation, necessitating revolutionary technological breakthroughs to address this critical capability gap [4]. Traditional emergency systems exhibit four fundamental deficiencies: first, slow response times—ground transportation constraints limit access to disaster core areas within critical golden rescue time windows; second, limited coverage—complex terrain and adverse environments render numerous affected areas rescue “blind spots”; third, inefficient resource allocation—absence of real-time information support results in severe supply–demand mismatches in relief distribution; fourth, outdated coordination mechanisms—poor information flow among multiple agencies and administrative levels leads to substantial resource waste. The devastating loss of life and property caused by these issues during major disasters is deeply concerning, urgently demanding innovative technological solutions to achieve transformative improvements in emergency response capabilities.

The rapid advancement of UAV technology presents unprecedented opportunities to address these challenges. The literature review by Macrina et al. [5] systematically analyzed recent progress in UAV-assisted routing, demonstrating that UAV technology has achieved remarkable advancement in recent years. In particular, the emergence of high-speed microprocessors has enabled intelligent autonomous control of multiple systems, making UAV applications increasingly prevalent in photography, construction, surveillance, and monitoring, with last-mile delivery and emergency response among the most promising applications. UAVs demonstrate five distinct advantages in emergency scenarios: first, rapid response capability—the ability to take off within minutes of receiving commands, unaffected by ground traffic conditions; second, three-dimensional maneuverability—the capability to traverse obstacles and reach areas inaccessible by conventional transportation means; third, cost-effective efficiency—significant operational cost reduction compared to traditional aircraft such as helicopters; fourth, flexible deployment—dynamic mission and route adjustment capabilities based on disaster conditions; fifth, enhanced safety—elimination of risks associated with deploying rescue personnel in hazardous areas. These advantages position UAVs as core technological enablers for the modernization transformation of emergency response systems.

However, relying solely on UAV technology cannot comprehensively address all emergency rescue challenges. Through text mining analysis of 5364 papers (1978–2023), Jahani et al. [6] discovered that UAV applications in logistics and supply chain management have evolved from single-tool implementations to complex multi-modal collaborative systems. Their research employed a Latent Dirichlet Allocation (LDA) model to identify ten distinct research topics, revealing four promising future research directions, with the most significant being the integration of UAV-based logistics with multi-modal transportation to achieve efficient traffic management. Ground–air integrated collaboration has become an inevitable trend due to several factors: strong complementarity—ground vehicles offer high payload capacity and extended endurance, while UAVs provide high speed and exceptional flexibility, creating synergistic advantages; optimal system efficiency—collaborative scheduling achieves synergistic effects; enhanced reliability—multi-modal transportation provides redundant backup capabilities, improving system robustness. This integration represents not only an inevitable outcome of technological development but also an urgent requirement of emergency management practice.

Recent years have witnessed breakthrough advances in key technologies for UAV delivery in emergency scenarios. The innovative research by Pan et al. [7] demonstrated how system innovation can achieve substantial performance improvements. Their proposed ULRB framework (UAV Landing and Recharging on Buses) addresses two major industry challenges—insufficient delivery range and battery degradation—by integrating logistics UAVs with public transportation systems. Using extensive evaluations with large-scale real-world bus trajectory datasets, the results were remarkable: UAV delivery range extended by an average of 5.54 times (reaching 7.73 times in extreme cases), and battery aging decelerated by an average of 3.26 times (reaching 4.16 times in extreme cases). This research not only demonstrates the tremendous potential of technological innovation but also, more importantly, provides a novel system design paradigm—achieving capability enhancement through cross-domain integration.

UAV delivery in emergency scenarios represents not merely an aircraft technology challenge but a complex system optimization problem. The joint optimization framework proposed by Sajid et al. [8] fully embodies this concept. They established a mixed-integer linear programming (MILP) model that considers the effects of ancillary processes and varying payload on flight time for UAV-assisted delivery system routing and scheduling problems. Through developing a hybrid genetic and simulated annealing (HGSA) algorithm combined with a UAV-oriented MinMin (UO-MinMin) scheduling algorithm, they achieved dual objectives of travel time minimization and task completion time optimization.

Table 1. Summary of key reference contributions in the Introduction section.

Research Focus	Main Contributions	Technical Methods	Application Scenarios	Representative Studies
COVID-19 and climate disaster compound risks	First systematic analysis of compound climate risks during pandemic, revealing limitations of traditional emergency systems	Interdisciplinary risk assessment framework	Global pandemic and extreme weather event interactions	[1]
Compound event management in COVID-19 era	Identified key deficiencies and challenges in emergency management, proposed improvement recommendations	Critical analysis methods, case studies	Sri Lanka monsoon and pandemic compound management	[2]
Operations research applications of UAV disaster response	Established theoretical framework for UAV disaster response operations research, defined three operational types	Systematic literature review, classification analysis	Information collection, material delivery, communication support	[3]
Emergency responder capability assessment	Revealed fundamental deficiencies, including insufficient interoperability of traditional emergency communication systems	Network-Centric Operations (NCO) concept	Homeland security, multi-agency coordination	[4]
UAV-assisted path planning	Systematically reviewed UAV path planning technology progress, established complete technical classification system	Bibliometric analysis, technical classification	Last-mile delivery, collaborative path planning	[5]
UAV logistics supply chain text mining	Revealed UAV application evolution from single-tool to multi-modal collaborative systems through large-scale text mining	LDA topic modeling, text mining (5364 papers)	Logistics supply chain, multi-modal transportation integration	[6]
UAV and public transport system integration innovation	Proposed ULRB framework, achieving 5.54× UAV delivery range extension and 3.26× battery aging deceleration	System integration design, large-scale real-world data validation	Urban logistics, cross-domain integration	[7]
UAV-assisted delivery system optimization	Established MILP model considering ancillary processes and payload variations, developed HGSA-UO-MinMin algorithm	Mixed-integer linear programming, hybrid genetic simulated annealing algorithm	Delivery route scheduling, task optimization	[8]

summarizes the main contributions of the eight core references cited in Section 1.1 of the Introduction. Based on the aforementioned problem-oriented and demand-driven approach, this paper aims to (1) comprehensively review the technological development trajectory of UAV delivery and ground transfer scheduling in emergency scenarios, constructing a complete knowledge system framework; (2) conduct in-depth analysis of current technological development bottlenecks, identifying key factors constraining industrial application; (3) proactively propose future technological development pathways, providing scientific foundations for relevant decision-making; and (4) provide systematic reference guidelines for government agencies, research institutions, and industry stakeholders.

1.2. Review Methodology

1.2.1. Literature Search Strategy

This review employs a multi-database systematic search methodology to ensure comprehensive acquisition of high-quality research literature related to UAV delivery and ground transfer scheduling in emergency scenarios. The primary databases searched included the Web of Science Core Collection, the IEEE Xplore Digital Library, ScienceDirect, the Engineering Index (EI Compendex), China National Knowledge Infrastructure (CNKI), and Google Scholar, ensuring comprehensive coverage of both international and domestic literature. Additionally, major academic conference proceedings were searched, including papers from authoritative conferences such as the IEEE Intelligent Transportation Systems Conference (ITSC), the International Institute for Operations Research and the Management Sciences Annual Meeting (INFORMS), and the Transportation Research Board Annual Meeting (TRB), as well as technical reports from research institutions and government agencies.

To ensure systematic and comprehensive search coverage, this research employed a combined approach of forward and backward citation tracking, identifying and acquiring additional relevant studies through analysis of key citation relationships in the literature. Furthermore, attention was paid to the latest publication dynamics of major research teams and journals in this field, ensuring the inclusion of cutting-edge research findings.

1.2.2. Search Terms and Scope Definition

The search strategy employed a combination of controlled vocabulary and free-text terms, constructing multi-level search expressions. Core search terms included “UAV/Unmanned Aerial Vehicle”, “Emergency Response”, “Disaster Management”, “Logistics Delivery”, “Path Planning”, “Vehicle Routing”, “Collaborative Optimization”, “Ground-Air Integration”, and “Intelligent Scheduling”. The temporal scope primarily focused on research literature published from 2010 to 2025, reflecting theoretical advances and practical applications during the rapid development period of UAV technology. Geographic scope was unrestricted, but special attention was paid to technological application cases in complex emergency environments and specific requirements of China’s emergency management system.

1.2.3. Literature Inclusion and Exclusion Criteria

Inclusion criteria: (1) peer-reviewed journal articles and conference papers, as well as technical reports from authoritative institutions; (2) research topics focused on core technological domains such as UAV emergency delivery, ground–air collaborative transportation, and emergency logistics optimization; (3) studies with clear technological innovation points or practical application value; (4) research providing quantitative analysis results or empirically validated theoretical models; (5) publicly published literature in English or Chinese.

Exclusion criteria: (1) studies involving purely theoretical discussions lacking technological implementation pathways; (2) literature focusing solely on single technological components without addressing system integration; (3) duplicate publications or journal-extended versions of conference papers (retaining higher-impact factor versions); (4) conceptual discussions lacking sufficient technical detail; (5) research primarily targeting military applications with limited applicability to civilian emergency scenarios; (6) literature published too early, with technological approaches proven infeasible.

1.2.4. Literature Screening and Comprehensive Analysis Process

The literature screening employed a two-stage assessment approach: The initial screening stage involved relevance assessment based on titles and abstracts, independently conducted by two researchers, with disagreements resolved through discussion. The refined screening stage involved full-text reading and quality assessment, focusing on evaluating research scientific rigor, innovation, and practicality, ultimately including 78 high-quality publications.

Given the narrative review nature of this study, the comprehensive analysis process emphasized systematic organization of technological development trajectories and thematic classification rather than quantitative meta-analysis. Based on literature content characteristics, an analytical framework was constructed that encompassed eight core technological dimensions: UAV emergency delivery system architecture and optimization, ground–air integrated transportation coordination mechanisms, emergency logistics and rescue operation optimization strategies, UAV path planning and scheduling algorithm innovation, ground vehicle routing and UAV collaborative optimization, emergency response decision support system construction, low-altitude economy and urban air traffic management systems, and intelligent transportation system integration and coordination. To ensure analytical objectivity and comprehensiveness, this study particularly focused on comparative analysis among different technological approaches, identifying current research hotspots and gaps. Through systematic bibliometric analysis, important information regarding temporal evolution trends, major research institution distribution, and characteristics of highly cited publications in this field was identified, providing readers with a panoramic understanding of this research domain.

1.3. Paper Structure

The paper structure is organized as follows: Section 2 reviews the relevant literature across eight core technological dimensions, presenting a comprehensive overview of the technological development; Section 3 conducts in-depth analysis of current challenges, revealing underlying mechanisms of constraining factors; Section 4 proactively proposes future development directions, depicting technological evolution roadmaps; Section 5 summarizes the entire paper, distilling core viewpoints and policy recommendations.

2. Analysis of Key Technology Systems and Application Scenarios

UAV delivery and ground transfer scheduling in emergency scenarios involve deep integration and system-level coordination across multiple technological layers. This section systematically analyzes the technological architecture, methodological innovations, and practical applications in this field along eight core technological dimensions, constructing a comprehensive technology system framework.

2.1. UAV Emergency Delivery System Architecture and Optimization

UAV emergency delivery systems serve as the foundational support for the entire technology framework, with their core focus on constructing flight platforms and delivery mechanisms capable of adapting to complex emergency environments. The research by Ghelichi et al. [9] represents the latest advances in this field, proposing stochastic optimization-based solutions to address the core challenge of demand uncertainty in disaster-affected areas.

The innovation of this research manifests across three levels: First, in theoretical modeling, the authors developed a chance-constrained programming (CCP) formulation that selects platform locations whose utility distribution produces minimum α-percentiles, ensuring system performance at specified probability levels. Second, in algorithmic design, they employed finite-horizon stochastic dynamic programming solutions, decomposing complex problems into three manageable stages—identifying optimal platform combinations, developing greedy-based approximation algorithms, and utilizing sample average approximation (SAA) methods for optimal platform set selection. Third, in a practical validation, through a case study in Central Florida, they demonstrated that this approach better handles demand fluctuations compared to traditional deterministic methods, maintaining stable service levels under uncertain conditions.

Energy efficiency optimization represents another critical challenge in UAV emergency delivery operations. The research by Dukkanci et al. [10] provides breakthrough solutions from an energy management perspective. They introduced the Energy-Minimizing and Range-Constrained Drone Delivery Problem (ERDDP), marking the first study to explicitly calculate energy consumption as a function of UAV speed. Core contributions of this research include establishing a nonlinear model that simultaneously considers launch point selection, customer assignment, and flight speed optimization; reformulating through second-order cone programming to enable previously intractable nonlinear problems to be handled using commercial optimization software (IBM ILOG CPLEX Optimization Studio version 12.7.1); and introducing perspective cutting techniques that significantly enhance model solution efficiency. Computational results demonstrate that speed optimization achieves energy conservation while satisfying service time constraints, which holds significant importance for extending UAV operational duration in emergency scenarios.

Low-altitude airspace congestion management represents a systemic challenge that must be addressed for large-scale UAV deployment. The research by She and Ouyang [11] fills theoretical gaps in this domain. They pioneered the introduction of traffic flow theory into three-dimensional low-altitude airspace, establishing an equilibrium model for self-organized UAV traffic flows. Research findings indicate the following: in high-density UAV operational environments, congestion effects lead to dramatic system efficiency degradation; through rational airspace design and flow control, congestion-induced delays can be reduced; ground-based distribution centers and airborne fulfillment centers each demonstrate advantages under different density conditions. This research provides theoretical foundations for large-scale UAV scheduling in emergency scenarios.

Multi-UAV collaborative systems represent the cutting-edge direction of technological development. The hybrid genetic algorithm proposed by Peng et al. [12] achieves genuine multi-aircraft coordination. Their innovations include developing a scheduling framework supporting simultaneous operations of multiple heterogeneous UAVs; designing a complete algorithmic workflow encompassing population management, heuristic initialization, and population education; and implementing parallel operational modes, where multiple UAVs simultaneously take off from different locations and execute diverse tasks. Performance evaluation results demonstrate that multi-UAV collaboration significantly enhances delivery efficiency compared to single-UAV systems, holding crucial importance for large-scale material distribution in emergency scenarios.

The special requirements of humanitarian relief have driven the development of specialized systems. Lu et al. [13] applied multi-objective optimization theory to humanitarian logistics, proposing innovative models that simultaneously consider efficiency and equity. Their multi-objective humanitarian pickup and delivery vehicle routing problem with drones (m-HPDVRPD) incorporates two critical objectives: minimizing maximum collaborative routing time (ensuring overall efficiency) and maximizing minimum demand node satisfaction rate (ensuring equity). Through developing a hybrid multi-objective evolutionary algorithm with specialized local search operators (HMOEAS), they demonstrated in actual COVID-19 pandemic cases the advantages of truck–UAV collaborative modes over single-mode approaches in epidemic material distribution.

Figure 1 illustrates the four-layer architecture of the UAV emergency delivery system. The emergency response application layer provides specialized services, including medical emergency delivery, disaster relief operations, humanitarian logistics, and urban emergency response, directly addressing the needs of different types of emergency scenarios. The intelligent coordination and decision layer achieves ground–air integration coordination, AI-enhanced decision support, and real-time adaptive scheduling, providing intelligent management capabilities for the system. The technical operation layer encompasses seven core technical modules: path planning, route optimization, energy management, load balancing, traffic management, quality control, and security assurance, forming the technical foundation for system operations. The physical resource and sensing layer consists of four major physical components: UAV fleets, ground vehicle networks, communication infrastructure, and emergency infrastructure, providing the hardware foundation for system operation. The four layers are organically integrated through standardized interfaces, supporting multi-modal collaborative delivery missions in complex emergency environments and ensuring that the system can provide efficient, reliable, and scalable emergency supply delivery services when facing various emergency situations.

2.2. Ground–Air Integrated Transportation Coordination Mechanisms

Ground–air integrated transportation coordination represents a critical technological breakthrough for achieving efficient emergency response, with its core focus on breaking traditional transportation mode boundaries and achieving multi-dimensional resource optimization.

The specificity of medical emergency scenarios has driven the development of specialized coordination systems. Using Istanbul as a case study, Özkan [14] investigated the transportation challenges of blood products—materials extremely sensitive to both time and temperature. The research established a multi-objective integer programming (MOIP) model that simultaneously optimizes the number of UAVs used and total flight distance, while considering UAV range, payload weight and volume constraints, and satisfying hospital blood requirements and distribution center supply capabilities. Through analysis of real data from the Turkish Red Crescent in 2019, the research found that vertical take-off and landing UAVs can bypass ground traffic congestion in urban environments, significantly reducing blood product delivery times, and that through multi-objective meta-heuristic algorithm applications, Pareto-optimal solution sets can be identified within acceptable computational times, providing decision-makers with multiple trade-off options.

Hub selection for rapid medical delivery networks is crucial for system design. Escribano Macias et al. [15] proposed an innovative bi-stage operational planning approach. The first stage employs trajectory optimization algorithms considering multiple flight phases, including takeoff, cruising, hovering, and landing, for precise energy consumption calculation; the second stage combines battery management heuristic algorithms for hub selection and routing planning. In a hypothetical response mission for the 1999 Taiwan Chi-Chi earthquake, research demonstrated that through optimizing hub locations, UAV fleets can provide emergency relief to 20,000 people within 24 h; compared to conservative energy estimation methods, this approach reduces battery inventory requirements; and multi-objective optimization considering mission duration and allocation fairness achieves an excellent balance between efficiency and equity.

Comprehensive design of rescue systems requires consideration of multiple factors. Zhao et al. [16] designed a complete UAV rescue system from a systems engineering perspective. Through analyzing problem characteristics and establishing reasonable assumptions, they employed transportation planning theory to establish optimization models with the objective function of minimizing idle-space buffer material volume while considering UAV payload constraints. Research results indicate that each ISO cargo container configured with five Type B UAVs and one Type C, one Type F, and one Type H UAV can provide 188 days of rescue requirements, achieving 71.4% container space utilization. This standardized configuration scheme provides practical guidance for rapid emergency material deployment.

Multi-stage scheduling system design reflects temporal dimension optimization. The priority-based multi-stage emergency material scheduling system proposed by Gong et al. [17] holds significant innovative importance. They constructed a three-tier “storage warehouses–transit centers–disaster areas” network, integrating the large-scale transportation capabilities of trains with the flexible delivery advantages of UAVs to achieve multi-modal coordination. The research developed a Tabu Genetic Algorithm combined with the Branch and Bound (TGA-BB) method, integrating the global search capability of genetic algorithms, precise solution mechanisms of branch and bound, and local search avoidance characteristics of Tabu search. Across eight test instances at different scales, TGA-BB outperformed other algorithms in both average response time and average runtime, reducing the average response time by approximately 52.37% compared to TGA-PSO in Instance 7 and shortening the average runtime by about 97.95% compared to TGA-SA in Instance 2.

The special requirements of pandemic control have catalyzed innovative applications. The UAV-based network system proposed by Kumar et al. [18] during the COVID-19 pandemic demonstrated the tremendous potential of technology in public health crises. The system integrates multiple critical functions: thermal screening for temperature monitoring; large-area disinfection using spraying devices; public announcements through speakers; transportation of medical supplies and samples. The push–pull data acquisition mechanism in the research is particularly suitable for remote and highly congested epidemic areas, playing important roles in regions where wireless or internet connectivity issues exist or COVID-19 transmission risks are high. Actual deployment results show that UAV systems can cover 2 km areas within 10 min for disinfection, thermal imaging collection, and patient identification, while avoiding direct human contact risks.

The establishment of theoretical frameworks lays foundations for technological development. The Flying Sidekick Traveling Salesman Problem (FSTSP) proposed by Murray and Chu [19] holds milestone significance. They first formalized the collaborative operations between UAVs and traditional delivery trucks as mathematical models, demonstrating that this collaborative mode can achieve efficiency improvements compared to independent operations. The research provided two mixed-integer linear programming formulations and two simple yet effective heuristic solution methods, with numerical analysis revealing trade-offs between using UAVs with faster flight speeds versus those with longer endurance. This pioneering work provided theoretical foundations and methodological guidance for subsequent research.

Scientific site selection for takeoff and landing points represents a fundamental component of ground–air integrated coordination. Wu’s research [20] established terrain and meteorological constraint models for helicopter takeoff, landing, and operations, as well as radar detection blind zone models, proposing optimization methods for helicopter landing site selection. The research defined air–ground attributes of facility points, proposed four rescue modes based on connection methods, established a two-stage site selection model under aircraft performance constraints, and designed A* algorithms for solution. The results demonstrate that ground–air collaborative emergency scheduling models can significantly reduce rescue costs and time while improving rescue efficiency. This research provides an important reference for ground infrastructure layout of emergency UAV systems: through establishing multi-constraint landing site optimization models that integrate terrain, meteorological, and performance factors, optimal allocation of rescue resources and maximization of scheduling efficiency can be achieved.

2.3. Emergency Logistics and Rescue Operation Optimization Strategies

Emergency logistics optimization represents a core component of enhancing disaster response efficiency, requiring optimal decision-making under extreme uncertainty and time pressure.

The expansion of multi-UAV systems brings new optimization opportunities. Murray and Raj [21] extended FSTSP to multi-UAV scenarios, with their proposed Multiple Flying Sidekicks Traveling Salesman Problem representing a significant advancement in this field. Their research revealed the complexity of multi-UAV coordination: combinatorial explosion issues in task allocation; difficulty in handling synchronization constraints; scheduling complexity for charging and maintenance. Through developing heuristic methods containing three sub-problem solutions, the research achieved effective solutions for practical-scale problems. Extensive numerical testing indicates that potential time savings from using multiple UAVs exhibit diminishing marginal returns. This finding provides important guidance for emergency resource allocation.

The development of adaptive algorithms enhances system flexibility. The Adaptive Large Neighborhood Search (ALNS) metaheuristic algorithm developed by Sacramento et al. [22] represents the frontier of algorithmic innovation. Core innovations of this algorithm include destroy–repair operators specifically designed for UAV constraints, flexible launch and recovery mechanisms, and precise fuel consumption calculations included in cost functions. In extended studies of capacitated multi-truck scenarios, the algorithm successfully solved instances with 100 customers. Sensitivity analysis revealed the impact of key parameters: every 10% increase in UAV speed improves system efficiency by 6–8%; payload capacity is a more critical constraint factor than endurance; rational UAV quantity configuration can significantly enhance system performance.

Consideration of time dependency makes models more realistic. Wang et al. [23] first introduced time-dependent road travel times in truck–UAV coordination, representing an important step toward practical applications. Their continuous-time model based on fluid queuing theory can accurately characterize time-varying characteristics of traffic congestion, dynamically adjust routing strategies to avoid congested periods, and achieve dynamic coordination between trucks and UAVs. In a validation with actual traffic data from Chongqing, China, the model demonstrated excellent performance with significantly enhanced system adaptability to terrain constraints. The research also conducted a sensitivity analysis for four road congestion states, providing guidance for scheduling strategies under different traffic conditions.

Enhanced prediction capabilities provide data support for optimization. The research by Wang et al. [24] compared the performance of deep learning versus traditional methods in urban road travel time prediction. Through designing 10 prediction scenarios with different sliding time window lengths and prediction scales, the research found that LSTM-DNN models achieved the best Mean Absolute Percentage Error (MAPE) values across all 30 min sliding time window prediction scenarios and that deep learning models generally outperformed shallow learning models in MAPE metrics; however, in Root Mean Square Error (RMSE) metrics, shallow learning models such as random forest performed better in most scenarios. This multi-dimensional performance evaluation provides a scientific basis for prediction model selection in emergency scenarios.

2.4. UAV Path Planning and Scheduling Algorithm Innovation

Path planning and scheduling optimization serve as the “brain” of UAV systems, directly determining task execution efficiency and success rates.

The integration of inventory routing pioneered new optimization dimensions. Najy et al. [25] first combined inventory routing problems with UAV delivery, representing significant innovation in operations research. Their proposed IRP-D (Inventory Routing Problem with Drones) model simultaneously considers dynamic changes in inventory levels, multiple UAV dispatches, and collaborative operations between trucks and UAVs. Through precise branch-and-cut algorithms and heuristic methods based on basic IRP solutions, the research demonstrated significant cost savings—compared to traditional inventory routing methods, this integrated optimization approach is particularly suitable for emergency material prepositioning and distribution.

Energy management has become a core consideration in algorithm design. Ramasamy et al. [26] proposed innovative solutions for coordinated path planning problems of multiple fuel-constrained UAVs. Their three-tier heuristic approach includes using K-means clustering to determine UGV waypoints; optimizing UGV routes through Traveling Salesman Problem (TSP) formulations; and employing Vehicle Routing Problem (VRP) formulations with capacity constraints, time windows, and allowed missed visits to solve UAV routes. Research found that constraint programming solvers are 7–30 times faster than mixed-integer solvers but 4–15% sub-optimal. Through Monte Carlo methods evaluating the impact of mission distribution, cluster numbers, and UAV quantities, quantitative guidance for system configuration was provided.

Artificial intelligence methods have brought revolutionary breakthroughs. The reinforcement learning-based truck–UAV coordinated delivery system proposed by Wu et al. [27] demonstrates the tremendous potential of AI technology. Their innovations include decomposing delivery problems into customer clustering and routing sub-problems, employing encoder–decoder frameworks combined with reinforcement learning without requiring manually designed heuristic rules, and designing different problem contexts specific to trucks and UAVs. Experimental results demonstrate that this method has good generalization capability and can be efficiently applied to problems at different scales. This end-to-end learning approach provides new insights for handling complex emergency scenarios.

The evolution of operations research reveals domain development trends. The review study by Galindo and Batta [28] systematically examined OR/MS research progress in disaster operations management. Their analysis shows that the field is evolving from purely engineering approaches toward multidisciplinary methods encompassing earth sciences and sustainability sciences; that integrated optimization of pre-positioning, real-time response, and recovery phases has become a new trend; and that uncertainty handling and robust optimization methods are increasingly being emphasized. Research gaps identified by the study point to future development directions.

Systematic research on prepositioning strategies provides strategic guidance. Sabbaghtorkan et al. [29] conducted a statistical analysis of the relevant literature from 2000 to 2018, identifying eight risk mitigation methods. Using fuzzy TOPSIS for strategy evaluation, they found that the combination of strategic reserves and flexible transportation is most effective; collaborative frameworks can significantly improve resource utilization efficiency; models considering demand-side costs are more realistic; and uncertainty in funding, budget, and infrastructure represents major challenges. The research particularly noted the potential role of social media in disaster preparedness, providing new insights for emergency response informatization.

Trajectory planning in complex environments represents a core technology for emergency rescue flights. Hang [30] conducted systematic research on rescue trajectory planning in complex low-altitude flight environments, constructing a comprehensive evaluation model for complex low-altitude rescue flights based on the fuzzy analytic hierarchy process and proposing trajectory pre-planning methods based on two-dimensional and three-dimensional airspace grids and a multi-agent-based four-dimensional trajectory dynamic planning model. The research balances the safety, timeliness, and efficiency of rescue flights, comprehensively considering factors such as geographical environment, airspace characteristics, safety situation, rescue units, accident locations, and aircraft performance, achieving conflict-free trajectory planning for multiple aircraft through ant colony optimization algorithms and improved A* algorithms. This research provides a complete theoretical framework and algorithmic foundation for path planning of emergency UAV systems in complex environments, with its multi-dimensional dynamic planning method holding particular value for enhancing emergency response timeliness.

2.5. Ground Vehicle Routing and UAV Collaborative Optimization

Collaborative optimization between ground vehicles and UAVs represents the key to achieving system efficiency maximization, requiring solutions to complex synchronization and coordination challenges.

Integration of environmental factors brings sustainable development perspectives. Rodríguez-Espíndola et al. [31] integrated environmental concerns into collaborative decision-making for flood response. Their research innovations include disaster preparedness systems combining multi-objective optimization and geographical information systems; using cartographic models to avoid selecting flood-vulnerable facilities; simultaneously optimizing facility location, inventory prepositioning, resource allocation, and relief distribution. In the actual case of the 2013 Acapulco, Mexico, flood, research demonstrated that no single organization could independently cope with disasters of such scale; excessive numbers of government-deployed organizations led to high costs without achieving optimal service levels; and the proposed system outperformed current methods in both cost and service-level aspects.

Equity considerations have transformed optimization objectives. Liu et al. [32] integrated efficiency, equity, and effectiveness metrics into humanitarian facility location. Their distributionally robust facility location model considers random failure of network nodes and edges, uncertainty on both supply and demand sides, and equity requirements for coverage rates. Through robust chance-constrained methods, the research developed conic and linear approximations solved within outer approximation frameworks, embedding three acceleration techniques: branch-and-cut, in-and-out algorithms, and Benders decomposition. Case studies demonstrate that this model has clear advantages over traditional scenario-based approaches, particularly showing more stable performance under extreme events.

Research on coordination mechanisms reveals the value of cooperation. Balcik et al. [33] conducted in-depth analysis of coordination practices in humanitarian relief chains. Research findings include the following: numerous participants bring coordination challenges, as each organization has different missions, interests, and capabilities; information sharing is the foundation of effective coordination; resource pooling can significantly improve efficiency; trust building requires long-term efforts. They evaluated the applicability of commercial supply chain coordination mechanisms in relief environments, providing theoretical guidance for cross-organizational cooperation.

Inventory coordination demonstrates the importance of systems thinking. The stochastic programming model proposed by Davis et al. [34] determined how to position and distribute supplies in cooperative warehouse networks. Model innovations include constraints considering service equity, incorporating traffic congestion caused by evacuation behavior, and setting time constraints for effective response. Through extensive computational studies, they described conditions where prepositioning is beneficial and explored relationships between inventory placement, capacity, and coordination within networks. Research shows that effective utilization of short-term information (such as hurricane forecasts) can significantly improve operational-level supply allocation efficiency.

Innovative utilization of computational resources expands system capabilities. Zhu et al. [35] proposed an innovative scheme for UAV computational task scheduling using urban parking resources. In post-disaster rescue scenarios, they organized surviving parked vehicles into clusters to collaboratively compute applications uploaded by UAVs. Through constructing offloading strategies based on deep reinforcement learning, the system can intelligently interact with the environment to achieve optimization objectives, effectively improve task completion rates, and significantly reduce task execution costs. Simulation experiments demonstrate that this scheme outperforms baseline approaches in both task completion rate and execution cost, providing new insights for emergency computing in resource-constrained environments.

Optimized design of ground transportation infrastructure provides important support for ground–air coordination. In their research on non-conventional lane design at urban road intersections, Wang and Yang [36] proposed that optimally configuring the number and turning functions of approach and exit lanes to adapt to changing traffic demands, combined with optimal traffic signal timing, can significantly improve intersection capacity, reduce delays, and enhance safety. The research indicates the need to construct demand-responsive dynamic lane function conversion mechanisms and develop refined, precise spatiotemporal coordination control methods. These dynamic optimization concepts for ground traffic organization hold important reference value for ground–air coordination: in emergency scenarios, ground vehicles need to provide takeoff and landing space and material transfer support for UAVs. By borrowing dynamic configuration concepts from non-conventional lanes, flexible adjustment and optimal configuration of emergency corridors can be achieved, improving coordination efficiency between ground traffic and aerial delivery, providing fundamental support for constructing efficient ground–air integrated emergency response systems.

2.6. Emergency Response Decision Support System Construction

Decision support systems represent the key to integrating various technologies into comprehensive emergency solutions, requiring the achievement of information fusion, intelligent analysis, and decision assistance.

Breakthrough advances in path optimization have enhanced system efficiency. The multi-strategy fusion dynamic path planning method proposed by Xu et al. [37] represents the latest technological level. Their MSF-MTPO algorithm integrates five key innovations: an adaptive extended neighborhood A* algorithm dynamically adjusts the search range according to obstacle distribution; a bidirectional search mechanism searches simultaneously from start and end points; inflection point trajectory correction eliminates redundant inflection points; local tangent circle smoothing enhances trajectory smoothness; the artificial potential field method achieves dynamic obstacle avoidance, with successful verification on P200 UAV platforms proving its practicality.

Systematic application of computational intelligence has accelerated technological progress. The review study by Zhao et al. [38] comprehensively analyzed applications of computational intelligence in UAV path planning. Through systematic review of major journal and conference papers, the research found the following: computational intelligence methods clearly outperform traditional approaches in online and 3D problems; hybrid algorithms combine advantages of different methods; problem characteristics determine algorithm selection; trade-offs exist between real-time performance and optimality. This analysis provides guidance for researchers in selecting appropriate methods.

Innovative applications of optimization theory provide new solution approaches. Chen et al. [39] reconstructed artificial potential field methods as constrained optimization problems, effectively addressing limitations of traditional methods by introducing additional control forces and applying optimal control theory. Based on the detailed derivation processes of discrete UAV dynamic models, combined with path tracking validation using six-degree-of-freedom simulation models of quadrotor helicopters, the effectiveness of the improved methods was proven—the computed paths were shorter and smoother, with dead point problems effectively resolved.

Applications of deep learning have achieved end-to-end optimization. The large-scale complex environment UAV autonomous navigation method based on deep reinforcement learning proposed by Wang et al. [40] demonstrates the tremendous potential of AI. They formalized the problem as a Partially Observable Markov Decision Process (POMDP) and designed novel online DRL algorithms within an actor–critic framework based on two rigorously proven policy gradient theorems. Experimental results show that this method enables UAVs to autonomously perform navigation in virtual large-scale complex environments and can generalize to more complex and larger-scale three-dimensional environments and that the proposed algorithm outperforms existing state-of-the-art methods.

Optimization of visual detection tasks demonstrates specific application requirements. Phung et al. [41] developed enhanced discrete particle swarm optimization algorithms for UAV visual surface inspection. Through performance improvements via deterministic initialization, random mutation, and edge exchange, combined with GPU parallel computing frameworks, computation time was significantly reduced while maintaining unchanged hardware requirements. Experimental data from office building and bridge detection validated algorithm effectiveness, providing practical tools for infrastructure inspection.

Flexible collaboration mechanisms enhance system adaptability. The collaborative truck multi-drone routing and scheduling problem studied by Salama and Srinivas [42] introduced important innovations: allowing UAV launch and recovery at non-customer locations; considering three key decisions—customer assignment, vehicle routing, and activity scheduling; supporting both cyclic and acyclic UAV operations. Through optimization algorithms hybridizing simulated annealing and variable neighborhood search, numerical analysis demonstrates that using flexible sites significantly improves delivery efficiency compared to existing methods that restrict truck stopping locations.

Data-driven situational awareness serves as an important foundation for decision support systems. Wang et al. [43] demonstrated the tremendous potential of multi-source data fusion in traffic management through urban traffic state-sensing research. The research indicates that ETC systems have evolved from simple toll collection tools to comprehensive traffic management platforms, featuring unique advantages such as precise vehicle identification, extensive spatiotemporal coverage, and stable data quality. Through constructing multi-source data fusion frameworks, effective complementarity among ETC data, floating car data, and video detection data were achieved, significantly improving traffic state estimation accuracy. The development trends revealed by this research—from passive response to proactive prediction, from single functions to comprehensive services—provide important insights for emergency response decision support systems: through integrating multi-source heterogeneous information such as UAV sensor data, ground vehicle trajectory data, and traffic infrastructure data, more comprehensive and accurate emergency situational awareness and decision support capabilities can be constructed, achieving intelligent allocation and precise scheduling of emergency resources.

Multi-departmental collaborative decision-making represents a core challenge for emergency response systems. Wang et al. [44] proposed a multi-departmental collaborative emergency decision model based on multi-granularity semantic phrases to address the complexity of multi-departmental collaborative operations in low-altitude emergency rescue after emergencies. This model uses semantic phrases by decision departments to represent preference information of evaluation objects, achieves consistency transformation through conversion functions, and obtains the relative importance of decision departments, collaboration coefficients between departments, and the weight vectors of key indicators. Based on multi-departmental collaboration considerations, evaluation indices for different combination schemes are calculated to determine optimal solutions. Earthquake rescue case validation demonstrates that this collaborative decision model can comprehensively consider uncertainty in evaluation environments and mutual influences of multi-departmental collaborative operations, making decision processes and results more realistic. This provides important theoretical foundations and methodological guidance for cross-departmental coordination of emergency UAV systems.

2.7. Low-Altitude Economy and Urban Air Traffic Management Systems

The low-altitude economy, as an emerging economic form, is gradually becoming a global focus, providing important industrial foundations for emergency UAV system development. Sun et al. [45] conducted in-depth analysis of core elements in low-altitude economic development and the important role of high-tech innovation, noting that while the low-altitude economy has tremendous development potential, it still faces challenges such as incomplete legal frameworks and insufficient infrastructure construction. The research explored pathways for advancing low-altitude airspace management reform through developing low-altitude airspace digitalization systems and constructing hierarchical and classified airworthiness management systems for unmanned aircraft systems. These findings provide direction for the large-scale deployment of emergency UAV systems: establishing comprehensive low-altitude airspace digital management systems, constructing rapid approval and flight management mechanisms suitable for emergency scenarios, and promoting integrated development of emergency rescue and the low-altitude economy.

Precise handling of time window constraints enhances service quality. The branch-and-price-and-cut algorithm developed by Yin et al. [46] represents the latest advance in exact algorithms. They addressed truck–drone delivery problems with time window constraints, where drones can take off from trucks, independently serve one or more customers, and return to trucks at other nodes along truck routes. Through enhanced bidirectional labeling algorithms solving challenging pricing problems, combined with subset-row inequalities tightening lower bounds and enhancement strategies improving pricing problem solution efficiency, extensive numerical studies evaluated algorithm performance and assessed benefits of truck–drone delivery relative to truck-only delivery, providing insights for management decisions.

Uncertainty management ensures system robustness. The robust drone–truck delivery problem studied by Yang et al. [47] directly confronts challenges of road traffic uncertainty. Since uncertainty in ground traffic networks can not only cause service promise failures but also expose drones to danger, they focused on mitigating these risks when designing routing plans. Through developing precise branch-and-price methods, research demonstrated the following: robust solution variance reduced by up to 58%; the feasibility ratio (on-time performance) improved by up to 90%; mean values remained at comparable levels. These insights indicate that robust routes can be executed more frequently in practical usage.

Special requirements of urban emergency response have driven the development of specialized systems. The dynamic truck–UAV collaboration strategy proposed by Long et al. [48] specifically targets characteristics of urban environments. Their DTU collaboration strategy, based on urban emergency management characteristics, developed integrated truck–UAV collaborative scheduling models, proved their NP-hardness, and developed Tabu search-based integrated scheduling algorithms. Comprehensive experiments demonstrate the following: the DTU strategy performs excellently in urban areas with high-density road networks; the system shows resilience to different degrees of road network disruption; it significantly outperforms parallel and flying sidekick strategies in large-scale emergency response scenarios; performance advantages become more pronounced as demand density increases.

Enhanced large-scale deployment capabilities expand application scope. The research by Madani et al. [49] addressed complex scenarios where drones can visit multiple customers per dispatch. Their introduced practical attributes include the following: allowing trucks to launch and retrieve drones at both customer and non-customer nodes; supporting both cyclic and acyclic drone operations; handling multiple visits and multiple launch–retrieval locations. Through developing effective variable neighborhood search methods that innovatively categorize neighborhoods and maintain their individual impacts through adaptive selection schemes, sensitivity analysis revealed influences of key drone parameters, providing guidance for system design. The review by Wang et al. [50] constructed a complete theoretical framework for emergency facility location problems. Their classification includes the following: deterministic models applicable to cases with known parameters; stochastic models handling uncertainty with known probability distributions; dynamic models considering time-varying characteristics; robust models addressing worst-case scenarios. The research provides methodological guidance for locating distribution centers, warehouses, shelters, and medical facilities, holding important value for constructing emergency UAV ground support networks.

2.8. Intelligent Transportation System Integration and Coordination

Intelligent transportation system integration provides technological foundations and information support for ground–air integrated coordination.

Innovative assessment methods for humanitarian relief networks. The collaborative truck–drone system developed by Zhang et al. [51] for post-disaster transportation network assessment fills gaps in this field. The system includes camera-equipped drones that can launch from trucks to collect node and link information, with trucks used for drone recovery and charging. Through decomposing the problem into path-based master problems and two sub-problems solved within a column generation framework, research found the following: the algorithm can obtain optimization gaps of less than 10% for all terminated instances within predetermined time limits; actual cases from Istanbul’s Kartal district demonstrated model practicality; management insights were developed for humanitarian relief agency applications.

Big data-driven intelligent identification enhances decision quality. The research by Song et al. [52] demonstrates the application potential of deep learning in emergency logistics. Through big data-driven deep learning models, the system achieved the following: real-time processing of logistics identification and routing; efficient configuration of six optimized routes within 57.19 km; effective management of emergency flow control. This data-driven approach provides new paradigms for intelligent emergency response.

Urban airspace design ensures operational safety. The research by Doole et al. [53] addressed traffic organization issues in highly constrained airspace. Through applying road analogies of two-way and one-way streets, imposing horizontal structures, and employing heading-altitude rules to vertically segment cruising traffic, fast-time simulation experiments evaluated safety, stability, and efficiency under different traffic demand densities. The results demonstrate the following: vertically segmented altitude layers combined with horizontal constraints represent effective ways to organize drone traffic; transition altitude settings are crucial for turning safety; rational airspace design can support high-density operations.

Optimized scheduling for on-demand urban air traffic. Kleinbekman et al. [54] developed specialized scheduling systems targeting eVTOL aircraft characteristics. Unlike fixed-wing aircraft or helicopters in commercial aviation, eVTOL aircraft have different flight dynamics, limited battery energy, and limited landing spots at vertiports. They utilized mixed-integer linear programming to compute optimal required arrival times, safely separating aircraft based on remaining battery charge and vertiport capacity to achieve minimum delays. The proposed vertiport terminal area airspace design concept, combined with existing energy-efficient trajectory optimization tools, provides foundations for safe and efficient UAM operations.

Data collection optimization for IoT devices. Samir et al. [55] studied trajectory planning problems for UAVs collecting data from time-constrained IoT devices. In smart city applications, low-resource IoT devices need to upload sensor data before hard deadlines, otherwise data becomes outdated and loses value. Through jointly optimizing UAV trajectory and wireless resource allocation, the number of served IoT devices is maximized. The research developed high-complexity branch, reduce, and bound algorithms to find global optimal solutions, as well as effective sub-optimal algorithms based on successive convex approximation. Extensive simulations demonstrate clear advantages of the proposed methods compared to greedy algorithms based on distance and deadline metrics.

Development trends of next-generation traffic control systems. The research by Wang and Yang [56] analyzed key technologies in China’s urban traffic control systems. The research indicates the following: traditional systems over-rely on IT technology while neglecting fundamental traffic control theories; real-time interaction provides new support for traffic optimization control; rich traffic control interaction conditions and comprehensive data should be utilized to create next-generation systems. New systems should demonstrate high refinement, precision, better responsiveness, and enhanced intelligence—characteristics that hold important reference value for ground–air coordination in emergency scenarios.

Application prospects of graph neural networks. The comprehensive survey by Jiang and Luo [57] analyzed applications of graph neural networks in traffic forecasting. Through a systematic review of 207 papers, they found that graph convolutional networks and graph attention networks are the most promising solutions, achieving state-of-the-art performance across various traffic forecasting problems. These technologies can be effectively applied to traffic flow prediction and path optimization in emergency scenarios.

Comprehensive integration of sensor technologies. Guerrero-Ibáñez et al. [58] conducted in-depth analysis of sensor technologies in intelligent transportation systems. Modern society faces transportation problems, including congestion, safety, and pollution. Through seamless integration of vehicles and sensing devices, their sensing and communication capabilities can be leveraged to achieve intelligent transportation systems. The research explored how sensor technologies integrate with transportation infrastructure and how safety, traffic control, and infotainment applications benefit from multiple sensors deployed across different ITS elements, providing technical pathways for constructing comprehensively perceptive emergency response systems.

Enhanced collaboration through V2X communications. The research by Hobert et al. [59] demonstrates the role of V2X communications in supporting cooperative autonomous driving. Two key cooperative features—cooperative sensing and cooperative maneuvering—enable vehicles to exchange information collected from local sensors and coordinate maneuvers. First-generation V2X communication systems were primarily designed for driver warning applications focused on road safety and traffic efficiency, not targeting autonomous driving use cases. The research described target use cases, identified communication requirements, analyzed missing features in current ETSI standards, and provided specification sets for standard revisions and extensions supporting cooperative autonomous driving. These technological advances lay foundations for ground–air collaborative communications in emergency scenarios.

Table 2. Quantitative performance comparison of key UAV–ground vehicle collaborative methods in emergency scenarios.

Method/System	Performance Metrics	Quantitative Outcomes	Application Scenario	Research Study
ULRB Framework (UAV–Bus Integration)	Delivery range extension; battery aging reduction	5.54× average range extension (up to 7.73×); 3.26× battery aging deceleration (up to 4.16×)	Urban logistics delivery	[7]
TGA-BB Algorithm	Response time; runtime efficiency	52.37% response time reduction vs. TGA-PSO; 97.95% runtime reduction vs. TGA-SA	Multi-stage emergency material scheduling	[17]
UAV Disinfection System	Coverage area and time	2 km area coverage within 10 min	COVID-19 pandemic response	[18]
Standardized UAV Rescue Configuration	Container utilization; service duration	71.4% container space utilization; 188-day rescue capability	Disaster relief operations	[16]
Constraint Programming vs. Mixed-Integer Optimization	Computational efficiency; solution quality trade-off	7–30× faster processing; 4–15% optimality reduction	Multi-UAV fuel-constrained coordination	[26]
Robust Drone–Truck Delivery	Solution robustness; on-time performance	58% variance reduction; 90% feasibility ratio improvement	Uncertain traffic conditions	[47]
ALNS Metaheuristic	Speed–efficiency relationship	6–8% efficiency improvement per 10% UAV speed increase	Capacitated multi-truck scenarios	[22]
LSTM-DNN Prediction Model	Travel time prediction accuracy	Best MAPE performance in 30-min sliding window predictions	Urban traffic prediction	[24]

summarizes quantitative performance outcomes of key UAV–ground vehicle collaborative methods reported in emergency delivery research. The comparison demonstrates significant improvements achieved through different technological approaches, with delivery range extensions of up to 7.73 times, response time reductions exceeding 50%, and coverage capabilities reaching 2 km areas within 10 min. These quantitative results validate the effectiveness of collaborative optimization strategies in emergency scenarios and provide benchmarks for system performance evaluation.

Through systematic analysis of the above eight core technological domains,

Table 3. Summary of key research on UAV delivery and ground transfer scheduling technologies in emergency scenarios.

Technology Domain	Key Technical Features	Primary Application Scenarios	Technology Maturity	Major Contributions and Innovations
UAV Emergency Delivery System Architecture and Optimization	Stochastic optimization, energy management, airspace congestion control	Disaster relief, uncertain demand delivery	Medium	Chance-constrained programming (CCP), second-order cone programming reformulation, 3D low-altitude airspace equilibrium models
Ground–Air Integrated Transportation Coordination Mechanisms	Multi-objective optimization, bi-stage planning, collaborative operations	Medical emergency delivery, blood product transportation	Higher	Multi-objective integer programming (MOIP), trajectory optimization algorithms, Flying Sidekick Traveling Salesman Problem (FSTSP)
Emergency Logistics and Rescue Operation Optimization Strategies	Multi-UAV coordination, adaptive algorithms, time dependency	Large-scale delivery, complex environment operations	Medium	Multiple Flying Sidekicks TSP, Adaptive Large Neighborhood Search (ALNS), time-dependent routing
UAV Path Planning and Scheduling Algorithm Innovation	Intelligent algorithms, computational intelligence, optimal control	Complex environment navigation, obstacle avoidance planning	Higher	Multi-strategy fusion (MSF-MTPO), adaptive neighborhood A*, improved artificial potential field methods
Ground Vehicle Routing and UAV Collaborative Optimization	Robust optimization, multi-criteria evaluation, coordination mechanisms	Resource prepositioning, flood response, facility location	Medium	Distributionally robust models, multi-objective optimization and GIS integration, chance-constrained methods
Emergency Response Decision Support System Construction	Deep reinforcement learning, parallel computing, flexible coordination	Autonomous navigation, visual detection, last-mile delivery	Lower	POMDP modeling, GPU parallel DPSO, flexible launch–recovery mechanisms
Low-Altitude Economy and Urban Air Traffic Management Systems	Exact algorithms, robust planning, dynamic collaboration	Urban delivery, uncertain environments, emergency response	Medium	Branch-and-price-and-cut algorithms, robust optimization, dynamic truck–UAV collaboration (DTU)
Intelligent Transportation System Integration and Coordination	Network assessment, deep learning, airspace design	Post-disaster assessment, emergency logistics, urban airspace	Lower	Column generation framework, big data-driven models, constrained airspace design

summarizes major research achievements in UAV delivery and ground transfer scheduling technologies for emergency scenarios. From a technological maturity perspective, UAV path planning and ground–air integrated transportation coordination technologies are relatively mature, with numerous successful practical application cases. while emergency response decision support systems and intelligent transportation system integration technologies remain in developmental stages, primarily at theoretical research and small-scale validation levels. From application scenario perspectives, medical emergency delivery and disaster relief represent the primary current application domains with relatively sufficient technological validation, while urban airspace management and large-scale collaborative operations still face numerous challenges. Overall, this field has formed a relatively complete technological system, but further development is needed in system integration, standardization, and industrialization aspects.

3. Current Issues and Challenges

3.1. Regulatory Framework and Standardization Deficiencies

The primary challenge facing UAV delivery in emergency scenarios is the inadequacy of regulatory frameworks and the absence of industry standards. The comprehensive study by Stöcker et al. [60] demonstrates that although countries have gradually established national legal frameworks since the early 2000s, existing aviation regulations are primarily designed for traditional manned aircraft, with insufficient consideration of UAV-specific characteristics. Their exploratory investigation of global UAV regulations reveals that while all UAV regulations share a common objective—minimizing risks to other airspace users and to people and property on the ground—they exhibit significant variations across all comparative variables.

Particularly in emergency situations, existing regulatory frameworks present the following critical issues:

(1)

Absence of clear provisions for special flight rules and simplified approval procedures, severely impacting emergency response timeliness;

(2)

Substantial differences in regulatory requirements among different countries and regions, hindering cross-border emergency rescue operations;

(3)

Beyond explicit legal frameworks, market forces such as industry design standards and reliable information about UAVs as public goods are expected to influence future development, but current standardization efforts are progressing slowly.

This absence of regulation and standardization not only restricts large-scale technology deployment but also, more critically, may delay rescue operations during crucial emergency moments, causing irreversible losses.

The scalability challenges facing UAV–ground vehicle collaborative systems extend beyond regulatory constraints to encompass fundamental performance limitations revealed through existing research. While regulatory frameworks present significant deployment barriers, as discussed above, quantitative analysis from the reviewed studies demonstrates additional technical constraints that compound these challenges.

System performance optimization shows promising but limited scalability potential. The ULRB framework proposed by Pan et al. [7] demonstrates significant individual system improvements with UAV delivery range extensions averaging 5.54 times normal capacity, reaching up to 7.73 times in optimal conditions, while simultaneously reducing battery aging by 3.26 times on average. However, these performance gains are primarily validated for single or small-scale UAV operations rather than large fleet deployments required for major emergency scenarios.

Algorithmic efficiency improvements offer mixed scalability prospects. The TGA-BB algorithm developed by Gong et al. [17] achieves substantial performance enhancements, reducing average response time by approximately 52.37% compared to TGA-PSO and runtime by 97.95% compared to TGA-SA. Similarly, the research by Sacramento et al. [22] demonstrates that speed optimization strategies achieve 6–8% efficiency improvements per 10% increase in UAV speed, but these improvements face diminishing returns as operational complexity increases with fleet size.

Advanced path planning algorithms demonstrate significant technical progress but highlight coordination complexity challenges. The enhanced dynamic artificial potential field (ED-APF) approach proposed by Jayaweera and Hanoun [61] addresses critical limitations in UAV path planning for reconnaissance and look-ahead coverage support for mobile ground vehicles. This method formulates path planning as both a follow and cover problem, adopting vertical sinusoidal paths that adapt relative to ground vehicle position and velocity while extending reconnaissance capabilities beyond ground sensor ranges. The ED-APF technique demonstrates superior performance compared to general artificial potential field methods in dynamic and obstacle-populated environments, with validation conducted using Robot Operating System (ROS) and Gazebo-supported PX4-SITL simulations. However, the method’s effectiveness is primarily demonstrated for individual UAV–ground vehicle pairs, and scaling such sophisticated coordination algorithms to multi-UAV fleets presents exponential complexity growth challenges that remain unresolved for large-scale emergency deployment scenarios.

Coverage and operational capacity metrics highlight current system limitations. The UAV-based network system by Kumar et al. [18] can effectively cover 2 km areas within 10 min timeframes for emergency operations, while the standardized rescue configuration proposed by Zhao et al. [16] achieves 71.4% container space utilization supporting 188-day rescue operation requirements. However, scaling these capabilities to city-wide or regional emergency response levels remains unvalidated, with robust optimization approaches by Yang et al. [47] showing variance reductions of up to 58% and feasibility improvements of up to 90%, primarily demonstrated in limited-scale scenarios.

3.2. Insufficient Adaptability to Adverse Environments

Emergency scenarios are often accompanied by extreme adverse environmental conditions, posing severe challenges to UAV system reliability. Shakhatreh et al. [62] indicate in their comprehensive survey that key research challenges facing UAVs in civilian applications include charging challenges, collision avoidance and swarming challenges, and network- and security-related challenges. Research reveals that most path planning algorithms assume ideal flight conditions, lacking real-time adaptation mechanisms for dynamic meteorological conditions.

Specifically, insufficient adaptability to adverse environments manifests across multiple levels:

(1)

Under extreme weather conditions such as strong winds, heavy rain, and low temperatures, existing UAVs experience dramatic reductions in flight stability and navigation accuracy, with battery performance also significantly degraded;

(2)

Electromagnetic interference and communication disruption issues in complex terrain environments remain unresolved, particularly in mountainous areas and urban canyons;

(3)

There is insufficient research on the impact of environmental factors such as smoke, dust, and toxic gases at disaster sites on sensor performance.

Research indicates that civilian infrastructure is expected to dominate over USD 45 billion in UAV usage market value, but if adverse environment adaptability issues cannot be resolved, this enormous market potential will not be fully realized.

3.3. Cybersecurity and Data Protection

Cybersecurity threats to emergency UAV systems are becoming increasingly serious, creating a key bottleneck constraining their large-scale application. The security analysis study by Yaacoub et al. [63] thoroughly explores attacks, limitations, and recommendations for UAV systems. Research reveals that both the probability and frequency of malicious UAV use are high, with potentially very dangerous and devastating effects.

The research provides detailed analysis of emerging threats from UAVs in cyberattacks, including exploitation of UAV vulnerabilities within communication links and security risks in intelligent devices and hardware (including smartphones and tablets). Through implementing simulated attacks, the authors demonstrate realistic attack scenarios, detailing how to execute attacks on given UAVs following hacking cycles. This research reveals multiple security challenges facing emergency UAV systems: existing research severely lacks a security architecture design for distributed systems; security mechanisms such as data transmission encryption, identity authentication, and access control urgently need improvement; substantial sensitive information involved in emergency scenarios (including disaster data, personnel locations, resource allocation, etc.) lacks effective protection measures. The research emphasizes the need to develop detective, protective, and preventive countermeasures to address these growing security threats.

3.4. Human–Machine Interaction and Usability

Existing systems exhibit severe deficiencies in human–machine interaction design, failing to meet actual operational requirements in emergency scenarios. Through a case study of Beijing South Railway Station, Chen et al. [64] developed weight-based multi-criteria evaluation methods, finding that transfer performance assessment of large transportation hubs requires simultaneous consideration of quantifiable factors and subjective perceptions, which operate at different evaluation scales and boundaries.

The insights from this research hold significant importance for emergency UAV systems: emergency personnel often need to make rapid decisions under high-pressure environments, while existing systems primarily focus on technical performance optimization with insufficient consideration of operators’ cognitive load; complex operational interfaces and decision processes may lead to operational errors or delays, which could cause catastrophic consequences in emergency scenarios; system design needs to fully consider the requirements of operators with different technical levels, ensuring that even non-professional personnel can quickly operate the system during emergencies. The research demonstrates through applications of Multi-Level Grey Evaluation (MGE) and the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) that comprehensive evaluation methods are crucial for enhancing system usability. This is particularly useful for designers and operators evaluating transfer performance and service levels of large transportation terminals, and the same principles should be applied to human–machine interface design in emergency UAV systems.

Emergency UAV–ground vehicle collaborative systems present unique human–machine interface challenges that extend beyond traditional usability concerns. Cognitive load management becomes critical as operators must simultaneously monitor multiple UAV status indicators, ground vehicle positions, environmental conditions, and mission progress while making time-sensitive decisions under high stress. Effective interface design requires clear information hierarchies that prioritize essential data while minimizing unnecessary visual complexity.

Situational awareness maintenance presents significant challenges when coordinating aerial and ground assets across different operational domains. Operators require integrated displays that combine real-time video feeds, asset positions, mission status, and environmental hazards in coherent visualizations. The interface must clearly distinguish between UAV and ground vehicle capabilities and current operational states to prevent confusion during critical operations.

Automation-level optimization requires a careful balance between system autonomy and human control. While automation reduces operator workload, excessive reliance may decrease situational awareness when systems encounter unexpected conditions. Emergency interfaces must provide clear automation status indicators and intuitive manual override capabilities to maintain operator understanding and control authority.

Error prevention mechanisms are essential given the high-stakes nature of emergency operations. Interface design must incorporate confirmation protocols for critical actions and provide clear, actionable feedback without creating operational delays. The challenge involves balancing comprehensive error prevention with the speed requirements inherent in emergency response scenarios.

These interface design challenges demonstrate that emergency UAV systems require integrated approaches combining technical capabilities with human factor considerations to achieve optimal operational effectiveness.

3.5. Cost-Effectiveness and Sustainability

Although technical feasibility has been widely validated, the lack of cost–benefit analysis severely hinders the actual deployment of emergency UAV systems. Through a systematic literature review of 111 interdisciplinary publications (2013–March 2019), Kellermann et al. [65] conducted an in-depth analysis of UAV application prospects in parcel and passenger transportation. The research identified 2581 relevant citations, subdivided into anticipated barriers (426), potential problems (1037), proposed solutions (737), and expected benefits (381).

The research reveals the complexity of cost–benefit assessment:

(1)

The debate is characterized by major technical and regulatory issues and barriers that are considered to prevent or impede UAV use for parcel and passenger transportation;

(2)

Clear economic expectations coexist with quite complex and differentiated social and environmental impact concerns;

(3)

Careful scrutiny of the most prevalent transportation-related promises (traffic reduction, travel time savings, and environmental mitigation) reveals an urgent need to provide scientific evidence for promises associated with UAV transportation use.

Regarding sustainability, the research indicates the following: UAV system procurement costs, maintenance expenses, and personnel training investments are quite substantial, requiring deeper evaluation of economic viability compared to traditional delivery methods; issues such as battery endurance, charging infrastructure construction, and equipment upgrades all need comprehensive consideration; the complexity and uncertainty of environmental impact assessment make sustainable development goals difficult to quantify. The research concludes that the debate on UAV transportation requires further qualification, with greater emphasis on social benefits and public participation, which holds important guidance significance for UAV system deployment in emergency scenarios.

Beyond economic considerations, emergency UAV systems face significant ethical and social acceptance challenges that can substantially impact deployment success. Privacy concerns arise from UAV surveillance capabilities and sensitive data collection during emergency operations, particularly when systems collect disaster site imagery, personnel locations, and potentially personal information about affected populations [63]. Social acceptance varies significantly across demographics and cultural contexts, with concerns about surveillance, noise pollution, and potential misuse creating resistance even in emergency contexts. The research by Chen et al. [64] demonstrates how subjective perceptions operate alongside quantifiable factors, highlighting the importance of addressing public concerns through transparent communication and community engagement.

Social equity considerations emerge as critical challenges, as high-tech emergency systems may create disparities between well-funded urban areas and resource-constrained regions. As highlighted in humanitarian logistics research [13], emergency response systems must balance efficiency objectives with equity considerations to ensure a fair distribution of services. Additionally, accountability frameworks for automated emergency decisions remain underdeveloped as UAV systems become more autonomous and AI-enhanced, raising questions about legal and moral responsibility when automated systems make critical decisions affecting human welfare. These ethical and social challenges require proactive community engagement and comprehensive frameworks for emergency UAV deployment to achieve the social license necessary for effective long-term implementation.

3.6. Implementation Barriers and Deployment Challenges

Beyond cost considerations, emergency UAV systems face multiple implementation barriers that significantly impact deployment feasibility and operational success. Technical barriers include insufficient battery endurance for extended operations, limited payload capacity constraining mission scope, and communication range limitations in remote disaster areas. The research reveals that existing UAVs experience dramatic performance reductions under adverse weather conditions such as strong winds, heavy rain, and low temperatures, with battery performance also significantly degraded [62]. Additionally, electromagnetic interference and communication disruption in complex terrain environments remain unresolved, particularly affecting operations in mountainous areas and urban canyons.

Infrastructure and logistical barriers present substantial challenges for large-scale deployment. Charging infrastructure construction requires significant investment and strategic planning, particularly for remote emergency operations where traditional power sources may be unavailable. Equipment maintenance and repair capabilities must be distributed geographically to ensure rapid response times, creating additional logistical complexity and cost burdens. The standardized rescue configuration research [16] demonstrates that while optimal resource utilization can be achieved (71.4% container space efficiency), the complexity of coordinating multiple UAV types and ground support systems creates substantial operational challenges.

Regulatory and policy barriers extend beyond basic legal frameworks to encompass operational approval processes, insurance requirements, and liability frameworks. The comprehensive regulatory analysis by Stöcker et al. [60] reveals that while countries have established national legal frameworks, existing regulations lack provisions for special emergency flight rules and simplified approval procedures, severely impacting response timeliness. Cross-jurisdictional coordination challenges become particularly acute during large-scale disasters that span multiple administrative boundaries, where different regulatory requirements may conflict or create operational delays.

Human resource and training barriers require substantial investment in specialized personnel development. Emergency UAV operations demand highly trained operators capable of managing complex multi-modal coordination under high-stress conditions. The human–machine interaction challenges identified in our analysis [64] demonstrate that existing system interfaces fail to meet operational requirements, potentially leading to operational errors or delays that could cause catastrophic consequences. Training programs must address not only technical operation skills but also emergency-specific decision-making, multi-agency coordination, and system troubleshooting under adverse conditions.

4. Future Development Directions

4.1. Distributed Decision Architecture

Future research should focus on developing distributed decision architectures to enhance system robustness, scalability, and response speed. Zhang et al. [66] proposed BCPay—a blockchain-based fair payment framework for cloud computing outsourcing services, providing innovative insights for distributed system design. This research first achieved secure fair payment mechanisms without relying on any third parties (trusted or untrusted), offering important implications for distributed coordination in emergency scenarios.

Core innovations of the BCPay framework include the following: first, achieving “soundness” and “robust fairness”, where fairness can resist eavesdropping and malleability attacks; second, demonstrating extremely high efficiency in terms of transaction numbers and computational costs; third, as illustrative applications, the research further constructed blockchain-based provable data possession schemes for cloud computing and blockchain-based outsourcing computation protocols for fog computing.

These technological breakthroughs provide new possibilities for distributed decision-making in emergency UAV systems: blockchain-based decentralized coordination mechanisms can ensure that individual UAV nodes can still autonomously coordinate to complete tasks when central control systems fail; edge computing-enhanced real-time decision capabilities enable each UAV to become an independent decision unit, significantly improving system response speed; autonomous negotiation multi-agent systems can achieve dynamic optimization of resource allocation in complex emergency environments. Future research should deeply explore how to combine these distributed technologies with special requirements of emergency scenarios, constructing truly decentralized emergency response networks.

4.2. Robustness and Reliability Enhancement

Robust optimization methods for uncertain environments are key to enhancing emergency system reliability. The Variable Neighborhood Search (VNS) strategy proposed by Menéndez et al. [67] provides efficient solutions for order batching problems. In warehouse operational management optimization problems, this research minimizes the time required to collect all orders by grouping received orders into batches, surpassing existing technology levels in both quality and computational time.

Key contributions of the research include the following: first, proposing multiple strategies based on VNS methodology to solve problems; second, confirming method superiority through non-parametric statistical tests; third, demonstrating how single pickers can complete each batch collection without exceeding capacity limits. These achievements have direct reference value for emergency UAV scheduling.

In emergency scenarios, robustness enhancement needs to unfold across multiple dimensions: distributionally robust optimization can handle multiple uncertainties in demand, supply, and environment; multi-stage stochastic programming can address dynamically changing emergency demands; adaptive algorithms can adjust strategies based on real-time feedback; fault tolerance mechanisms ensure that single-point failures do not cause system collapse; redundant design provides necessary backup capabilities; autonomous recovery capabilities enable systems to quickly rebuild functionality after partial damage. Future research should integrate these robustness technologies into unified frameworks, constructing emergency response systems capable of stable operation under extreme conditions.

4.3. Cross-Domain Collaboration Mechanisms

Establishing standardized cross-domain collaboration mechanisms is a necessary condition for achieving large-scale emergency response. The research by Michailidis et al. [68] provides new perspectives for cross-domain collaboration by proposing a security-aware computation offloading framework for mobile edge computing (MEC)-enabled IoT networks operating in environments with aerial and ground eavesdroppers.

Research innovations include the following: first, deploying UAVs as both aerial MEC servers and relays, forwarding partial tasks to ground access points for computation; second, further enhancing computation offloading capabilities by integrating Reconfigurable Intelligent Surface (RIS) units near access points; third, jointly optimizing transmission power allocation, time-slot scheduling, task allocation, and RIS phase shifts to maximize minimum Secure Computation Efficiency (SCE).

This research provides important insights for cross-domain collaboration in emergency scenarios: unified API standards and data formats need to be established among different organizations and systems; communication protocol compatibility is the foundation for achieving seamless integration; security must be the primary consideration for cross-domain collaboration; optimization algorithms need to handle non-convex complex problems. The iterative algorithms proposed in the research, through Dinkelbach and Block Coordinate Descent (BCD) methods, separately handle fractional objective functions and coupled optimization variables, providing effective tools for solving complex optimization problems in cross-domain collaboration. Future work should establish more comprehensive cross-domain collaboration standard systems, ensuring that resources from different sources can be rapidly integrated and efficiently utilized in emergency scenarios.

4.4. Emerging Technology Integration

Deep integration of emerging technologies will bring revolutionary changes to emergency UAV systems. First is traffic state sensing and multi-source data fusion, where data is crucial for any algorithm or platform—if data problems occur, the application effectiveness of algorithms and platforms will also suffer [69]. Therefore, the importance of multi-source data fusion and analysis for emergency rescue is self-evident.

The research by Ye et al. [70] further demonstrates the application prospects of artificial intelligence in emergency communications. Their DQN-based shaped reward function method improves training efficiency by designing enhanced reward functions, effectively alleviating sparse reward and prolonged training time problems in reinforcement learning algorithms. Experimental results show that this method effectively shortens training time while improving convergence rates.

These studies point to directions for technology integration in emergency UAV systems: artificial intelligence and machine learning technologies can achieve intelligent decision-making and autonomous control; Internet of Things technology can construct comprehensive sensing networks; 5G/6G communications provide high-speed, low-latency data transmission; edge computing enables real-time processing and rapid response; digital twin technology supports system simulation and prediction; augmented reality technology improves operational interfaces; quantum computing provides new solution capabilities for complex optimization problems. Future research should explore deep integration of these technologies, constructing intelligent, networked, and autonomous next-generation emergency response systems.

4.5. Practical Application Validation and Industrialization

Strengthening validation in practical application scenarios and advancing industrialization is the necessary path for technological development. The data-driven smart manufacturing framework proposed by Tao et al. [71] provides important reference for emergency UAV system industrialization. The research indicates that advances in internet technology, Internet of Things, cloud computing, big data, and artificial intelligence have profoundly impacted manufacturing, with the volume of data collected in manufacturing growing and big data providing tremendous opportunities for today’s manufacturing paradigm transformation to smart manufacturing.

This holds important implications for emergency UAV system industrialization: standardized performance evaluation systems need to be established to ensure technology reliability and effectiveness; large-scale demonstration application projects can validate actual technology effects and identify potential problems; improvement of industry–academia–research cooperation mechanisms can accelerate technology transfer; business model innovation is key to achieving sustainable development; supporting policy and regulatory frameworks are necessary conditions for industrialization.

Particularly important is the research emphasis on big data empowerment: through collecting and analyzing massive data from emergency response processes, system performance can be continuously optimized; predictive maintenance can identify and resolve potential problems in advance; intelligent scheduling systems can make optimal decisions based on historical data and real-time information; digital twin technology supports virtual validation and optimization. Future work should establish complete data-driven emergency response systems, promoting continuous technology improvement and industrialization processes through sustained data collection, analysis, and feedback.

Our systematic analysis across the eight core technological domains reveals several critical research gaps that represent priority areas for future investigation. In UAV emergency delivery system architecture and optimization, current research primarily focuses on single-objective optimization under ideal conditions, with significant gaps in multi-objective optimization frameworks that simultaneously consider energy efficiency, mission reliability, and cost-effectiveness under uncertain emergency environments. While stochastic optimization approaches have shown promise in addressing demand uncertainty, comprehensive frameworks integrating weather variability and infrastructure damage assessment remain underdeveloped.

Ground–air integrated transportation coordination shows substantial gaps in real-time synchronization mechanisms and dynamic resource reallocation strategies. Current research demonstrates successful coordination in specific scenarios, but scalable protocols for managing hundreds of UAVs and ground vehicles simultaneously during large-scale disasters are lacking. Cross-modal failure recovery mechanisms and adaptive coordination under partial communication loss represent critical unaddressed challenges that could significantly impact system reliability during actual emergency deployments.

Emergency logistics optimization reveals gaps in equity–efficiency trade-off frameworks and dynamic priority adjustment mechanisms. While multi-UAV coordination research [21] and adaptive algorithms [22] have advanced efficiency optimization, comprehensive consideration of fairness constraints and vulnerable population prioritization remains limited. Additionally, real-time demand forecasting integration with logistics optimization remains underdeveloped, particularly for compound disaster scenarios where demand patterns differ significantly from historical data.

Path planning and scheduling algorithms face unresolved challenges in collective intelligence implementation and swarm coordination under communication constraints. Individual UAV path optimization has advanced significantly with multi-strategy approaches [37] and computational intelligence methods [38], but distributed decision-making for UAV swarms in GPS-denied environments and dynamic obstacle-rich scenarios requires substantial additional research. The integration of machine learning approaches with traditional optimization methods for real-time replanning represents a critical research frontier.

Collaborative optimization between ground vehicles and UAVs shows significant gaps in heterogeneous fleet management and cross-platform resource sharing mechanisms. Current research demonstrates coordination benefits in specific applications but lacks comprehensive frameworks for managing mixed autonomous and human-operated vehicle fleets. Dynamic role assignment and task reallocation during mission execution require substantial additional research.

Decision support system construction faces critical gaps in explainable AI implementation and multi-stakeholder decision frameworks. While advanced approaches show promise in autonomous navigation [40] and optimization applications [41], interpretability requirements for emergency decision-making and integration with existing emergency management protocols remain underdeveloped. Human–AI collaboration frameworks specifically designed for high-stress emergency environments represent urgent research needs.

These identified research gaps directly inform the three-stage development roadmap presented in

Table 4. Future development roadmap for UAV delivery technology in emergency scenarios.

Development Stage	Time Period	Main Tasks	Key Technologies	Expected Outcomes
Technological Breakthrough and Theoretical Validation Period	2025–2027	Resolve core theoretical problems and key technological bottlenecks	Blockchain technology, edge computing, reinforcement learning algorithms, cryptographic protocols, 5G communications	Theoretical breakthroughs in distributed decision architectures, innovation in robustness optimization algorithms, formulation of cross-domain collaboration standards
System Integration and Scale Validation Period	2028–2031	Focus on engineering implementation and systematic integration of technologies	6G communications, quantum communications, digital twins, autonomous control, multi-source data fusion	Multi-modal collaborative system integration, large-scale demonstration applications, improvement of standardization systems, establishment of regulatory frameworks
Industrialization Promotion and Comprehensive Application Period	2032–2035	Achieve large-scale commercial application and social value creation of technologies	Quantum computing, brain–computer interfaces, holographic communications, autonomous decision AI, sustainable energy technologies	Maturation of business models, comprehensive large-scale deployment, advancement of international cooperation, realization of significant social benefits

, with theoretical breakthrough requirements (2025–2027) addressing fundamental algorithmic and coordination challenges, system integration needs (2028–2031) focusing on scalability and interoperability gaps, and industrialization requirements (2032–2035) addressing practical deployment and standardization challenges. The entire development roadmap reflects a complete technological development path from theoretical validation to engineering implementation to large-scale application, with each stage interconnected and progressively advancing, providing clear guidance for the leapfrog development of emergency rescue technologies.

4.6. Standardized Performance Metrics

Based on the comprehensive analysis of existing research, we propose a standardized performance evaluation framework for UAV–ground vehicle collaborative systems in emergency scenarios. This framework addresses the current lack of unified evaluation criteria that hinders systematic comparison across different technological approaches and impedes evidence-based system selection in emergency management.

The proposed framework encompasses five core metric categories. Operational efficiency metrics include response time (measured as the time interval from emergency request initiation to first service delivery), coverage efficiency (calculated as the ratio of successfully served demand points to total emergency demands within specified time windows), and resource utilization rate (percentage of available UAV and ground vehicle capacity effectively deployed). Service quality and equity metrics encompass the demand satisfaction rate (ensuring balanced service distribution across affected populations), service reliability (measured through successful delivery completion rates and system uptime), and the geographic equity index (evaluating fair coverage distribution across different disaster-affected areas). Economic and sustainability metrics cover total operational cost per service unit delivered, the cost–benefit ratio compared to traditional emergency response methods, and energy efficiency expressed as service units delivered per energy unit consumed. System robustness metrics evaluate the environmental adaptability score (performance degradation under adverse conditions), scalability index (system capacity expansion capabilities), and fault tolerance rate (ability to maintain operations despite component failures). Safety and security metrics include the safety incident rate per operational hour, the cybersecurity resilience score, and the human–machine interface usability index.

Each metric should be evaluated on a normalized 0–100 scale to enable cross-system comparisons. As this review provides a foundational framework, future research should focus on determining appropriate weighting schemes for different emergency scenarios. For instance, medical emergency response may prioritize response time and service reliability, while large-scale disaster relief operations may emphasize coverage efficiency and resource utilization. The development of scenario-specific metric weighting represents a critical area for subsequent empirical validation and practical implementation studies.

This standardized framework enables systematic comparison across different UAV-UGV collaborative approaches, supports evidence-based decision-making in emergency management system deployment, and facilitates technology transfer between research and practice. The proposed metrics collectively provide comprehensive system evaluation while maintaining sufficient flexibility for adaptation to diverse emergency response contexts.

4.7. Global Implementation Experiences and Regional Adaptations

Emergency UAV system deployment exhibits significant regional variations reflecting diverse geographical, regulatory, and socioeconomic contexts across different continents.

Nordic medical emergency applications demonstrate advanced healthcare system integration. Claesson et al. [72] analyzed 3165 out-of-hospital cardiac arrest cases in Stockholm County, revealing that drones arrived before emergency medical services in 32% of urban cases (1.5 min savings) versus 93% of rural cases (19 min savings). Their GIS-based optimization and test-flight validation provide empirical benchmarks for AED delivery system design.

Asia–Pacific regulatory challenges are exemplified by Australia’s pharmaceutical delivery barriers. Hogan et al. [73] identified aviation regulations, rather than pharmaceutical regulations, as primary deployment constraints in rural medication delivery, highlighting institutional barriers that limit healthcare accessibility improvements despite technological capabilities.

African development context applications provide resource-constrained environment insights. Wang’s [74] field study in Malawi examined medical supply delivery where healthcare infrastructure is underdeveloped but airspace regulations are relaxed, raising important ethical questions about technology deployment standards in development contexts.

North American clinical integration demonstrates advanced emergency cardiac care applications. Zègre-Hemsey et al. [75] analyzed U.S. drone-delivered AED systems, noting that despite technological feasibility, accessibility challenges persist, with 40% non-utilization rates even when AEDs are within 100 m proximity.

South American operational integration experiences are illustrated through Brazil’s military drone implementation. Janot and de Araujo de Assis [76] analyzed Brazil’s Air Force drone assimilation from 2008 onwards, revealing how local forces adapt technology while building institutional capabilities and security practices.

Cross-regional patterns reveal that rural deployments consistently outperform urban applications globally, regulatory harmonization represents a universal challenge, and successful implementation requires adaptation to local contexts while maintaining core safety standards. These diverse regional experiences demonstrate both universal deployment principles and the necessity of context-specific adaptation strategies for emergency UAV systems.

5. Conclusions and Discussion

This paper comprehensively reviews the research status, challenges, and development trends of key technologies for UAV delivery and ground transfer scheduling in emergency scenarios. Through systematic analysis of eight core technological domains and in-depth study of important studies in the literature, the following significant conclusions are drawn:

(1)

The Technology System Has Initially Formed but Still Requires Improvement

Research demonstrates that UAV delivery technology in emergency scenarios has formed a relatively complete technological system, covering all levels from single-aircraft optimization to system integration. Theoretical research is relatively mature with active algorithmic innovation, particularly achieving breakthrough progress in key technologies such as path planning, task allocation, and collaborative control. Hybrid optimization algorithms, multi-UAV coordination, artificial intelligence enhancement, and real-time adaptation have become mainstream trends in current technological development. However, the completeness and systematicity of the technology system still need improvement, particularly with regard to obvious shortcomings in cross-domain collaboration, system integration, and standardization, requiring further research and development.

(2)

Application Validation Has Made Positive Progress but Large-Scale Deployment Faces Challenges

Existing research has made significant progress in technological feasibility validation, with multiple technologies successfully validated in practical scenarios. Medical emergency delivery achieved rapid transportation of blood products; UAVs demonstrated unique terrain adaptability in disaster relief; contactless delivery played important roles during pandemic control. These successful cases provide strong support for technology promotion. However, it must be recognized that the transition from small-scale experiments to large-scale deployment still faces numerous challenges, including airspace management, security assurance, cost control, personnel training, and other aspects, requiring systematic solutions.

(3)

Critical Challenges Urgently Need Breakthroughs

Research identified five critical challenges: regulatory framework deficiencies severely constrain the legalization and standardized development of technology; insufficient adaptability to adverse environments limits practical applications in emergency scenarios; cybersecurity threats pose serious challenges to system reliability; complex human–machine interaction affects operational efficiency and safety; lack of cost–benefit assessment hinders investment decisions and commercialization processes. These challenges are interconnected and mutually influential, requiring coordinated solutions from multiple dimensions, including technological innovation, management optimization, and policy support—breakthroughs in any single aspect cannot achieve comprehensive improvement.

(4)

Future Development Directions Are Clear with Broad Prospects

Based on current technological foundations and development trends, future research should focus on five directions: distributed decision architectures will enhance system robustness and scalability; robustness enhancement technologies will ensure reliable operation under extreme conditions; cross-domain collaboration mechanisms will achieve optimal resource allocation and efficient utilization; emerging technology integration will bring revolutionary capability improvements; industrialization validation will promote practical applications and commercial development of technologies. These directions complement each other, jointly constituting a complete picture of future technological development.

The theoretical contributions of this research are primarily manifested in the following: first, the construction of a complete technological framework for UAV delivery and ground transfer scheduling in emergency scenarios, clarifying logical relationships among various technological modules; second, the identification of key bottlenecks constraining technological development, pointing out directions for subsequent research; third, the proposal of development pathways based on emerging technology integration, providing new insights for technological innovation.

Regarding practical significance, this research provides a scientific basis for government departments to formulate relevant policies, provides technological assessment frameworks for enterprise investment decisions, provides a reference for research institutions to determine research priorities, and provides technical guidelines for emergency management departments to deploy UAV systems. Particularly in the current context of frequent global disasters and increasing public safety incidents, this research holds important practical significance for enhancing emergency response capabilities and safeguarding people’s lives and property.

While this review focuses primarily on civilian emergency response applications, the technological foundations of UAV–ground vehicle collaborative systems demonstrate significant potential for dual-use applications in defense logistics and military operations. The adaptable nature of these technologies enables seamless transition between humanitarian and defense contexts, offering strategic advantages in both domains.

Military reconnaissance and logistics support represents a natural extension of emergency response capabilities. The research by Nowakowski et al. [77] demonstrates how commercial UAVs can effectively support autonomous ground vehicle missions through advance reconnaissance in unknown environments. Their work with the TAERO manned–unmanned vehicle system highlights critical applications, including visual target tracking, real-time data transmission between robotic platforms, and route verification for autonomous operations. These capabilities directly translate to defense logistics scenarios where supply convoys require advance reconnaissance to identify potential threats, optimal routes, and suitable delivery locations in contested or unfamiliar terrain.

Advanced coordination algorithms developed for emergency scenarios show particular promise for military task allocation. Ma et al. [78] propose adaptive depth graph neural network (AD-GNN) approaches combined with biomimetic algorithms for dynamic task allocation among UAVs and ground vehicles in complex urban environments. Their research addresses reconnaissance, combat, relay, and electronic warfare tasks, achieving operational efficiencies above 85% in search and rescue operations and 90–95% in disaster management scenarios. The adaptive nature of these algorithms enables real-time optimization based on changing tactical situations, threat levels, and mission priorities—capabilities that are equally valuable in military logistics operations where supply chain requirements constantly evolve based on operational demands.

Cross-domain technology transfer between civilian emergency response and defense applications offers mutual benefits. Technologies developed for disaster relief operations, such as robust communication protocols, autonomous navigation in GPS-denied environments, and collaborative decision-making frameworks, directly enhance military logistics capabilities. Conversely, defense-oriented developments in secure communications, threat detection, and operational planning can strengthen civilian emergency response systems. This bidirectional technology flow accelerates innovation in both domains while ensuring that investments in emergency response infrastructure provide additional strategic value for national security applications.

The dual-use nature of UAV–ground vehicle collaborative technologies suggests that future development initiatives should consider both civilian emergency response and defense logistics requirements to maximize technological impact and resource utilization efficiency.

With continuous technological development and growing application demands, UAV delivery and ground transfer scheduling technologies in emergency scenarios will inevitably play increasingly important roles in ensuring public safety and enhancing emergency response capabilities. Future research should pay more attention to the following:

(1)

Strengthening interdisciplinary integration, integrating knowledge from multiple disciplines including transportation engineering, computer science, management science, and social sciences into research.

(2)

Promoting deep industry–academia–research cooperation to accelerate technology transfer and application.

(3)

Conducting large-scale empirical research to validate technology effectiveness in real emergency scenarios.

(4)

Establishing international cooperation mechanisms to promote collective enhancement of global emergency response capabilities.

In conclusion, UAV delivery and ground transfer scheduling technology in emergency scenarios represents a research field full of opportunities and challenges. Through continuous technological innovation, systematic engineering practice, and comprehensive policy support, this technology will inevitably make important contributions to constructing safer, more efficient, and intelligent emergency response systems, providing strong support for the sustainable development of human society.