Implementation and Critical Factors of Unmanned Aerial Vehicle (UAV) in Warehouse Management: A Systematic Literature Review

Abstract

Unmanned aerial vehicles (UAVs) have proven to be a key solution for nearly automated or smart warehouse operations, enabling receiving, picking, storage, and shipping processes to be timely and more efficient. However, there is a relative scarcity of review studies specifically on UAV-based warehouse management. Research knowledge and insights on UAV applications in this field are also limited and could not sufficiently or practically support decision-making on commercial utilization. To leverage the potential applications and current situation of UAVs, this study provides a systematic literature review (SLR) on UAV adoption in warehouse management. SLR approach was critically conducted to identify, select, assess, and summarize findings, mainly on the two descriptive research questions; what are the past applications of UAV, and what are critical factors affecting UAV adoption in warehouse management? Five key critical factors and 13 sub-factors could be observed. The results revealed that hardware (e.g., payloads, battery power, and sensors) and software factors (e.g., scheduling, path planning, localization, and navigation algorithms) are the most influential factors impacting drone adoption in warehouse management. The managerial implications of our research findings that guide decision-makers or practitioners to effectively employ UAV-based warehouse management in good practice are also discussed.

1. Introduction

Nowadays, unmanned aerial vehicles (UAVs), commonly known as drones, have received great attention and commercial utilization in several domains. The significant advantages of UAVs to industries are that they are commercially affordable, easily controllable and deployable, precisely and accurately manoeuvrable, and efficient surveillants. From a broad perspective, UAV and some of its characteristics specifically benefit the focused tasks or activities. For example, the utilization of UAVs for detecting coastal landslides could rapidly map large coastal areas and simultaneously create high-resolution images [1]. In the agriculture domain, drones could provide aerial spraying ability, especially the positive impact of airflow from its rotor, which could bring notable achievement in a field spraying task [2]. Similar to other domains, UAVs are identified as a key innovative solution that plays a vital role in the warehouse management field since it is estimated that the improvement of inventory counting could be faster 119 times than the traditional approach [3]. Regarding the distinctive benefits of UAVs, they have been widely suggested and applied to various inventory and traceability tasks; [4,5] concluded the areas of UAV application in warehouses such as inventory audit, cycle counting, stocktaking, intra-delivery, surveillance, etc.

In recent years, several studies proposed new initiatives for drone adoption and implementation in warehouse management. For instance, the integrations of UAVs, barcodes, computer vision (CV), and machine learning (ML) were proposed for automated warehouse operations such as barcode detection and decoding [6,7]. Moreover, in the meantime, for improving self-positioning UAVs and automatic identification and data capture (AIDC), there are other studies, including [8], integrating drones with light detection and ranging (LIDAR) devices, together with radio-frequency identification (RFID).

As mentioned above, the adoption of drones in warehouse management tasks has been keenly interesting by academics and practitioners. Although there appear to be a large number of studies related to UAV applications in warehouses over recent years, existing research studies still focused on different domains (e.g., construction, humanitarian, and agriculture) or emphasized broader scopes than that of the warehouse management domain. To the best of our knowledge, there is a lack of a review study that concentrates on the scope of UAV-based warehouse management, commonly interconnected with supply chain management (SCM) and logistics management. Rather than that, the intensive analysis of the key influential factors of UAV adoption and in-depth information which could potentially support the authority’s decision is also limited.

From the literature, most existing review research on UAV adoption focuses more on a wider or narrow area. In a wider area, UAV applications are studied and implemented in many topics, for instance, autonomous cars and traffic control in the transportation sector, warehouse robots, search and rescue, and video games. This also means that research scholars tend to discuss UAVs’ adoption and their applications from a global viewpoint. Thus, findings or suggestions were provided broadly and might not be beneficial across other and different domains. In a narrow area, however, the analysis or discussion of the most recent research is often presented in the warehouse management field and with more specific tasks, such as stocktaking, regular surveillance, or product delivery. UAV adoption is implemented locally and does not take into account the whole warehouse management procedure. Those studies might lack the consideration of interconnections among different but related elements affecting the whole process of the focused domain.

For those reasons, we have examined previous UAV applications and capabilities to better grasp the most updated circumstances and existing gaps and establish a research agenda for UAV-based warehouse management activities. The value-added contributions of this study can be listed as follows:

This review research encounters the latest UAV adoption in a specific SCM domain by considering the entire warehouse management processes, e.g., inventory management, intra-logistics of items, and inspection and surveillance.

• The critical factors and their related sub-factors for UAV adoption and implementation are identified and analyzed to perceive an element that significantly affects UAV applications in warehouse management;
• The research contributes to delivering shareholders’ value and suggestions on devising a measure of UAV adoption by focusing on different warehouse operations together with diverse critical factors.

The purpose of this study is to gain insights into the aforementioned points, which can be summarized into two major research questions (RQs) as follows:

RQ1: What are the past applications of UAVs in warehouse management tasks?

RQ2: What are the critical factors of UAVs affecting the adoption and implementation in warehouse management?

To yield novel knowledge on the above RQs, a systematic literature review (SLR) approach was adopted in this study. The SLR method could assist scholars in performing the review approach accurately and reproducibly while reducing risks and errors caused by scholars’ bias. The method can improve the reliability and acceptability of work. Following the SLR procedure, the necessary data on UAV-based warehouse management has been accumulated and summarized systematically. It is assumed that the analyzed information would help practitioners easily decide on what factors should be prioritized and take at the first concerns for adopting, implementing, and manipulating UAVs for each warehouse management task.

The remainder of the study’s structure is organized as follows: the next section provides the SLR methodology for reviewing articles. After that, the major subsequent section presents the obtained statistical results from the literature. The descriptive analysis results and findings on both RQs were also reported and discussed in the same section. Finally, the last section shows the conclusions and limitations of this study. For readability purposes, we have also listed all acronyms used in the paper in

Table 1. List of acronyms used in the papers.

Acronym	Definition	Page Number
AGV	automated guided vehicle	16
AIDC	automatic identification and data capture	2
AR	augmented reality	16
CF01	critical factor 01-Hardware	14
CF02	critical factor 02-Software	14
CF03	critical factor 03-Integrated systems and others	14
CF04	critical factor 04-Area or distance of operation	14
CF05	critical factor 05-Mission time	14
CF06	critical factor 06-Costs	14
CF07	critical factor 07-Drone operation	14
CF08	critical factor 08-Warehouse operation	14
CF09	critical factor 09-Environment	14
CF10	critical factor 10-Items and inventories	14
CF11	critical factor 11-Organization	14
CF12	critical factor 12-Legislation and standards	14
CF13	critical factor 13-Society and mental	14
CFs	critical factors	7
CV	computer vision	2
GNSS	global navigation satellite systems	15
GPS	global positioning systems	12
IEEE	Institute of Electrical and Electronics Engineers	9
IMU	inertial measurement unit	21
LEDDAR	light-emitting diode detection and ranging	15
LIDAR	light detection and ranging	2
MDPI	Multidisciplinary Digital Publishing Institute	9
ML	machine learning	2
RFID	radio-frequency identification	2
RQs	research questions	2
SCM	supply chain management	2
SLAM	simultaneous localization and mapping	21
SLR	systematic literature review	2
Sub-CFs	sub-categories of the critical factors	14
UAVs	unmanned aerial vehicles	1
UGR	unmanned ground robot	16
UGV	unmanned ground vehicle	16
UWB	ultra-wideband	21
VTOL	vertical takeoff and landing	13
WoS	web of sciences	4

. The page numbers on which each acronym is first mentioned are also given.

2. Systematic Literature Review (SLR) Methodology

To systematically review articles related to RQs, we followed the SLR approach’s comprehensive guidelines, which were previously applied in several review studies [9,10,11]. The SLR procedure consists of 3 major phases: planning, conducting, and reporting, which can be further divided into 7 minor steps, as depicted in Figure 1.

2.1. SLR Necessity

In recent years, many studies have increasingly paid attention to the utilization of UAVs in warehouse management tasks. Nevertheless, the applied warehouse activities, critical success/failure factors, and potential observations were still varied and mainly scattered. With analysis and summarization of those past results and findings, scholars and practitioners can often realize current situations, significant development possibilities, and improvement opportunities. Therefore, to address the abovementioned issue, this study aims to collect evidence from past studies and further analyze it in academic and practical contexts regarding the RQs. The SLR approach was used in this study to respond to the well-defined RQs systematically and dependably. Based on the guidelines in [11], we first validated that there is no similar work in past literature. We have used the following keywords and syntaxes: (“warehouse” OR “inventor*” OR “stock”) AND (“unmanned aerial vehicle” OR “UAV” OR “drone”) AND (“systematic review” OR “literature review” OR “systematic literature review” OR “SLR”). Regarding the mentioned query string, results from the 2 high-quality academic databases, SCOPUS and Web of Science (WoS), presented that there is no literature review and SLR study considering the same purposes and scopes of our study. Therefore, based on the significance of research objectives, questions, and underexplored topics, it is necessary to answer the RQs of this study.

2.2. Research Questions (RQs)

As presented in the introduction section, this study identifies 2 major RQs. To better show the study’s goals, the motivations of RQs are explained in

Table 2. RQs and Motivations.

Research Question	Motivation
RQ1: What are the past applications of UAVs in warehouse management tasks?	To understand the state of the art in using UAVs for warehouse operations.
RQ2: What are the critical factors of UAVs affecting the adoption and implementation in warehouse management?	To identify critical factors for the success and failure implementation of UAVs in warehouse management

.

2.3. Review Protocol

This section shows a set of core processes for searching and selecting studies to extract and analyze data. The first process is the search process which focuses on identifying effective search strings and keywords. It is acknowledged that the lack of consistency in concepts and terminology could be a significant problem in searching and finding relevant articles. Therefore, in this step, we reviewed the concepts and terminology of UAVs and warehouse management. After that, we identified major keywords regarding the defined RQs concurrently with their synonyms, acronyms, related terms, and descriptions, as presented in

Table 3. The search terms.

Major Keywords	Synonym, Acronym, and Related Term	Description
unmanned aerial vehicle	UAV, drone	The direct term (“unmanned aerial vehicle”), acronym (“UAV”), and alternative (“drone”) are used to cover articles related to UAVs.
warehouse	inventor*, stock	The direct keyword “warehouse” is used to detect warehouse management activities from searched articles. Moreover, other alternative terms related to the warehouse are also applied to search all relevant articles as much as possible. The “inventor*” is used to cover inventor, inventory, and inventories, and “stock” is adopted to cover both stock and stocks.

.

Together with the search terms, we also apply Boolean operators “AND” and “OR” to make an inclusive search. The “AND” operator is applied between major keywords, while the “OR” operator connects the major keywords with their synonyms, acronyms, and related terms. Therefore, the final search string applied in this study is presented as follows: (“unmanned aerial vehicle” OR “UAV” OR “drone”) AND (“warehouse” OR “inventor*” OR “stock”). The selection of studies is executed in 3 major subsequent processes, which are the primary search, secondary search and, finally, snowball tracking. The first step is executed by searching articles indexed in 2 well-known research databases, SCOPUS and Web of Science. The searched articles were published from 2006 (the first year that the FAA (Federal Aviation Administration) legally permitted a commercial drone) [12] to the end of December 2022. All obtained articles from the 2 renowned sources are combined, after which the duplicates are removed from the first lists.

The secondary search is performed by evaluating and screening studies obtained from the primary search. This step filters the irrelevant studies by applying the exclusion and inclusion criteria described as follows. The exclusion criteria exclude the studies unavailable for downloading a full-text paper, and studies classified as short or editorial articles were filtered out. All studies that are not written in the English language were also removed. Finally, we eliminated studies that are not classified in the research domains of “engineering,” “computer science,” “social sciences,” “decision sciences,” “business, management, and accounting,” “economics, econometrics, and finance,” and “multidisciplinary.”

Apart from the exclusion criteria, the inclusion criteria focus on the relevance of studies to the RQs and their quality. To select high-quality papers for delivering reliable data and results [9], firstly, we select only peer-reviewed documents which are classified as journal articles and conference papers. Secondly, we then evaluate the relevance of filtered studies to RQs by examining their titles, abstracts, introductions, and conclusions.

In the last step, we apply a snowball technique to maximize the inclusion of relevant studies. This method could significantly support search paradigms and the identification of additional references. Identifying new articles is based on the studies cited by the article being examined. The method to review and extract papers’ data and information from this snowball process is similar to the approach applied to studies screened by the primary and secondary search. Detailed information on the data extraction process is described below.

To systematically extract data from all papers filtered and selected by the 3 steps mentioned above, we use a form created by Microsoft Excel to record data uniformly. The data shown in

Table 4. Data extraction from each paper.

Type of Data	Collected Data
General data	Article title, author name, publication source, year of publication
Specific data	RQ1: type of UAV, warehouse’s task or activity utilized UAV, capability, and advantage of UAV RQ2: critical success/enabling and failure/challenge factor

are used to analyze further results and findings.

Regarding various characteristics and influential critical factors of UAVs and warehouses, we have followed definitions and characteristics of the focused aspects identified in past well-known studies to extract and classify data consistently. For the categories of UAVs, we classify drones based on their weight following the categorization proposed by [13], as listed in

Table 5. UAV categorization.

Designation	Weight Range
Nano drone	W ≤ 200 g
Micro drone	200 g < W ≤ 2 kg
Mini drone	2 kg < W ≤ 20 kg
Small drone	20 kg < W ≤ 150 kg
Tactical drone	150 kg < W ≤ 600 kg
Strike drone	W > 600 kg

.

For the warehouse management tasks, we have relied on the categories of warehouse operations defined by [14]. The operations can be classified into 3 primary groups: inventory management, intra-logistics of items, and inspection and surveillance. The descriptions of warehouse tasks are detailed in

Table 6. Warehouse management tasks and descriptions.

Warehouse Operation	Description
Inventory management
Inventory audit	An inventory audit is a process of cross-checking actual physical inventory levels/records against financial records to ensure accurate inventory accounting. Auditors and third parties could complete this process to identify any problem related to the counting stock process, inventory storage, and accounting.
Stock management	Inventory management is a systematic approach to sourcing, storing, and selling inventory—both raw materials (components) and finished goods (products).
Cycle counting item search	Cycle counting is an inventory control method allowing businesses to confirm that physical inventory counts match their inventory records. This method involves performing a regular count of physical items in different areas of the warehouse without counting the entire inventory and then recording the adjustment of specific products.
Buffer stock maintenance	Buffer stock refers to extra inventory kept on hand in case of manufacturing delays or an unexpected increase in demand. Thus, buffer stock maintenance is the process of calculating and maintaining the right amount of buffer stock to have available to help keep carrying costs low while ensuring customer orders are fulfilled on time.
Stocktaking	Stocktaking (or stock counting) is about manually checking and recording all the inventory businesses currently have. To conduct stocktaking, the entire business might need to shut down for a time to allow for each item to be physically counted. This process also includes checking the stock types and verifying the conditions of stock movement and their whereabouts.
Intra-logistics of items
Drone-based delivery	Intra-logistics of the item is defined as the logistic processes within a warehouse or factory using a drone, including warehousing, material or information flow, transportation, and express delivery of items, tools, and spare parts within the warehouses. It also referred to picking and commissioning final products for customers, e.g., drone-based delivery service.
Inspection and surveillance
Monitoring and inspection	Warehouse monitoring and inspection is the pre-planned safety process or proactive approach for identifying potential risks to the safety, integrity, and quality of stored commodities in a warehouse. The process aims at ensuring employee safety, inventory security, and optimized workflows and procedures.
Regular surveillance	Regular surveillance is an alternative inspection process that pinpoints preventing/prohibiting the primary threats to warehouses, i.e., theft and other unwanted behaviours. This process is an objective evaluation to determine how well the quality procedures are followed in day-to-day production and is often integrated with the surveillance system.

.

Besides, the critical factors (CFs) can be categorized into 2 major types: success and failure factors. However, 1 CF can be classified as either success or failure depending on users’ situations or characteristics. Therefore, in this study, we extract factors that significantly impact the success or failure of UAV adoption and implementation in warehouse operations as CFs.

Finally, the data and information extracted from selected papers are further performed for quantitative and qualitative analysis. Key analysis results have been reported and discussed based on the set-out RQs. The review protocol of this study was drawn in Figure 2 to summarize and illustrate a structured methodology for capturing knowledge or insights on this topic and all relevant variables. The results of SLR regarding our proposed review protocol are presented in the next section.

3. SLR Results and Discussion

After utilizing the SLR methodology presented in Figure 1, articles obtained and screened from the pre-defined procedure can be concluded and presented in Figure 3.

The primary search was carried out at the end of December 2022. From this search, 1561 potential studies containing 908 and 653 articles were collected from the SCOPUS and WoS databases, respectively. However, the studies from these two renowned databases mostly came from the same sources. Therefore, regarding the redundancy, we removed 427 duplicates from the first search; this reduced the potential articles to 1134. Based on the exclusion and inclusion criteria identified in the previous section, the potential papers were further filtered to 804 and 98 articles, respectively. Finally, the snowball technique was conducted to maximize the acquisition of relevant studies, and seven articles were added to our SLR. The amount of final reviewed papers is 106. Compared to other SLR studies (e.g., [15,16]) analyzing CFs and published in high-quality journals, the number of reviewed articles seems adequate and acceptable.

3.1. Descriptive Results

3.1.1. The Distribution of Publications

The distribution of all reviewed articles corresponding to the time dimension can be depicted in Figure 4. The chart clearly shows a recent high interest in UAV adoption in warehouse activities. More than half of the articles (59.43%) were published in the last three years (between 2020 and 2022). Therefore, this finding confirms the suitability of reviewing, analyzing, and summarizing this topic. The study’s findings would indicate past and current situations and opportunities for improvement in this research area.

3.1.2. Publication Sources

Out of 106 articles illustrated in Figure 5, 49 studies (46.23%) were published in academic journals, whereas 52 papers (49.06%) and five papers (4.72%) remained that were conference proceedings and book series, respectively. Generally, a number of studies were published in different journals and conference proceedings, so most publication sources (65.09%) could publish only one paper on this recent topic.

Table 7. Number of studies published by publication sources.

Source	Type of Publication	No. of Studies
Sensors	Journal	5
European Journal of Operational Research	Journal	3
Applied Sciences	Journal	3
International Conference on Unmanned Aircraft Systems (2021)	Proceeding	2
ACM International Conference Proceeding Series	Proceeding	2
AIAA Scitech 2019 Forum	Proceeding	2
IEEE Access	Journal	2
IEEE International Conference on Emerging Technologies and Factory Automation	Proceeding	2
IEEE International Conference on Intelligent Robots and Systems	Proceeding	2
IEEE Robotics and Automation Letters	Journal	2
IEEE Transactions on Automation Science and Engineering	Journal	2
IEEE Transactions on Instrumentation and Measurement	Journal	2
IEEE Transactions on Intelligent Transportation Systems	Journal	2
IEEE Transactions on Systems, Man, and Cybernetics: Systems	Journal	2
Proceedings of the American Control Conference	Proceeding	2
Others		69

shows that only 15 academic sources could publish more than two papers per journal or conference proceeding. Most of these sources were affiliated with the Multidisciplinary Digital Publishing Institute (MDPI) or the Institute of Electrical and Electronics Engineers (IEEE).

3.1.3. Authorships and Collaborations

There were 382 different authors studying the research related to the adoption of UAVs in warehouse management. As shown in

Table 8. Number of articles per author.

No. of Articles per Author	No. of Authors (Name of Authors)	Percentage
1	346	90.58%
2	29	7.59%
3	2 (Pieter Suanet, and Emmeric Tanghe)	0.52%
4	3 (Dzmitry Tsetserukou, Ivan Kalinov, and Wout Joseph)	0.79%
5	0	0.00%
6	1 (Jeroen Hoebeke)	0.26%
7	1 (Eli De Poorter)	0.26%
Total	382	100%

, most authors (90.58%) contributed one paper on this topic. Only 9.42% (36 of 382 authors) of authors could publish more than one article, and most of them (29 authors or 7.59% of all authors) published two articles. Interestingly, two authors intensively contributed to this complex topic. Eli De Poorter and Jeroen Hoebeke published seven and six articles in this research field, respectively. Both of them worked in the same research group (IDLab Group) affiliated with the Department of Information Technology at Ghent University. Five other authors could publish more than two articles. Three of them (Pieter Suanet, Wout Joseph, and Emmeric Tanghe) also worked in the IDLab Group. In contrast, the two remaining authors (Dzmitry Tsetserukou and Ivan Kalinov) were with the Skolkovo Institute of Science and Technology. Therefore, regarding the small proportion of highly published authors found from the analysis, this finding signifies the limitation of continuity in this research topic.

To understand the collaboration of authors in this complex and integrated study, we have analyzed the number of authors per article shown in Figure 6. From the information, we realize that almost all articles (94.34%) require the collaboration of authors, and most studies (28.30%) were conducted by three authors and followed by the collaboration of four authors (20.75%). Similar to our findings in past work [17], the collaboration of authors is the key characteristic of UAV adoption integrated with other research methods or technologies. Therefore, this research field is fundamentally multidisciplinary and highly requires exchanges of knowledge among various researchers from different areas.

As presented above, findings obtained from the descriptive analysis could show the trends and general data of articles. However, it is unable to reveal in-depth details, specific insight or information on UAV adoption in the warehouse management task. To fulfil the ultimate objectives of this study, we intensively reviewed and analyzed all qualified articles in accordance with the RQs previously specified. The significant findings obtained from our analysis based on RQs are discussed in the subsequent sections.

3.2. RQ1: “What Are the Past Applications of UAVs in Warehouse Management Tasks?”

The applications of UAVs in warehouse management are quite varied. However, our analysis classified warehouse management tasks into three major operations, which could be further classified into eight applications, as mentioned in the previous section. From 106 studies, we have summarized significant information regarding the application of UAVs in warehouse management activities, as presented in

Table 9. Application of UAV in warehouse operations.

Warehouse Operation	Application Area	No. of Articles 1	Percentage
Inventory management (54 articles or 47.37%)	Inventory audit	1	0.88%
	Stock management	32	28.07%
	Cycle counting item search	11	9.65%
	Buffer stock maintenance	1	0.88%
	Stocktaking	9	7..89%
Intra-logistics of items (47 articles or 41.23%)	Drone-based delivery	47	41.23%
Inspection and surveillance (13 articles or 11.40%)	Monitoring and inspection	5	4.39%
	Regular surveillance	8	7.02%

¹ Some articles applied UAV to more than one application area.

.

In accordance with RQ1, Table 9 presents the UAV applications in the eight significant warehouse operations. Findings show that inventory management is the most studied task (54 articles or 47.37%), which conforms to the past findings (e.g., [5] emphasizing UAVs’ highest potential for warehouse application. The second most studied major application is the intra-logistics of items (47 articles or 41.23%). This logistics-related task operated by UAVs has been widely studied regarding its challenges and improvement opportunities, especially UAVs’ payload and battery capabilities. The least investigated warehouse operation by drone was the inspection and surveillance. The topic was unattractive to most researchers since this operation is typically recognized as an uncritical task of warehouse management. Moreover, UAV surveillance technology is mature and successfully implemented in several real industries (e.g., [18,19,20]).

Although the above findings could present the past situations of UAV applications in major warehouse operations, these results still are unable to provide in-depth details of warehouse sub-tasks. Therefore, we deeply analyzed the usage of UAVs in sub-areas of application as depicted in the second column of Table 9. For the sub-applications of UAVs in warehouse tasks, drone-based delivery was the most studied area (45 articles or 40.91%). In this major operation, most studies focused on optimizing scheduling and path planning (e.g., [21,22]), as well as improving the localization and navigation of UAVs in warehouses (e.g., [23,24]). The scheduling and path planning tasks aimed to optimize the number of utilized drones, whereas the studies related to the localization and navigation of UAVs focused on the solutions of indoor positioning without the usage of a global positioning system (GPS). The heavy attention on these subjects highlights the drawbacks and potential for the advancement of UAV use in warehouse operations. The second most explored sub-area of drone application was stock management (32 articles or 28.07%). The studies in this sub-category applied UAVs for managing warehouses and stocks. Several papers consider and improve warehouse-related systems with the adapted UAVs. For example, the study [25] focused on the improvement of UAV design for stock management activity, and another research [4] improved the inventory data transfer of UAVs in real-time conditions. High attention to these warehouse operations indicates their challenges and high improvement demands.

Apart from the enriched studied categories above, two sub-operations of warehouse management adopting UAVs were highly underexplored. Those are inventory audits and buffer stock maintenance (one study per sub-operation or 0.88% of articles). For the inventory audit, the paper of Barlow and colleagues (2019) [26] proposed a novel algorithm to minimize the mission time of inventory audit as well as warehouse activity. In accordance with buffer stock maintenance, another study [27] applied deep neural networks to estimate the volume of wood stock. From the evidence, these two sub-areas of UAV applications were studied by each paper only. Therefore, this finding highlights the opportunities for improvement on these underexplored topics.

From the perspective of UAV-related information, the extracted results from the reviewed articles manifest that most studies (89 papers) did not present the sizes and weights of UAVs. Consequently, types of UAVs were mostly unidentified in these articles. A few remaining studies classified drones into distinct categories in accordance with the number of rotors, and all these drones can be categorized as multi-rotor UAVs. Among these papers, most applied UAVs are quadcopter or four-rotor drones (19 papers), hexacopter (two articles), and multi-rotor (one article), respectively. It is obvious that these UAVs are identically classified as multi-rotor drones or rotary-blade drones. Regarding the designs of UAVs, none of these studies applied other drone types: fixed-wing UAVs and fixed-wing hybrid vertical takeoff and landing (VTOL). The wide adoption of multi-rotor UAVs in many studies comes from their distinctive advantages as stated in [28,29,30]), including (1) ease of control and manoeuvre, (2) vertical takeoff and landing abilities, (3) stability, and (4) reasonable price.

There are 18 articles that prominently identified the type, specification, or model of the UAVs. Among this group of articles, a small drone was the most applied UAV (seven articles or 38.89% of papers specifying UAVs). In comparison, micro drones were the second most adopted UAV (six articles or 33.33% of papers specifying UAVs). One can see that most drones (94.44%) applied for warehouse management were lighter than tactical drones. The high utilization of UAVs that are lighter than 150 Kilograms largely depends on the fundamental characteristics of warehouse management related to indoor operations and confined spaces [31]. Regarding these characteristics, the heavier drones (tactical drone and strike drone) are not suited for most warehouse management activities, except intra-company delivery between long-distance outdoor warehouses (e.g., [32]). Although the low-weight UAVs are mostly and suitably applied for warehouse operations, the lowest-weight UAV or nano drone is different since it has some significant limitations, especially the extremely limited payload and flight range capability. For these reasons, none of the reviewed studies applied nano drones to operate warehouse tasks. Therefore, from our findings and analysis, we recommend the application of low-weight UAVs and multi-rotor UAVs for warehouse management operations. Nevertheless, it could be varied by irregular situations (e.g., outdoor operations, long-range delivery, and high-weight inventory).

The comprehensive literature review results from this section show the past and current status of UAV applications in warehouse management tasks. Furthermore, these novel findings could highlight the gaps in the existing body of knowledge that can be applied for studying in future works. To better understand the success and failure of UAV applications in warehouse management, in the next section, we examine the factors affecting the success and failure of UAV adoption in warehouse management.

3.3. RQ2: “What Are the Critical Factors of UAV Affecting the Adoption and Implementation in Warehouse Management?”

From the extensive literature review, we identify the CFs in adopting and implementing UAVs in warehouse management operations. This research operation is conducted following the second research question (RQ2). Regarding the numerous factors found in the reviewed articles, we filtered, selected, and summarized factors indicated as a major variable of experiments. Thus, a set of discovered factors were identified for significant impacts on UAV applications before being classified as CFs. To reduce the number of CFs to a manageable level, analogous CFs were combined. For CFs that are too narrow and specific, they were amalgamated into the more generic CFs. From these approaches, we have concluded CFs and then classified them into major CFs, as previously indicated in the SLR methodology section. The sub-categories of CFs were classified by their characteristics and corresponding with the major CFs. Finally, the summarized CFs are concluded and presented in

Table 10. List of the critical factors in implementing UAVs for warehouse management.

Major CFs	CFs	Acronym	References 1
Technology	Hardware	CF01	[4,7,14,18,21,24,25,28,29,30,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84,85]
	Software	CF02	[3,4,6,7,14,18,21,22,23,24,26,27,28,29,30,32,33,34,36,37,38,39,43,44,45,47,49,51,55,58,59,61,62,63,65,66,67,68,70,72,76,77,79,80,82,83,84,85,86,87,88,89,90,91,92,93,94,95,96,97,98,99,100,101,102,103,104]
	Integrated systems and others	CF03	[29,33,36,61,62,69,81,97,105,106,107]
Operation	Area or distance of operation	CF04	[18,29,32,33,36,37,38,39,40,41,43,46,57,58,59,60,65,67,70,73,75,76,77,78,79,82,83,84,85,91,96,105,108,109]
	Mission time	CF05	[4,26,28,29,32,36,39,41,45,46,53,54,69,70,74,77,84,85,87,96,105,108,110,111]
	Costs	CF06	[7,25,29,32,34,37,41,42,48,49,54,55,56,58,67,76,78,79,85,96,110,111,112,113]
	Drone operation	CF07	[7,18,32,36,37,39,40,49,50,54,63,66,67,69,70,71,73,76,77,84,85,106,110,111,112,114]
	Warehouse	CF08	[4,14,28,32,33,35,41,47,51,54,57,60,65,69,71,81,84,104,108,115]
	Environment	CF09	[14,18,24,32,38,40,41,44,46,55,62,74,79,82,87,91,96,116,117]
	Items and inventories	CF10	[27,28,29,33,46,58,70,74,77,78,79,91,111,117]
Organization	Organization	CF11	[25,36,77,82,114]
Legislation and standards	Legislation and standards	CF12	[14,55,56,79,96]
Society and mental	Society and mental	CF13	[14,50]

¹ Some articles highlighted more than one CF or sub-category of the critical factors (Sub-CF); For example, the article [83] denoted hardware as CFs and emphasized more than one Sub-CF of the hardware, including the weight of the UAV and energy consumption.

.

As shown in Table 10, there are 13 identical CFs among five major CFs of UAV successful adoption in warehouse management. The explanations and discussions on CFs and sub-categories of the critical factors (Sub-CFs) are presented as follows.

3.3.1. Technology

The technological factor is commonly recognized as an aerial platform which is identified as a major component of UAV [118]. This major factor can be divided into three CFs: hardware (CF01), software (CF02), integrated systems and others (CF03), as stated in Table 10.

• Hardware

Hardware (CF01) is denoted as the most examined element. There are three types of hardware related to UAV operations in warehouses, including (1) drone hardware, (2) warehouse hardware, and (3) network hardware. Note that the operations of UAVs in warehouses are drone-based delivery that is often employed with a long-range or long-time task in a large indoor warehouse environment. Several studies in recent years emphasized the improvement and development related to CF01, including indoor navigation sensors, battery size, energy consumption (power system), as well as payload capability of UAVs. In the indoor warehouse, UAVs are limited to utilizing the fundamental navigation systems [72], i.e., global positioning systems (GPS) and global navigation satellite systems (GNSS), as opposed to outdoor operations. For these reasons, several studies highlighted the sensor as a critical factor for navigating indoor warehouses (e.g., [37,44,48]). Moreover, LIDAR, also known as a 3D laser scanner, is one of the most recommended sensors. Concerning the current limitation of autonomous indoor navigation of UAVs, various on-board sensors, such as light-emitting diode detection and ranging (LEDDAR), ultrasonic distance meter, and gyroscope, are recommended as the built-in functions of commercial drones [81].

Payload capability is another critical technological factor of UAV-based warehouse operations. Most studies consider it as a variable of studies’ designs or experiments (e.g., [32,59,84]). UAVs’ payload directly affects warehouse management missions’ capability [60]. If UAVs can carry higher inventory weights, it could reduce the time or number of flight rounds and increase warehouse operations’ efficacy. Hence, the payload could directly benefit inventory delivery and stocktaking tasks [65]. The payload capacity of UAVs refers to the ability to cover wide ranges of inventory weights allowing business units to meet various industry requirements. The higher and more flexible payload capability would enhance the possibility of UAV utilization in commercial or industrial sectors.

It is worth reminding that payload was widely identified as the trade-off issue with another highly examined factor, i.e., the power system of drones [5]. Specifically, battery and energy consumption are affected by UAVs’ payload, drone velocity, and some other factors, i.e., area or distance of operation (flight range) [79]. It is an undeniable fact that the potential of UAV applications directly depends on these three key attributes. For this reason, the power system of UAVs is then identified as CFs in several studies. A lot of concerns related to the power system of UAVs in warehouse management have been raised by many research studies. Most studies emphasized the battery’s size and weight, which are significant for the UAV’s capability, allowing long-flight range, higher speed, and more loading capacity. Whereas some studies highlighted the significance of power systems and concerns in other dimensions, such as ease of battery replacement [25], charging time and charging facilities [53], as well as plug-in charging capability [49]. There are other aspects that are rarely mentioned or included in CF01 in a few previous works, for instance, the design of the UAV [50], the remote controller [32], the material of the UAV’s body or structure [61], and an improved GPS system [35].

Basically, drones must be integrated with warehouse technology or other hardware components. For this reason, CF01 should also be discussed in terms of warehouse hardware. This part is related to equipment, parts, or components of inventory in warehouses. Although UAVs can be applied efficiently and are extremely useful for warehouse operations, only the adoption of unmanned aircraft could not achieve some significant missions of warehouse management, e.g., inventory delivery and stock counting. From the literature review, only two pieces of warehouse hardware could be found: barcode tags and RFID tags. Both of them were utilized for identifying and providing encoded data and information on inventories (e.g., [35,55]). Specifically, RFID could be further used for identifying the positions of stocks [52]. Importantly, it can be used for supporting navigation and position reference of UAVs (e.g., [51,72,81]).

The network is another element included in CF01 and is always related to the success or failure of UAV adoption in warehouses. Several studies identified that network and communication systems play significant roles in smooth and fast data transmission and navigating UAVs. From these tasks, there are three components of the network system highlighted as influenced factors in previous studies, including wi-fi (e.g., [33,68]), antenna (e.g., [35,51]), and network connection (e.g., [64,79]).

• Software

UAVs and warehouses are complex systems that fundamentally compose hardware and software components. Software (CF02) has been identified as the critical technological factor of UAV adoption in warehouses by several articles. The most studied and suggested part of CF02 is software algorithms which can be further classified into UAV algorithms and warehouse management algorithms. The UAV algorithms are highly discussed and mainly rely on two algorithms: commands or instructions for (1) localizing and navigating UAVs in the indoor environment and (2) scheduling and path planning of UAVs.

As mentioned in the hardware section, hardware sensors installed on UAVs are unable to navigate and position objects in indoor warehouses accurately and safely. For this reason, many research scholars have studied, developed, or suggested algorithms for accurately controlling drones to prevent their collisions in the indoor environment (e.g., [30,34,82]). Nevertheless, it could be observed that most of the past developed UAV algorithms were just the initial generations. They still required further improvement for the high precision of navigation and localization. Furthermore, algorithms should be improved simultaneously along with the integrated UAV hardware components [29,55].

The latter UAV algorithm aims to generate and manipulate the flight schedule and hovering paths of drones. Note that the success and failure of UAV applications depend on costs and the efficiency of operations. The development of UAV’s route and schedule optimization algorithms is ultimately necessary [79]. Regarding UAV and warehouse management constraints, the newly invented or improved algorithms were introduced in various dimensions. Based on the time dimension, some studies proposed optimization algorithms considering flight time, delivery time, computational time, or overall mission time, as could be observed in [26,36,70,73,102]. In another dimension, the UAV algorithms were developed by focusing on energy consumption [83,103], path length [59,66], path quality [97], mission costs [67], task distribution [49,72], or flight accuracy [38]. Although UAV path and schedule planning algorithms have been progressively proposed in the past decade, they are still varied and only depend on specific situations or strictly focused scenarios. Enlightened by this limitation, it is required to develop generic path and schedule planning algorithms concerning a variety of variables and applicable within a wide range of constraints. There are also some other software algorithms related to UAV functions that were studied in past research, including stabilization [7], image or video recognition [61], and disturbance prediction [94].

• Integrated systems and other technologies

The integration of hardware, as well as software among different systems or various technologies, was identified as CF03 [29,36,105]. Nevertheless, the underlying studies in this aspect are still far fewer than the abovementioned CF01 and CF02. In this category, mainstream studies strongly pointed out the integration of different systems, technologies, or equipment associated with the mobile automated guided vehicle (AGV) and UAV [61], unmanned ground robot (UGR) and UAV [62], unmanned ground vehicle (UGV), UAV, and RFID [69], UAV and RFID [97], blockchain and drone [33], as well as augmented reality (AR) and UAV. These studies emphasized that the success of UAV adoption and implementation in warehouse operations could derive from the practical application of UAVs. To achieve the industrial requirements and practices, it is necessary to integrate UAV systems with other systems or technologies that can address the UAVs’ flows and solve real-world business problems.

3.3.2. Operation

As discussed in RQ1, UAV operations in warehouse management can be categorized into three tasks: inventory management, intra-logistics of items, and inspection and surveillance. Several operational CFs were examined and suggested in the reviewed studies to achieve these three operations.

• Area or distance

Massive studies considered area or distance (CF04) in which the UAV flight has been operated as one of the core experimental variables. CF04 could be influenced by and affect other UAV operation aspects, such as the drone’s velocity, the number of drones, and delivery time. Compared with traditional operations, UAVs could greatly advantage firms with large warehouse areas, long distances, or high trajectories of logistic operations [3]. On the other hand, UAVs failed to operate effectively in small warehouses with narrow corridors, limited space, or low-altitude shelves. In most industrial situations, drones cannot fly and hover autonomously because warehouses periodically change their layouts, shelf sizes, and levels that directly affect areas of UAVs’ operations [72]. From past findings and suggestions, the success of UAV adoption in warehouses largely depends on either warehouse size or flexibility. Nevertheless, the recent development of drone hardware and software could help firms tackle the limitations related to flight paths, areas, and layouts of warehouses.

• Mission time

Mission time is one of the major objectives of warehouse and logistics management and could be identified as a CF05 of UAV adoption. Several studies applied CF05 as a key performance indicator for assessing both capability and feasibility of UAV implementation in warehouse management. The lower the mission time is achieved, the higher the increased adoption potential. The success of utilizing UAVs in warehouses depends on the performance of CF05. For this reason, previous studies concentrated on the improvement of CF05 by proposing different UAV solutions, e.g., increasing the speed of UAVs [45], expanding the number of drones [71], decreasing charging time [29], reducing energy consumption or the amount of battery charging, and enhancing the capability of pathing and scheduling algorithms (presented in the software factor).

• Costs

Costs (CF06) are usually a critical factor of most business functions, as was indicated in nearly 21% of the focused articles. Unfortunately, most studies did not investigate a specific type of cost. Instead, they emphasize the significance of the overall costs of warehouse management to the success of UAV adoption. However, some previous articles have suggested types of CF06, including UAV and hardware costs, inventory costs, UAV maintenance costs, and delivery costs.

• Drone operation

Drone operation (CF07) was the second most studied operational factor. The most suggested Sub-CF of drone operation is the amount of UAVs or fleet size. The large fleet size of UAVs could potentially support business firms in reducing mission time (CF05) by allocating tasks or inventories [84,85]. However, the drone fleet size is based on several factors, such as loading capacity, flight time, and the number of delivered inventories [71].

The control of drones is considered a part of CF07 that focuses on both pilots and flight control. Currently, the control of UAVs in indoor environments without a fully automated control system still requires a pilot with a high capability level and experience [37]. This situation is because indoor warehouses consist of several obstacles or unfriendly environments affecting the UAVs flights. Some of the threats of controlling the drones were often reported, for example, small corridors, narrow shelf range, and unavailable GPS signal. On the contrary, some recent developments of automated control systems for UAVs are increasing interest and are highly required to avoid human errors. Nevertheless, some limitations of current technologies and constraints on commercial UAV products impacting the control of drones need to be explored. Significant constraints and variables of drone control suggested in past studies consist of cruising speed [37,69,70], noise control [50], and altitude control [18].

Another Sub-CF of CF07, which is rarely mentioned by some articles, is the downtime and breakdown of UAVs [7,114]. The results obtained from the previous studies imply that spare UAVs and maintenance plans should be provided. This operation could reduce downtime, increase performance efficiency, and prevent unsuccessful implementation of UAVs in warehouse management.

• Warehouse

UAVs must cooperate with the existing warehouse operations, technologies, or systems to achieve warehouse management’s fundamental goals or objectives. From the literature review, the significant success factors of integrating UAV with warehouse operations (CF08) could be observed: warehouse layouts, types of warehouses and characteristics, and warehouse management processes.

The type of warehouses and their characteristics are evenly suggested in the most recent studies (e.g., [41,57,60,71]). They are related to space or safety distance, shelf or item position, and rack level. The second most mentioned Sub-CF is management processes or systems directly related to CF08. However, the delivery method and complexity of inventory operations were not often mentioned. Specifically, previous studies suggested that higher product variety [84] and more complex delivery methods [32] are the crucial aspects negatively affecting the success of drone adoption in warehouses.

• Environment

The friendly and supportive environment (CF09) is another highly suggested factor in previous drone and warehouse-related studies. CF09 can directly influence UAV’s efficiency, capability, and performance in warehouse operations. Some previous works roughly discussed the environment of UAV application in warehouse management (i.e., [32,38,46]). In contrast, most studies identified specific environmental aspects, which are light and illumination [14], ground surface [82], noise and disturbance [41], humidity and ambience [14], as well as climate and weather [79].

Among various aspects, light and illumination in warehouses are the most concentrated Sub-CFs fall within CF09. Specifically, insufficient light and poor illumination would go against the precise navigation of drones and the quality of inventory photos or barcodes captured by UAVs. CF09 can affect warehouse activities such as cycle counting, stock-taking, and drone-based delivery. Another highly suggested Sub-CFs in this area is the climate or weather. Poor weather conditions result in unstable and unprecise flights, especially in outdoor environments. All these Sub-CFs of CF09 would further negatively affect warehouse operations’ performance, leading to unsuccessful UAV adoption in warehouses.

• Items and inventories

A stock item is the least suggested CFs in a group of operational factors. The items and inventories (CF10) were only mentioned in two aspects: (1) the number of items and (2) the size and weight of items. About 10.37% of the focused studies emphasized the latter aspect that directly influences the capability of UAV carriage. Large-sized and heavy-weight items make the delivery process difficult and increase the costs of warehouses. It is believed that the more weights and sizes of items the UAV can handle, the more unsuccessful UAV adoption. Therefore, firms aiming to adopt UAVs in warehouse operations should carefully consider their products or inventory dimensions and weights. Another CF10-related factor is the item quantity. The high quantity of items requires a large fleet of drones or a high frequency of operating flights which directly increases the costs of organizations. However, as discussed in the technological software factor, this problem can be solved by utilizing intelligent scheduling algorithms.

3.3.3. Organization

The organization factor (CF11) is a general process or an organization system unrelated to warehouse operations. Two organizational elements were suggested in the past related works, including organizational budgets and maintenance systems. However, compared to the other CFs, the frequency of suggestions regarding these two elements is extremely low. Some articles highlighted the significance of the adequacy of organizational budgets that directly impact the successful implementation, improvement, development, or investment of warehousing UAVs (e.g., [36,77,82]). Another sub-factor of CF11 emphasized in past studies is the maintenance system, i.e., a good maintenance plan and quick response to the breakdown of UAVs [25]. From this aspect, it is recommended that organizations develop maintenance systems incorporating the efficient management of spare drones [114] or spare parts, i.e., batteries [36].

3.3.4. Legislation and Standard

Legislation (CF12) is another less-concerned factor for determining UAV adoption in warehouses. In most cases, UAVs were applied to operate indoor warehouses, which are private or commercial areas. This means the UAV-related study [96] emphasized its less restrictive regulations. Regarding the lack of specific laws for UAV adoption in indoor and private areas, some renowned companies, such as that in [119], decided to wait for indoor regulatory approvals before applying UAVs for warehouse management. Apart from the law, following the UAV adoption best practice or standard is an alternative solution for UAV-based warehouse management. Business firms should consider a warehouse standard to make the warehouse environment conform to the international standard. In response to this concern, one of the related studies [55] applied the government and international warehouse standards to the simulated warehouse environment for drone operations.

3.3.5. Society and Mental

Most studies on drone adoption in warehouse management overlook the analysis and discussion of society and mental factors (CF13). From our intensive review, only two articles [14,50] mentioned essential aspects related to CF13. The first study [14] did not clearly explain its significance to UAV adoption. The study identified the society and mental as CF by referring to another relevant article [4], which highlighted the negative perceptions of people to UAVs from its first task as a military weapon and other concerns, including safety, intimidating appearance, noisiness, and privacy invasion. Similar to the finding found from the reviewed article [50], the noisiness of UAV was a critical mental-related aspect that decreased mental health and quality of life but increased insomnia and anxiety. To reduce the negative effects of noise on annoyance and low work performance of warehouse staff, the improvement of hardware design and control of UAVs were strongly encouraged [50].

All information presented above explains the CFs suggested in previous related studies. We have intensively reviewed and intently concluded these to provide critical success or failure factors for adopting UAVs in warehouse management, as drawn in Figure 7. Moreover, to perceive more information and to highlight the possible improvement opportunities for drone adoption in warehouses, we further analyze and discuss CFs regarding their attention by considering the frequency of their suggestions in past studies. The results depicted in

Table 11. Number of suggestions classified by major CFs and CFs.

Major CFs (% of Suggestions)	CFs	CFs Ranking	No. of Suggestions	Percentage
Technology (53.56%)	CF01-Hardware	1	123	30.22%
	CF02-Software	2	84	20.64%
	CF03-Integrated systems and others	9	11	2.70%
Operation (43.24%)	CF04-Area or distance of operation	3	41	10.07%
	CF-5-Mission time	5	24	5.90%
	CF06-Costs	7	22	5.41%
	CF07-Drone operation	4	27	6.63%
	CF08-Warehouse	6	23	5.65%
	CF09-Environment	5	24	5.90%
	CF10-Items and inventories	8	15	3.69%
Organization (1.47%)	CF11-Organization	10	6	1.47%
Legislation and standards (1.23%)	CF12-Legislation and standards	11	5	1.23%
Society and mental (0.49%)	CF13-Society and mental	12	2	0.49%

present several novel perspectives, which are further analyzed for future improvement opportunities to provide advantages to academics and practitioners as follows.

From the findings presented in Table 11, the most suggested major CFs is technology (53.56% of all suggestions), and the high attention comes from its two CFs, including hardware (30.22%) and software (20.64%), respectively.

In past studies, the highest number of which mentioned CF01, hardware, was criticized for the payload, power consumption, and sensors. Limited or low capacity of drones’ payload and battery power negatively affects UAVs’ capability and warehouse operating costs. We have reviewed the above-filtered articles and other relevant studies to provide possible and potential solutions for these issues. It is found that several studies suggested the improvement of drones’ hardware by reducing or controlling the weights of sensors, batteries, or other parts (e.g., [65,118,120]) and enhancing rotor performance as well as improving UAVs’ carrying or lifting capability (e.g., [4,32,74]). Another major solution to the payload-related problem was the improvement and management of UAV operations. To reduce the carried weights per flight, some studies suggested (1) the management of flight schedules by increasing the number of operating flights [66] or expanding fleet sizes [49,84,85] and (2) the adjustment of warehousing operations by reducing or controlling weights and sizes of carried items [28,29,70].

The improvement and management of UAVs’ batteries and energy consumption is another suggestion for CF01. Low battery capacity and high-power consumption directly influence the high frequency of UAV operating flights and mission times. To improve these problems, most drone and warehouse-related studies suggested the management of battery and UAV operations, including (1) adding more charging units [79], (2) optimizing speed, travelling paths, and operating schedules [79,82], (3) utilizing battery swap strategy [49], and (4) wired power UAV [5]. Apart from these managerial suggestions, there are other possible solutions for improving UAV hardware related to battery and power systems, for instance, wireless charging techniques, hybrid power systems, and electro-magnetic field-based techniques, which were suggested in a drone-related study [121] in other domains.

Another highly suggested drone hardware is sensors. Due to most warehouse activities operated by UAVs being executed indoors, UAVs’ fundamental navigation sensors (GPS and GNSS) and systems are unavailable. This limitation affects the precision and accuracy of UAV positions. To address this issue, onboard sensors, which can effectively function as indoor navigation and localization of UAVs, were emphasized as CF. The most suggested sensor by recent studies was a LIDAR sensor (six papers out of the 15 papers mentioned sensors). This sensor has been identified as having vast advantages to drone operations in indoor or limited-space warehouses, mainly because it is lightweight and inexpensive [44]. Apart from LIDAR, there are other types of sensors and technologies applied for supporting indoor positioning, including LEDDAR, RFID, inertial measurement units (IMU), sonar, simultaneous localization and mapping (SLAM) cameras and stereoscopic cameras, ArUco markers, ultra-wideband (UWB) tags, ultrasonic distance meters, barometers, magnetometers, accelerometers, and gyroscopes. However, the number of studies suggesting and experimenting with this group of technologies is very low. Intuitively, the perception of their effectiveness and feasibility is still very limited. Therefore, there are still a lot of opportunities for studying more on these mentioned technologies or other innovative technologies for improving the navigation of drones in indoor warehouses.

The second most mentioned CFs of technology is software (CF02). The past studies and their critiques were mainly related to the development of drone algorithms which could be classified into two major groups: (1) scheduling and path planning algorithms and (2) localization and navigation algorithms.

The first group of drone algorithms aims to optimize flight schedules and paths. Since the operations of UAVs in warehouses are supposed to manage several constraints and parameters. Therefore, five core techniques can be applied to optimize the outputs, including representative techniques, cooperative techniques, non-cooperative techniques, coverage and connectivity, and security in path planning [122]. From our reviewed articles, a vast majority of articles only mentioned the significance of scheduling and path planning algorithms without the development of algorithms. Some studies have developed new algorithms, most of which applied representative and cooperative techniques. However, few studies focused on non-cooperative techniques and security in path planning. Interestingly, there is a lack of studies focusing on coverage and connectivity. Since each technique has merits and demerits, the applied techniques depend on the study problems, the focused situations, and the context. This finding highlights the unexplored and underdeveloped techniques and their related problems in path-planning activities. Furthermore, we also realized that the past developed algorithms still focused on specific parameters with a limited number of variables that were only related to the specific and focused problems. Therefore, until now, there is still a lack of development of an algorithm considering comprehensive and generic parameters of path planning activities in warehouse management.

On the other hand, the navigation and localization algorithms were mentioned and developed more often than a group of scheduling and path planning algorithms. The development of these algorithms depends on indoor navigation sensors, including inertial sensors, ultrasonic sensors, radio frequency sensors, optical flow sensors, camera motion, and tracking systems [123]. From the deeper reviews, we found that most of the navigation algorithms were developed based on the integration of two sensors which are optical sensors (LIDAR) and camera-based sensors. In contrast, another highly studied group of algorithms depends on only one sensor. The algorithms in this group supported two sensors, i.e., radio frequency and optical sensors, while there were few works proposing algorithms for inertial sensors. From the reviews, we found that the systems that integrate sensors along with the associated algorithms could provide higher navigation and localization precision and accuracy. However, drone algorithms’ development and adoption still depend on problems, environments, or situations.

Apart from the highly suggested factors, the nearly out-of-focused CFs are (1) legislation and standards (CF12) and (2) society and mental (CF13), which contain the lowest suggestions observed in the reviewed articles. For the legislation and standards, this factor was rarely suggested and underexplored since the characteristics of drone operations in the warehouse are mostly limited to private or commercial areas. However, regarding the past suggestions and the high potential of commercial adoption of UAVs in warehouse operations, some improvement opportunities can be further examined and studied, such as UAV legislation and standards in indoor, limited areas, confined spaces, or warehouse operations. Moreover, the regulations or standards considering the problems related to the lowest suggested factor, society and mental, could be another subject since there were some concerns and issues related to mental aspects identified in past studies. Nevertheless, before setting standards or legislation, the impacts of mental and society problems, such as noise, privacy invasion, and safety, should be foremost studied and investigated.

To this end, our key findings can be concluded along with some board summary suggestions for future research described in

Table 12. Key findings and summary suggestions.

RQs	Findings
RQ1: What are the past applications of UAVs in warehouse management tasks?
	UAVs employed in the warehouse management domain are often classified into a type of small drone with a weight lighter than 150 Kilograms. The analysis results from the literature show a wide adoption of multi-rotor UAVs, i.e., quadcopters or four-rotor drones. To select suitable drones for warehouse operations, it is necessary to consider the trade-off between satisfying business requirements and the cost of drone adoption; Among various warehouse management operations, drone-based delivery and stock management were the most studied areas indicating higher attention to UAV adoption for intra-logistics of items and inventory management. However, studies related explicitly to inventory audit and buffer stock maintenance are still neglected. These two underexplored topics exhibit large research opportunities and require more concern.
RQ2: What are the critical factors of UAVs affecting the adoption and implementation in warehouse management?
Hardware technology	Many suggestions on hardware technology for drone-based warehouse management have been given regarding drone hardware (i.e., indoor navigation sensors, battery size, power system, and payload capability) rather than warehouse hardware and network hardware. To expand this research area, there is a need to integrate warehouse hardware (i.e., barcode and RFID) and network hardware (i.e., WiFi, antenna, and network connection) with drone-based technology for more efficient warehouse management operations;
Software technology	Most studies focus more on UAV localization algorithms in indoor warehouses and UAV path planning. These algorithms required further improvement in the compatibility of software, hardware, and other equipment to achieve higher precision; It is rare to find software algorithms related to UAV stabilization, image or video recognition, and disturbance prediction supporting warehouse management operations. These topics shall remain open for in-depth analysis and research discussion;
Integrated system and others	UAVs alone might not successfully execute warehouse operations efficiently unless it is integrated with other systems or technologies such as AGV, UGR, barcode, RFID, and AR.
Area or distance	UAV-based warehouse management is practically operated and benefited only in large enough space that allows high trajectory movements. Comprehensive analysis of UAV mobility in indoor and small-sized warehouses with flexible layouts or shelf size constraints is challenging and needs extensive research and development;
Mission time	Lower mission time is one of the success indicators of employing drones in warehouse management that could be done by various schemes, including increasing the speed of UAVs, adding more number of the operated drones for a particular task, decreasing recharge time, reducing power consumption, as well as improving the path planning and scheduling algorithms. Scholars who aim to elaborate research in this area should address these strategies as alternative solutions to accomplishing the lower mission time requirements;
Cost	Despite most business units putting serious concern about the cost of drone adoption in warehouses, a detailed analysis of this critical factor seems to be ignored in recent research. This situation calls attention to statistical investigations and the classification of cost types to support budget planning or investment in UAV adoption;
Drone operation	The control of drones based on human-centric methods is recommended for managing unfriendly warehouse environments, warehouses with narrow spaces, or containing numerous obstacles. For security purposes, autonomous drone methods with intelligent flight and without human intervention should be applied to prevent harmful effects on humans or to avoid human errors during warehouse operations;
Warehouse	Crucial aspects of drone implementation in the indoor warehouse are related to the types of warehouses, layouts, space, safety distance, rack levels, and the position of shelves and items. Future experimental analysis or simulation research should consider these elements to fly a safety drone and decrease the negative impact of drones in warehouse operations;
Environment	Light and illumination is the most critical factor in operating drones in indoor warehouses, especially when capturing photos or detecting barcodes. Research scholars or practitioners should pay attention to the experimental and environmental settings affecting indoor localization;
Inventory or item	High quantity, large size, and heavy weight items directly influence the capability of UAV carriage and consume high costs. The simplest method of solving this issue is the utilization of intelligent scheduling algorithms rather than improving inventory objects or correcting drone hardware design;
Organization	A sufficient budget and a good maintenance plan are key success factors controlled by the organizations. Future research and investigation are highly encouraged to provide sufficient information and meaningful insights for the decision support of business leaders, drone-related professionals, and stakeholders. Extensive analysis of the return on investment of drone adoption and the strategic fit between drone adoption in the existing warehouse and the economic aspect of warehouse management should be considered;
Legislation	Although limitations imposed by government regulations primarily affect UAV operations, there is no specific law or regulations regarding UAV implementation in indoor warehouses. Therefore, manufacturers or drone practitioners in business firms should employ best practices and keep following international standards of drone-based warehouse management;
Society and Mental	The public perception of UAVs is that they are military weapons or toys rather than commercial vehicles. The real-world benefits of UAVs for a growing number of e-commerce markets and warehouse management should be promoted for widely accepted. This could be done by incorporating the improvement in hardware design and UAV control protocols concerning related individuals’ quality of life and mental health.

.

4. Conclusions

Due to the growth of global e-commerce in recent years, the appreciation of UAV or drone technologies for commercial utilization has become an emerging field. It continually gains higher attention from researchers and practitioners. Among numerous research in this area, this study has provided a detailed summary of the state-of-the-art UAV-based warehouse management solution. The SLR approach was performed to systematically review all related studies between 2006 (the year FAA initially and legally permitted a commercial drone) and 2022. We offer the ultimate research findings and important insights regarding (1) the past applications of UAVs and (2) a set of critical factors affecting their adoption and implementation in warehouse management.

In an attempt to uncover the first RQ, the preliminary result shows the feasibilities of UAVs that could be applied in different warehouse management tasks, i.e., inventory management (such as inventory audit, stock management, cycle counting, stocktaking, etc.), intra-logistics of items, and inspection & surveillance. By investigating the critical factors influencing the adoption of UAVs in RQ2, we found that each factor could be either a success or failure, depending on the situation and implementation contexts. Moreover, the successful adoption of UAV in warehouse management operations is relied on diverse critical factors and constraints, including technology (such as hardware, software, and integrated system), operation, organization, legislation, as well as society and mental. From our findings, a number of studies in this area generate signs of acceptance and positive perception of UAVs in the commercialized domain. UAVs adoption and implementation offer new opportunities to improve and optimize warehouse management operations and their sub-tasks, especially the tasks related to inventory audit and buffer stock maintenance (each accounted for only 8% of all articles). Compared with other warehouse operations (i.e., drone-based delivery and stock management), both tasks are far less concentrated by recent researchers and practitioners. This result highlights the way future research to better address problems and solutions on these two topics using drones.

Despite business that uses UAVs for fully automated or smart warehouses remaining challenged in dealing with more than a dozen factors, the overall impression of UAVs is very favourable and worth the investment. Among 13 CFs extracted from the focused articles, the technological factor was ranked the first concern in most research. This evidence implies its highly influential effects on various tasks, especially drone-based delivery and stock management. Due to the short-change cycle and rapid advancement in technology, it is still required for research expansion by focusing on technological improvement or an innovative solution with a better performance. However, new research dimensions exploring and discussing organization factors, legislation and standards factors, as well as social and mental factors, should not be ignored. From our study, a vast number of factors and sub-factors were discovered, but there appears to be a large difficulty in classifying small elements or sub-factors into groups with similar characteristics. New research on this topic is suggested by rearranging all relevant factors into a more appropriate classification. The most up-to-date insights and suggestions on different factors are highly encouraged. As such, the research implications would truly support firms and authorities to precisely and effectively implement UAVs in warehouse management operations.

In a future iteration, UAV adoption in a real indoor warehouse incorporating a deep learning-based model for barcode detection can be further explored. This future experiment should reflect research progress on hardware, software, and environmental constraints. Enlightened by this review study, we also hope to see a higher number of adoption, implementation, improvement, and better solutions of UAVs for the warehouse management domain.