Process Model for the Introduction of Automated Buses

Abstract

The early deployment of automated electric buses, as a sustainable future mobility concept, depends not only on technical development but also on comprehensive public transportation planning. Local authorities and transportation companies’ planners must strategically incorporate automated buses into the public transportation network on suitable routes. However, current approaches to transportation planning often neglect essential factors pertinent to automated buses, including legal regulations, the status of technological development, and the existing transportation infrastructure. Recognizing the paramount significance of addressing these considerations, this paper endeavors to adapt the public transportation planning process to accommodate the unique requirements of automated buses. To achieve this objective, this study incorporates the requisite input data and framework conditions specific to automated buses into the public transportation planning workflow. Moreover, it elucidates the resultant impacts on the various stages of the planning process and the utilization of mathematical optimization techniques. By employing the devised process model, it becomes feasible to comprehensively assess and evaluate not only the integration of conventional public transportation but also automated buses within a line network. This approach facilitates a comparative analysis of both modes of transportation in terms of costs and benefits, even during the early planning phases, ultimately identifying optimal routes.

1. Introduction

The effects of human-made climate change are omnipresent [1,2]. Rising temperatures, heatwaves, water shortages, and forest fires are just a few examples that result from steadily increasing greenhouse gas emissions [1,2,3,4]. Greenhouse gas emissions are assignable to five major sectors: energy systems, industry, buildings, transport, as well as agriculture, forestry, and other land uses [5,6,7,8,9]. The transport sector was responsible for 20% of global CO₂ emissions in 2021, through the burning of fossil fuels [3,8,9,10]. The largest share in the transport sector results from road transport, which is divided into passenger transport/mobility (60% of CO₂ emissions) and freight transport (40% of CO₂ emissions) [3,7,8]. Through various agreements and documents, such as the Kyoto Protocol, the Paris Agreement, and the Sustainable Development Goals, attempts are being made to implement measures worldwide to reduce greenhouse gas emissions [11,12,13]. For the transport sector, the following reduction measures are possible: use of electric light-duty vehicles, shift to public transport, shift to bikes, and use of electric heavy-duty vehicles, including buses [5]. However, transport is not only affected by climate change, but also has to cope with other challenges. These include, among others, demographic change, globalization, migration, urbanization, resource scarcity, and technological innovations [14,15,16]. Specifically for mobility, climate change, the reduction of CO₂ emissions, demographic change, urbanization, and megacities are the biggest challenges [17,18,19].

Therefore, future mobility must not only become climate-friendly, according to Skukla et al., but must change in general [5,20]. Future mobility is described in the literature as efficient, organized, safe, automated, electrically driven, connected, universally usable, and fair [21,22,23,24]. Examples include traffic network designs (bicycle lanes or 30 km/h zones), regulatory measures (tolls or parking bans), traffic technology and information (vehicle-to-everything or traffic light control), awareness raising (marketing), and improvements to public transport (sharing offers or park and ride) to reduce CO₂ emissions in the mobility sector [19,25]. Furthermore, current trends in the field of mobility include shared mobility, use of electric vehicles, cars on demand, and autonomous vehicles [19,26]. Accordingly, the use of automated vehicles in public transport can be a component of future sustainable mobility and contribute to the reduction of CO₂ emissions [26,27].

Automated buses combine the cost advantages of public transport with the flexibility of private transport, are safe, electrically operated, can be shared, and are usable by the general public [19,28,29,30,31]. According to the current state of development, the buses are referred to as automated buses or automated public transport, since they travel on a virtual line and use LiDAR and radar sensors as well as differential GPS for localization and obstacle detection [28,32,33]. In addition, there is often a driver on board [28,32,33]. That is why the buses are classified between “partially automated” and “fully automated” according to the automation stages defined by SAE [28,34,35,36,37]. However, Waymo already operates a fleet of Level 4 (no operator on board) vehicles in Arizona and San Francisco, and EasyMile also operates a Level 4 automated bus in Toulouse [38,39,40]. Further projects without safety drivers on board are planned in Berlin and Munich [41,42]. Nevertheless, according to the forecasts, highly automated driving with no driver on board (Level 4) will not be achieved until between 2025 and 2030, possibly even later [43,44].

In view of the challenges, especially in relation to climate change, the question arises how the use of automated buses can be made possible earlier. One of the main drivers for the use of automated buses is technological development, such as reliable detection and localization, resilient, robust, and safe system architecture, as well as data safety and security including safe validation and verification of systems [44,45]. However, as Level 4 automated buses depend on infrastructure and operational design domain (ODD), early identification of potential routes appropriate to the current state of technology is also critical to accelerating vehicle deployment [28,45,46,47]. The search for suitable routes for automated buses is part of public transportation and network planning, which is carried out by local authorities and transportation companies [44,48,49,50,51]. There are only a few research contributions to date on the public transport planning of automated buses [44,48,49,50,51]. Gao et al. developed an algorithm for the bus scheduling problem, Weng et al. investigated the Pareto optimal route search, and Hatzenbühler et al. developed optimization methods for network design for automated buses [52,53,54]. Furthermore, a design of transit networks based on agent-based modelling was evaluated [55]. Poinsignon et al. integrated automated vehicles into a public transport network from a cost perspective [56].

On the one hand, the above-mentioned statements show that automated buses pose new challenges to the process of public transportation planning. On the other hand, the literature research shows that currently problems in public transportation planning with automated buses, such as network design or route planning, have been partially investigated, but no holistic approach exists. Accordingly, there has been no transport planning approach until now that takes into account the framework conditions of automated buses, such as legal regulations, the current state of technical development, and the infrastructure of the transport network. However, this is necessary to identify suitable routes for automated buses in a service area at an early stage. As mentioned earlier, this is a component to enable early deployment of Level 4 automated buses and reduce CO₂ emissions in a timely manner. Therefore, the goal of this paper is to develop a public transportation planning process that considers the requirements of automated buses. This process model should support transport planners in the early identification of feasible routes for automated buses.

For this purpose, the following research questions are addressed:

• Which process models for the introduction of conventional buses already exist?
• What are the requirements for a process model for the introduction of automated buses?
• Can existing process models meet the requirements for the introduction of automated buses, or what adaptations are necessary?

In order to answer the research questions, a systematic literature search on transport planning and public transport planning was conducted first. Section 3 describes the methodology according to which the criteria and requirements of automated buses were determined and compared by means of a concept matrix. Necessary adaptations were derived from these results. Based on this, a process model for public transport planning of automated buses was developed, which is presented in Section 4. Finally, the results are discussed, future application possibilities are pointed out, and conclusions are drawn with an outlook on future research needs.

2. Literature Review

The literature review is divided into three parts. Due to the fact that the introduction of automated buses is part of transportation planning, the transportation planning process is investigated first. Second, the literature regarding public transport planning and network planning is reviewed. Finally, in the third subsection, related research areas are examined. The literature review starts by identifying literature within the databases of Scopus and Google Scholar. Relevant literature was selected by analyzing the title, abstract, and keywords. Then, the relevant literature was read and checked for fit. For transportation planning, the search terms “Transportation planning”, “Transportation planning process”, and “Traffic planning” were used. “Bus network design”, “Planning phases public transportation”, “Transit planning process”, “Transit network design” “public transport systems” were the search items for public transportation planning. On-demand public transport services, public transport planning in rural areas, operation of automated buses, logistics planning, and introduction of automated guided vehicles (AGVs) are among the related research areas.

2.1. Transportation Planning

There are many different approaches to transport planning around the world, a selection of which is presented below. In 1970, the US Department of Transportation, Federal Highway Administration, Bureau of Public Roads published a transportation planning process with seven steps. First, goals and objectives are developed. This is followed by an analysis of current conditions. In the third step, trends in population, mobility patterns, and employment, among other factors, are predicted. On this basis, network planning is carried out and, for example, public transport networks are planned. In the next step, estimated movements by mode and route are assigned to the developed networks. Before implementation can take place in the seventh step, the alternatives are evaluated in terms of costs, benefits, and effects [57].

Pas developed a procedure for urban transportation planning, which is divided into three main steps. The pre-analysis phase includes problem identification, formulation of goals, data collection, and generation of alternatives. In the technical analysis phase, trips are modelled in each alternative. Step-by-step, activities are predicted, travel routes are generated and distributed, modes and routes are selected, and emissions are modelled. Finally, in the post-analysis phase, the alternatives are evaluated, and a decision is made, implemented, and continuously monitored [58].

The basis of Meyer and Miller’s approach is to understand the problem. From this, challenges and opportunities for a transportation system can already be derived. The next step is to develop the vision. Based on this, goals and key figures for system performance are derived. The subsequent analysis step focuses on understanding the transportation system based on the collected data and identifying the individual elements. Furthermore, alternative strategies are also considered in this step. Subsequently, the alternatives are analyzed in terms of costs, benefits, and impacts, followed by a decision. During implementation, the defined measures are implemented and priorities are set. The final system monitoring compares the performance indicators of the system with the objectives. Furthermore, this model considers short-, medium-, and long-term planning horizons. In addition, higher-level priorities such as policy, regulations, or finance are incorporated into the process. This process model follows an iterative approach [59].

In the approach by Ognjenovic et al., the traffic planning process is divided into four fundamental stages. First, an information base is built, which includes data collection from static sources and traffic analysis. To collect the data, surveys or traffic counts can be used. Then, the data are correlated to each other and the traffic distribution over the day is calculated. The main component of the third step is to predict the traffic. For this purpose, key figures such as population, number of employees, and degree of motorization are analyzed and further processed in a traffic simulation. Finally, the planning alternatives are evaluated and the most advantageous variant is selected. During implementation, the realization of the plan is continuously monitored [60].

The US Department of Transportation developed a new seven-step planning process in 2018. The starting point is the vision and goals of a region, from which alternative improvement strategies are derived in the second step. The strategies are evaluated and prioritized. A long-term metropolitan transportation plan with measures and goals is then generated. From this, a transportation improvement program is developed, which contains short-term measures that fit the overall transportation plan. The individual measures are implemented and operated as projects when approved. Throughout all phases, the public is involved and has the opportunity to provide feedback. Furthermore, the budget, economic development, air quality, and environmental development are to be included as critical factors in the entire process [61].

In Germany, a five-stage process is followed in transportation planning. The traffic planning process usually starts with external triggers, for example, problem indications or suggestions for solutions from the public, which are taken up and reviewed in the orientation phase. In the problem analysis, guidelines and goals are developed and the current state and future development are analyzed. This is followed by the analysis of measures in which concepts for action are developed, effects are estimated, and variants are evaluated. The fourth phase is the weighing of alternatives with the decision for or against. In the last phase, the concepts are implemented. Impact monitoring, process evaluation, information, participation, and quality management take place in parallel to the planning phases [62].

Although there are differences between the individual process models, the planning process is fundamentally very similar. This is confirmed by further examples from Rahman et al. and Meyer [63,64]. Every planning process starts with an understanding of the problem, whereupon visions, goals, and guidelines are derived. Then, data are collected to determine the current state (as-is analysis) and future developments. After that, different alternatives for possible measures or even concepts for action are developed. In order for these to be evaluated, the effects of the measures must first be determined. In all procedures, the decision with subsequent implementation is made on the basis of the evaluation. In addition to the planning phases, the opinion of society is essential for the planning process, so that the needs of the population are met by the implementation [65]. Booth and Richardson described as early as 2001 the importance of involving the community in the planning process [66]. Furthermore, Wahl uses the example of Swedish municipalities to explain why, how, and when the public can participate in the planning process [67].

Due to the fact that the transportation planning process in Germany is advanced and contains all components of the other phases as well as the involvement of the society, the phases are used for the further procedure and for the model in Section 4 (see Figure 1). For the content of the phases, however, all procedures are included.

2.2. Public Transportation Planning Process

Regarding public transportation planning, a lot of approaches exist.

Table 1. Public transportation planning processes [48,49,68,69,70,71,72].

Ceder and Wilson	Desaulniers and Hickman	Liebchen and Möhring	Häll	Schöbel	Schnieder	Liu et al.
Network design Setting frequencies Timetable development Bus scheduling Driver scheduling	Network design Frequency setting Timetabling Vehicle scheduling Duty scheduling Rostering	Network planning Line planning Timetabling Vehicle scheduling Crew scheduling	Network design Frequency setting Timetabling Vehicle scheduling Crew scheduling	Infrastructure Lines and frequencies Timetable Vehicle routes and schedules Crew schedules	Network planning Line planning Capacity planning Timetable planning Vehicle scheduling Crew scheduling	Infrastructure planning Network design Frequency setting Timetable development Vehicle scheduling Crew scheduling & rostering

shows seven different approaches, starting with Ceder and Wilson in 1986, who divided the public transportation planning process into five steps: network design, setting frequencies, timetable development, bus scheduling, and driver scheduling [48]. Desaulniers and Hickman used the same phases but extended them to include rostering, where deployment is planned at short notice based on actual drivers available [49]. The approaches by Liebchen and Möhring in 2007 and Häll in 2011 again coincide with the approach of Ceder and Wilson [68,69]. According to Schöbel’s process model, public transportation planning starts with the infrastructure [70]. Then, lines and frequencies are planned [70]. The remaining three steps are analogous to the previous procedures [70].

Schnieder divides network and line planning into two work phases [71]. Furthermore, capacity planning is presented as an important step [71]. Finally, the approach by Liu et al. matches the approach by Schöbel in that infrastructure planning is the first step of work [72].

The phases of timetable planning, vehicle scheduling, and crew scheduling are represented in each approach. Furthermore, network planning is also present in each approach. According to Schneider, capacity planning includes the determination of frequencies, which is why public transportation planning is divided into the following five steps for this paper:

• Network planning
• Capacity planning
• Timetable planning
• Vehicle scheduling
• Crew scheduling.

The first step in network planning is traffic route planning, which also includes an infrastructure analysis [70]. Then, the number and position of the stops are determined [70,71]. Linking the stops with each other is part of the route network planning [48,70]. Based on the demand, the trip distance and the vessel size are varied within the capacity planning until the transport capacity and quality is sufficient [49,71]. The timetable for each line is then drawn up based on the route network, the frequencies, and the journey times of the buses [49,71]. Connection planning at the bus stops is also taken into account in order to reduce the waiting time for passengers [71,72]. Vehicle scheduling includes qualitative and quantitative calculation of the vehicles so that the timetable can be fulfilled [49,71]. For this purpose, a vehicle roster with time shares of the vehicles is created and the deployment curve is determined over the course of the day [48,71]. In addition, a vehicle reserve is also planned [71]. Within the framework of crew scheduling, the individual services are determined with regard to working-time laws [48,49]. In order to cover weekly and daily peaks in demand, duty roster planning is also carried out, in which each employee is assigned to a different shift schedule group [49,71]

Each of the individual phases has different target values according to which it is aligned [49]. For example, the stops are optimized according to costs, accessibility quality, and travel time [49,70,71], and the timetable is oriented to the quality goals of stability, number of vehicles, and capacity maximization [49,71]. The goals are in conflict with each other, which is why the planning cannot be trivially transferred into a mathematical model [48,49,71]. Moreover, there are also different planning horizons between the individual phases (long-term network planning and short-term crew scheduling) [73]. Furthermore, there are dependencies between the individual phases, so that the actual travel time, according to which network planning is optimized, can be determined only when the timetable is created [71]. Similarly, the number of vehicles is taken into account in timetable planning, but is not calculated in detail until vehicle scheduling [71,74]. Because of this, the individual planning steps are either run through iteratively or the planning steps are considered in an integrated manner and the effects are determined in this way [48,68,69,75].

Due to the fact that the solution of public transport planning is only possible with the help of mathematical models, a large number of publications deal with this topic. Durán-Micco and Vansteenwegen present an overview of a large number of publications in this context [76]. Since the current work focuses on the processual approach and not on the mathematical modelling, only some examples are listed below. Laporte et al. developed an optimization model that considers the steps of the transit planning model in an integrated manner (trip attraction and generation, trip distribution, mode choice, and traffic equilibrium) [77]. Guihaire and Hao applied different approaches to solve the transit network design and scheduling problem (TNDSP) mathematically and heuristically [74]. Through the combination of the three basic transit network problems—design (TNDP), frequency setting (TNFSP), and timetabling (TNTP)—an attempt was made to formulate an approach to the TNDSP [74]. Kepaptsoglou and Karlaftis further provide an overview of solution approaches to the transit route network design problem (TRNDP) [78]. In 2014, Kun developed a model to solve the transit network design with stochastic demand [79]. Bourbonnais et al. used a genetic algorithm with road network and public transit demand data for generating a solution to the transit network design problem (TNDP) and the transit network and frequencies setting problem (TNFSP) [80]. Liu et al. focused on a multi-objective optimization model for the urban electric transit network design [81]. For this, they used a Pareto artificial fish swarm algorithm [81]. To solve another multi-objective optimization of TRNDP of urban buses, Wang et al. used a heuristic approach [82]. Different modes of transport were assigned to different levels of network taking into account bus routes and city size [82]. Another approach by Heyken Soares investigated a node-based optimization procedure with zone-to-zone trips [83]. The resulting genetic algorithm showed significant improvements in route network optimization when using real-world data [83]. Jiang et al. developed a biobjective model for the integrated transit frequency and schedule design problem (ITFSDP) [84]. Operation costs and travel costs were minimized under the framework of a mixed-integer nonlinear programming problem using a branch and bound algorithm [84]. Nnene et al. presented a simulation-based optimization for designing public transport networks [85]. An activity-based simulation was used and it was shown by the example of Cape Town (South Africa) that efficient networks can be created [85]. The study by Yoon and Chow suggested an artificial intelligence-driven algorithm to combine transit network design with optimal learning [86]. Using New York City (USA) as an example, better results were achieved than with other methods [86].

2.3. Related Research Area

There is no specific process model for the introduction of on-demand transport in public transportation, but adjustments must be taken into account. A particular focus of on-demand transport is on linking individual passenger journeys and vehicle scheduling accordingly [87]. Solving these two requirements correspondingly while minimizing the waiting time is the challenge for which algorithms have already been developed [87,88,89]. Another focus is placed on the ordering and handling process, so that the relevant information on the place and time of the trip is known [88,89].

Major challenges in rural areas are the low population density and the long distances between localities, which makes it difficult to provide high-quality public transportation so that living conditions are similar to urban areas [90,91,92]. The focus in the planning of public transportation in rural areas is on the one hand on network planning, where different network forms such as radial and dispersed are possible [90,91]. On the other hand, timetable planning and vehicle rotation planning must be adapted to demand, which is why a solution approach has already been developed in Switzerland and algorithms for it have been tested [90,92]. However, a specific process model does not exist.

The use of electric vehicles in public transport poses a challenge to route planning [93]. For this reason, Emami et al. developed a procedure with criteria to identify suitable routes for electric buses [93].

Regarding the use of automated buses, there is also no separate process model, but project reports or journal articles have summarized the results and findings from the pilot operations with automated buses [28,33,35,94,95]. A special focus in the use of automated buses is on the selection of routes, since this depends on the vehicle technology on the one hand and on the infrastructure along the route on the other [28,33,94,95]. In addition, the legal framework conditions must also be taken into account [28,33,95]. Since the buses are electrically operated, the charging process must also be considered when selecting the route [28,33]. A first mathematical approach for network planning of automated buses was developed by Hatzenbühler et al. [54].

The subject of logistics planning is, among other things, the selection of vehicles/modes of transport for logistical transport activities and the definition of deployment plans for vehicles. For this reason, approaches from this topic area are also considered. With regard to logistics planning, there are several procedures that can be divided into the following phases: problem identification, defining goals and objectives, analysis and forecast of future conditions, measuring development with material flows, processes and layout, evaluation of different scenarios, selection of the best solution, and implementation [96,97,98]. Furthermore, logistics planning is also optimized in terms of transport costs, number of vehicles, and reduction of environmental impact, among other things [96].

AGVs are also used in logistics. The planning of AGVs is divided into six phases, according to VDI2710: system decision-making, detailed system design, procurement of the AGVs, resources scheduling, planning of modifications, and decommissioning. The aim of system decision-making is to make the decision for an AGV. The objective of system design is to plan the project, particularly with regard to a specification sheet. In this phase, the technical and organizational measures are again examined. The procurement phase has the goal of procuring the AGV as well as starting up and is divided into analysis of the vendor market, invitation to tender, offer evaluation, placing of order, specification, and realization. Subsequently, the operational planning has the goal to guarantee trouble-free operation of the AGV. If there are changes in the use of the AGV, adjustments are made in the operation of the AGV or in the operating environment. If the AGV is obsolete or uneconomical, or if there are changes in the area of operation, the system is decommissioned [99].

3. Method

In order to determine the applicability of existing public transport planning procedures for the introduction of automated buses and to identify gaps in current research, systematic literature analysis is used as a methodology [100]. According to vom Brocke et al., this methodology is divided into five steps: definition of the review scope, conceptualization of the topic, literature search, literature analysis, research agenda [101]. Since the literature review covers a wide range of different sources, the review scope should be determined first [101]. After that, the key words are worked out, which are then searched for [101]. Planning the implementation of an automated bus is part of transportation planning and public transportation planning, which is why these two topics form the review scope. This is complemented by related topics such as logistics. On the basis of the subject areas, key words were derived, such as “transportation planning process”, which were searched for in the systematic literature analysis. Both steps were explained at the beginning of Section 2. The literature search in databases like Scopus or Google Scholar is the third step [101]. The results have already been presented in Section 2.1, Section 2.2 and Section 2.3. After that, the literature is analyzed using a concept matrix [101,102]. Finally, based on the concept matrix, the research gap is derived [101]. The last two steps are explained below.

To create the concept matrix, the most important articles, in this case process models, were first identified. This is described in Section 3.1. Then, the concepts, in this case the criteria that must be met, were identified (Section 3.2). This is summarized in the concept matrix (Section 3.3), and thereupon the research gap is determined in Section 3.4.

3.1. Procedures to Be Investigated on the Basis of the Literature Analysis

In Section 2, different process models have already been described. The different approaches to transportation planning in Section 2.1 were combined into a process model for transportation planning. The approaches reviewed in Section 2, by Ceder and Wilson, Desaulniers and Hickman, Liebchen and Möhring, Häll, Schöbel, Schnieder, and Liu et al., were combined into an approach for public transportation planning. Furthermore, approaches presented in Section 2.3 regarding the use of on-demand transport, the use of public transportation in rural areas, and the use of electric vehicles and automated vehicles were supplemented by two approaches from logistics planning and the introduction of AGVs. In summary, the following procedures are analyzed in the concept matrix. For better presentation, the approaches are abbreviated in the concept matrix (see terms in brackets):

• Transportation planning based on Section 2.1 (Transportation planning)
• Public transportation planning based on Ceder and Wilson, Desaulniers and Hickman, Liebchen and Möhring, Häll, Schöbel, Schnieder, and Liu et al. (Public transportation planning)
• Approaches to the use of on-demand transport (On-demand transport)
• Approaches to the use of public transport in rural areas (Rural areas)
• Approaches to the use of electric vehicles (Electric vehicles)
• Approaches to the use of automated buses (Automated buses)
• Logistics planning according to Nuzzolo and Comi, Safran et al., and Sosunova et al. (Logistics planning)
• Planning of AGVs according to VDI2710 (AGV-VDI2710).

3.2. Criteria and Requirements for a Process Model for the Introductiont of Automated Buses

A process model for the introduction of automated buses must meet specific criteria, which are divided into three categories: general criteria, process criteria, and automated bus criteria. The following general criteria must be met by the process model:

• Effectiveness
• Efficiency
• Correctness
• Transparency
• Acceptance [62,64].

Effectiveness and efficiency refer to the fact that the defined goals are achieved as completely as possible and with a low use of resources [62]. Professionally suitable methods are to be used in planning so that the results are correctly worked out [62]. Furthermore, the methods and processes are to be explained in an understandable way so that they are transparent and comprehensible for external parties [62,64]. Stakeholders and those affected are to be involved in the planning process so that acceptance of the planning and implementation can be achieved among these stakeholders [62,64].

Implementation of the automated bus is part of transportation planning and public transportation planning. Therefore, as a process-related criterion, the entire transport planning process must be adhered to according to the presentation in Section 2.1, with five phases and accompanying processes (impact monitoring, process evaluation, information, participation, and quality management). Analogous to Section 2.2, the procedure of public transportation planning, consisting of network planning, capacity planning, timetable planning, vehicle scheduling, and crew scheduling, is to be used for the analysis of measures. The public transportation planning is to be carried out as iterative or integrated planning, and intermediate results are to be reviewed continuously. For public transport planning, demand must also be taken into account [48,49,68,69,70,71,72].

In summary, the following process criteria must be met:

• Adhere to phases of the transportation planning process
• Comply with phases of the public transportation planning process
• Conduct phases of public transportation planning in an iterative or integrated manner
• Conduct continuous impact assessment and evaluation
• Consider demand [48,49,68,69,70,71,72].

Pilot operations with automated buses show that route and station selection must take bus vehicle technology into account, as not every route is suitable [28,33,95]. In addition, the use of vehicles is affected by weather conditions, so this should be considered when choosing the transport route [28,33,95]. Even if the technological maturity of the vehicles reaches a certain level and enables new routes, the legal framework must also be in place so that the deployment can be carried out [28,33,95]. The use of automated buses has also already been tested on pedestrian and bicycle paths, which is why these paths should be added to traffic route planning [28,46]. Since road infrastructure has a major impact on the introduction of automated buses, an infrastructure analysis should be included in the procedure [28,33,95]. From automation level 4, the buses are on the road without a driver, which is why only stationary personnel must be considered in crew scheduling [28,34]. Automated buses are predestined for use in area operations on-demand [19,24]. This is another requirement for the planning process model. In addition, charging cycles must be taken into account, since automated buses are often operated electrically [28]. In summary, the following criteria for automated buses must be met by the planning process:

• Consider vehicle technology in traffic route and line planning
• Take vehicle technology into account when planning stops
• Take weather conditions into account when planning traffic routes and lines
• Consider legal framework conditions for automated driving in public transportation planning
• Consider pedestrian and bicycle paths in traffic route planning
• Consider infrastructure analysis in public transportation planning
• Consider only stationary personnel in personnel planning except in the operation control center (OCC)
• Consider flexibility (on-demand transport)
• Consider loading cycles (electric buses).

3.3. Concept Matrix for Evaluation of the Process Models

In the following, the previously mentioned approaches are reviewed with regard to their fit as planning procedures for automated buses. Accordingly, the criteria (Section 3.2) are evaluated on a three-level scale (requirement fulfilled, requirement partially fulfilled, requirements not fulfilled).

Table 2. General and processual criteria.

	Criteria	Effectiveness	Effectiveness	Correctness	Transparency	Acceptance	Adhere to Phases of the Transportation Planning Process	Comply with Phases of the Public Transportation Planning Process	Conduct Phases of Public Transportation Planning in an Iterative or Integrated Manner	Conduct Continuous Impact Assessment and Evaluation	Consider Demand
Procedures
Transportation planning		X	X	X	X	X	X			X	(X)
Public transportation planning		X	X	X	X	(X)		X	X	X	X
On-demand transport								(X)	(X)	(X)	(X)
Rural areas								(X)	(X)	(X)	(X)
Electric vehicles								(X)	(X)	(X)	(X)
Automated buses								(X)			(X)
Logistics planning		X	X	X	X		(X)	(X)			(X)
AGV-VDI2710		X	X	X	X		(X)	(X)			(X)

Legend: X: Requirement fulfilled; (X): Requirement partially fulfilled; Blank: Requirements not fulfilled.

shows the results for general and processual criteria.

The planning procedure of transportation planning according to Section 2.1 is standardized and thus fulfils all general criteria. In addition, an impact assessment and process evaluation are carried out across all phases. Demand is considered in the problem analysis phase. The phases of public transportation planning are not included [57,58,59,60,61,62].

All five phases of public transportation planning are included in the procedure of public transportation planning (Section 2.2). Furthermore, feedback as well as continuous impact assessment take place and the demand is considered. As a standardized procedure, the process should contain all general requirements. Acceptance is rated as partially fulfilled, since the sources do not directly address it, but key figures are taken into account in the evaluation, such as memorability of the timetable [48,49,68,69,70,71,72].

Approaches from on-demand transport, rural areas, and with electric buses are not standardized approaches and each include only parts of public transportation planning [87,90,93]. In these three research areas, algorithms are used that consider the questions in an integrated way and determine the effects [88,92,93].

Both approaches to the deployment of automated buses are project reports and neither are a standardized approach. Phases of public transport planning are carried out, but not completely and not iteratively. In addition, there is no impact assessment. Demand is considered in the form of a potential analysis [28,33,94,95].

The two procedures from logistics are standardized and use similar phases as the transportation planning procedure but with different names. Transport volume or demand (analogous to traffic demand) is calculated or vehicles and personnel are dimensioned (analogous to capacity planning), which is why some phases of public transportation planning are included. The extent to which employees are involved in the planning process in terms of acceptance is not described [96,97,98,99].

The transportation planning approach is described in very general terms, so the detailed requirements for automated buses are not addressed (see

Table 3. Automated bus criteria.

	Criteria	Consider Vehicle Technology in Traffic Route and Line Planning	Taking Vehicle Technology into Account When Planning Stops	Take Weather Conditions into Account When Planning Traffic Routes and Lines	Consider Legal Framework Conditions for Automated Driving in Public Transportation Planning	Consider Pedestrian and Bicycle Paths in Traffic Route Planning	Consider Infrastructure Analysis in Public Transportation Planning	Consider only Stationary Personnel in Personnel Planning Except in the OCC	Consider Flexibility (On-Demand Transport)	Consider Loading Cycles (Electric Buses)
Procedures
Transportation planning
Public transportation planning					(X)					(X)
On-demand transport									X
Rural areas									(X)
Electric vehicles							(X)			X
Automated buses		X	X	X	X	X	X			X
Logistics planning		(X)	(X)				(X)
AGV-VDI2710		(X)	(X)				(X)

Legend: X: Requirement fulfilled; (X): Requirement partially fulfilled; Blank: Requirements not fulfilled.

) [57,58,59,60,61,62]. Public transportation planning already takes charging cycles into account and includes legal framework conditions—although not for automated driving [48,49,68,69,70,71,72]. The approaches of on-demand transport and from rural areas refer exclusively to flexible on-demand transport [87,90]. Emami et al. focus only on the charging cycle and analyze the infrastructure to identify suitable routes for electric buses [93].

In the case of projects with automated buses, vehicle technology, weather conditions, legal framework conditions, and charging cycles are included in the route determination [28,33,94,95]. Furthermore, an infrastructure analysis is carried out [28,33,95]. In logistics planning and in the use of AGVs, the operating resources (vehicles) are selected first and only then are the processes developed [96,97,98,99]. This is comparable to the criterion that vehicle technology is taken into account in line and stop planning. Furthermore, the working environment is adapted to the new operating equipment, which has similarities to the infrastructure [96,97,98,99].

3.4. Research Gap

In summary, it can be seen from Table 2 and Table 3 that no approach meets all the criteria for a planning procedure for the introduction of automated buses. For this reason, a separate process model has been developed. To ensure that the basic criteria of transportation planning are included, the basic structure of the process model is based on the phases from Section 2.1. The five phases of public transportation planning are arranged in the phase of the measure investigation (see Section 2.2). As an innovation of this work, the criteria and framework conditions for the deployment of automated buses (see Table 3) are integrated into the public transport planning process. This means that the vehicle technology of the automated buses (performance and parameters of operation) is included as an input parameter in the traffic route planning and in the bus stop planning. Furthermore, the weather conditions, the infrastructure, and the legal framework in relation to automated driving are now also considered in the context of traffic route planning. Moreover, traffic route planning does not only refer to roads but also includes pedestrian paths and bicycle paths. Flexibility regarding on-demand traffic and charging cycles is added as new input data in the route network planning. The use of automated buses has an impact on the stationary staff of the OCC, which is why the workload for monitoring automated buses must be taken into account in crew scheduling. Finally, acceptance, which is already generally embedded in the process, needs to be extended with regard to automated buses. These new input data mean that the work steps in the individual phases change (e.g., infrastructure costs must be calculated in traffic route planning) and that the mathematical optimization must be expanded with regard to new criteria (e.g., personnel costs of the OCC in driver deployment planning). The individual input parameters and the changes in the overall process are described in the process model in Section 4. The process model is developed on the basis of the following assumptions:

• The process model is developed for the case that the buses have reached Level 4 of automation and can be used without a driver.
• The process model is designed for public transport planning that considers conventional buses and automated buses.
• The possibility of on-demand traffic is considered, but is not the focus of the study.

4. Results—Process Model for the Introduction of Automated Buses

The process model for the introduction of automated buses is shown in Figure 2. The public transportation planning is arranged in the third phase, together with the estimation of the evaluation and variant assessment. The individual phases of public transportation planning are explained below. In the orientation phase, it is briefly described which specifications are newly considered based on the introduction of automated buses. The problem analysis is extended so that new data (e.g., infrastructure analysis or analysis of the vehicle technologies) must be collected, which are required in the measure investigation. In the context of the study of measures, the process of public transport planning is first described in detail. In each phase, the criteria for the use of automated buses mentioned in Section 3 are then integrated. The resulting impact of the new input data in terms of changes to the work steps, mathematical models, and results are described at the end of each phase. Finally in the consideration, decision, and implementation phases, it is briefly discussed what changes occur in the phases as a result of the topic “introduction of automated buses” and what adjustments have to be made.

4.1. Orientation

The starting points of the transportation planning process are, on the one hand, deficiencies or problem indications from citizens or politicians that the current public transport service is not satisfactory, the traffic volume is too high, stops are not accessible, or emission limits are being exceeded. On the other hand, new opportunities can also arise through technological progress, such as automated buses that are operated electrically. These vehicles are expected to make public transport more affordable, reduce emissions, and solve the last mile problem [21]. Already at this point, it is recommended to collect traffic data and to gather problem descriptions as well as proposals for measures. What is new for the introduction of automated buses is that it is already recommended to conduct a market analysis with regard to the vehicles and their performance (including speed, width, duration of operation) as well as research regarding the legal framework for the introduction of automated vehicles. A market analysis is also suitable for researching costs and preparation time so that rough resource requirements in terms of time, personnel, and financial resources can be determined for the further procedure. On this basis, the decision to proceed with the transportation planning process is made.

4.2. Problem Analysis

The problem analysis starts with the development of guidelines and objectives. In the second step, the current situation is analyzed. In general, traffic volume, modal split, the existing public transport network, and structural data of the population within the area under consideration are analyzed [62]. For public transportation planning, the route network, transport demand, the means of transport, travel speeds, and the accessibility of the stops are also analyzed. With regard to the route network, pedestrian and bicycle paths should also be included in the analysis, as this is required for automated buses. The use of automated buses on a route depends on the technological maturity of the vehicles and the infrastructure of the road [28,33]. Therefore, it is recommended to perform a detailed market analysis of the buses in this step, which is not carried out for conventional buses. In addition to the suppliers, the technical performance characteristics of the buses and the circumstances under which they can be used are particularly relevant. The market analysis should also focus on projects or operations with automated buses, in order to be able to assess their use in public spaces. Besides the vehicle analysis, an analysis of the infrastructure data is indispensable at this point. For example, according to Soteropolous et al. and Beckmann et al., internet connectivity, speed limits, lane width, and traffic volume should be collected [46,103].

Along with the infrastructure analysis, it is recommended to identify possible infrastructure measures (e.g., introduction of a parking ban) and to determine costs for implementation. In order to determine the costs, the district or municipality obtains cost estimates from possible companies. Part of the problem analysis involves going deeper to research the legal framework of automated buses so that the area of application is also secured from the legal side. Since the demand is generally surveyed for all modes of transport, it is also recommended to survey the acceptance of citizens regarding the use of automated buses, since this new technology can trigger skepticism. Acceptance is best measured through surveys [104]. Finally, analysis of weather conditions is also recommended, as this influences the use of automated buses.

The inventory is followed by the analysis of correlations between the individual data and the conditions assessment. In the case of status assessment, the results are compared with the defined goals. A SWOT analysis (strengths, weaknesses, opportunities, and threats) is suitable as a methodology, and also indicates starting points for future measures.

4.3. Measures Investigation

Based on the preliminary work, the aim of the measures investigation is to develop an action plan with a bundle of measures and a procedure for implementation, as well as costs, time, and prioritization [62]. Within the scope of the study of measures, public transportation planning is carried out with the aim of identifying routes for conventional and automated buses, defining stops, creating a timetable, and planning the deployment of vehicles and personnel. It is determined that the entire public transportation planning process for automated buses will be iterative. Accordingly, there are feedbacks between the individual phases, which have already been described (Section 2.2). For example, there is feedback between stop and line planning. There are also feedbacks in the individual phases. One example is capacity planning, which is carried out repeatedly until the transport quality and capacity meet the targets.

Impact identification and evaluation is carried out in each phase of public transportation planning, because planning is always optimized according to the objectives of the respective phase. Furthermore, three milestones in the process are suitable for listing results and impacts and sharing them with political decision-makers. These include the completion of route network planning (phase 1), the completion of capacity and timetable planning (timetable—phase 3), and the end of the process after crew scheduling. Basically, the individual phases differ in the planning horizon and objectives; route network planning is part of strategic planning, timetable and capacity planning are part of tactical planning, and vehicle and crew scheduling are part of operational planning. This applies accordingly with regard to the fact that, among other things, changes to the route network plan cannot be decided at short notice. However, the shift planning can be adjusted, for example, monthly or weekly. Nevertheless, all phases should be run through once or, if necessary, several times so that a line network plan and a timetable are produced. For this reason, it is the opinion of the authors that an action plan, which is the result of this phase, should contain a route network plan, a capacity plan, a timetable, and vehicle and crew scheduling plans at the end of the public transportation planning process. The aim of the action analysis is to develop several action concepts that differ in terms of how the objectives of the individual phases are expressed. The goal of the variant evaluation is to compare the different characteristics of the objectives with each other. This step therefore plays a central role in the entire transportation planning process.

The procedure of public transportation planning (see Section 2.2) is used for the development of the action concepts, and the individual phases are described below.

4.3.1. Network Planning

The aim of network planning is to plan the future routes of the public transport service on the basis of the planning objectives [48,74]. This is achieved in three steps:

• Traffic route planning
• Stop planning
• Line network planning [70,71,72,75].

First of all, the traffic routes in the study area are determined [70,72,74,105]. This includes roads, footpaths, cycle paths, and also, where available, the existing public transport route network [70,74,105]. The goal is to determine the entire traffic route network for public transport vehicles [74]. For this purpose, the first step is to delineate and describe the entire network [105]. Then, in the second step, the categories of traffic routes are determined [106,107]. With this result, conventional bus planning continue with stop planning.

In principle, conventional buses can be used almost anywhere where motor vehicle traffic is present, but there are also restrictions with regard to weight, width, or public access to streets (residents only) [71]. In the case of automated buses, this analysis is much more complex, because automated buses depend on a variety of infrastructure characteristics, as Section 4.2 shows. Therefore, the use of conventional and automated buses per road is determined by analyzing the infrastructure in detail. This is the new second step of traffic route planning. In addition to the infrastructure analysis, the infrastructure costs of the automated buses are also determined (work step 3 in

Table 4. Overview traffic route planning following [70,71,72,74,105,106,107].

Input Data	Work Steps	Methods	Results
Route network incl. road, footpath, and cycle path network Public transport network Vehicle performance for conventional and automated buses (including speed, width, deployment options) Infrastructure analysis Infrastructure costs per measure Legal framework for automated buses Weather conditions	Delineation and description of the network Analysis of the use of conventional and automated buses Calculation of infrastructure costs per route for the use of conventional and automated buses Determination of the traffic route categories for the public transport vehicles	Evaluation model to determine feasibility and infrastructure costs. Traffic categories e.g., based on FGSV	Traffic route network for conventional and automated buses Travel times (also for automated buses) in the traffic route network Deployment ability of automated buses, effort for deployment, and infrastructure costs per street (evaluation model)

Input data, work steps, methods, and results that are specific to the automated bus are marked in green.

). An evaluation model is used as a new method to determine the extent to which a road can be used by a conventional or automated bus and what infrastructure costs are incurred. The result of the evaluation model is shown as an example in Figure 3. The usability of a roadway for an automated bus is shown in descending order from green (high usability) to red (low usability).

On the basis of the evaluation model, a decision is made about the use of an automated bus on a route. In the evaluation model, the infrastructure measures are evaluated monetarily on the one hand, but also non-monetarily if the quality of the cost calculation is not sufficient. Due to the early stage of planning, cost estimates are subject to certain assumptions and may deviate during implementation. An overview of the traffic planning is shown in Table 4, where input data, work steps, methods, and results that are specific to the automated bus are marked in green.

The stops are then planned with the aim of defining the number of stops in the area [75]. Optimization is based on the three divergent objectives:

• Access costs
• Access quality
• Travel time [48,49,71,74,75,77].

The access costs result from the number of stops that have to be built and maintained [75]. The stops provide the spatial accessibility in the service area [71]. The percentage of inhabitants living in the area served by a stop and the average distance from the stop to the passenger describe the quality of access [75,77]. The travel time consists of all time components (access to the start stop, travel time(s) possibly including transfer time, departure from the final stop to the destination) from the start address to the destination address [71]. Therefore, a large number of stops minimizes the departure distances to the stop, but the travel time is increased, because the buses have to stop more often [71,74].

In the first step of stop planning, the catchment areas are determined based on the means of transport and the location of the settlement area [71,74]. From the area served, which describes the linear distance to the stop, the actual walking distance and the resulting travel times to the stop can be calculated [71,74]. Stops are positioned (second step) on the basis of the area served by the stop, the traffic significance of the settlement area, and the traffic boundary conditions [71]. The traffic significance of a settlement area is based on demand [49,71]. Finally, a check is made with regard to the three objectives of access quality, access costs, and travel time, which are formulated mathematically [71]. A solution is iteratively calculated by optimization procedures [48,71,78]. From the first draft, which can be established intuitively, the development quality and the development costs can be determined directly [48,71]. The travel time can be checked only once the timetable has been created [71]. Therefore, the results are passed on to the line and network design, the capacity planning, and the timetable planning [71]. Based on the draft timetable, the travel times are then checked again and if they are insufficient, the stop planning and the whole process is repeated [71]. Suitable optimization methods include the approaches by Furth and Rahbee (2002), Saka (2001), Schöbel (2005), or Khondaker and Wirasinghe (2013) [75,108,109,110,111]. Results of this phase include number of stops, stop spacing, definition of stops (including intersection nodes or terminal nodes), lost time, waiting time, and development costs (see

Table 5. Overview stop planning following [48,49,71,74,75].

Input Data	Work Steps	Methods	Results
Settlement structure Traffic demand in traffic cells Acceptance of automated buses Traffic network plan Route network plan Means of transport Travel speeds and accelerations of public transport vehicles Performance data of automated buses (turning radius and boarding height) Accessibility and accessibility to the bus stop Suitability of bus stops for a turning operation	Choice of the mode of operation Determination of reasonable catchment areas (linear distance) Positioning of stops Verification of areal development	Mathematical modeling of the objectives “travel time” and “stop spacing” Optimization procedure for the objectives “access costs”, “access quality”, and “travel time”	Number of stops Stop spacing Lost time due to braking and acceleration Waiting times at stops for boarding and alighting Cost of stops Forms of operation offered Bus stop and/or door-to-door service

Input data, work steps, methods, and results that are specific to the automated bus are marked in green.

) [48,49,70,74,75].

Since automated buses can also be operated on demand, the operating modes in the service area are determined before the stops are defined. The choice is between scheduled, directional, and area operation. The modes of operation that best fit the demand for transport in the settlement structure are selected. This step is newly included in the process model for automated buses. For automated buses, new input parameters must be taken into account. The technical characteristics of the buses, such as the turning radius and the boarding height, must be added, as these change the exact positioning and design of the bus stop in the public space (work step 3). In addition, acceptance of the new technology influences demand, which affects the access quality objective. Access quality is also influenced by the legal framework conditions. The infrastructure costs of automated buses affect access costs, and bus performance affects travel time. These new input values have an impact on the mathematical modelling and must be included there.

For example, Furth and Rahbee select stops (from stop j to stop N, given that the stop preceding j is i) for a line from a given set of stop positions by minimizing the total cost to the operator and passengers [108]. The total costs here consist of net walking time cost per unit time (Z_w), riding delay cost per unit time (Z_r), and operating cost per unit time (Z_op) [108]. Accordingly, the objective function f is as follows [108]:

f (j; i) = min Z_w + Z_r + Z_op

(1)

Automated buses in this case would reduce operating costs Z_op compared with conventional buses and provide different access times and delay times.

In principle, the additional consideration of automated buses does not affect the execution of the steps. However, additional input data such as acceptance and performance of the vehicles are newly added. Furthermore, the previously created traffic route plan is included here, which was created on the basis of conventional and automated buses.

Within the framework of line network planning, the individual lines or area operations are first formed and then combined in a network, depending on the selected operating forms [48,71]. The planning design is evaluated by determining and assessing the following objectives:

• Economic efficiency of operations (including timetable kilometres, degree of timetable effectiveness, fare revenue, fuel consumption, ability to use automated buses/infrastructure costs)
• Improvement of the service (e.g., frequency of transfers, travel time)
• Improvement of operational quality [49,70,71,74,75,105]

Before the line planning can be carried out, an overview of different line forms and network forms is obtained in the first step [71]. Subsequently, the individual stops are connected with each other via direct connections, if possible [49,71,74]. Depending on the demand for stops, a line form is also selected [71]. There are interactions between route and stop planning, which have already been described above [105]. When planning routes, care must be taken to ensure that the route of public transportation is on roads that are suitable for the use of vehicles [105]. In an area operation, the stops do not have to be combined.

In the third step, the individual lines and area operations are linked together at selected stops and gradually condensed so that a line network is created [48,71]. Intuitive methods or optimization methods can be used as design methods for the line network [75]. Intuitive methods are divided into planning design and impact calculation [75]. The planning is based on certain design rules and the effects are determined on the basis of a computer-aided effect model [75]. A well-known method is by Friedrich [75,112]. Optimization methods transform the line network planning into a mathematical problem with an objective function that is optimized [70,75]. Gattermann et al., among others, developed some optimization procedures [75,113]. Based on the optimization methods, the fourth step is to determine and evaluate the effects on the three mentioned objectives.

In general, the four work steps do not change when considering automated buses. However, when planning the routes (step 2), special care must be taken to ensure that the autonomous buses travel on roads that are suitable for the vehicles. Also in this phase, new input parameters have an influence on the evaluation of the route network. Automated buses have a significant impact on economic efficiency, in more than one way. Infrastructure costs and lower speed have a negative impact on the cost structure, while the operating costs due to lower personnel costs and fuel consumption have a positive impact on the economic efficiency. This is explained by the example cited by Gattermann et al., creating a line network with sufficient quality until the demand for public transport can be served. For the cost calculation, the length of the line (cost_length), the number of edges (cost_edge), and fixed costs (cost_fixed) are added to the total costs of a line (cost_line). This example includes the infrastructure costs of the automated buses in the cost of the length of the route, which would increase the cost significantly [113].

cost_line = cost_length + cost_edge + cost_fixed

(2)

In summary,

Table 6. Overview line network planning following [48,49,70,71,74,75,105].

Input Data	Work Steps	Methods	Results
Route network Traffic route network (also for automated buses) Travel times (also for automated buses) in the transport route network Location of bus stops and bus stop layover times Suitability of stops for a turnaround. Traffic demand matrix Acceptance of automated buses Assignment of traffic cells to bus stops Mode of operation Deployability of automated buses, effort for deployment, and infrastructure costs per street (evaluation model) Fuel consumption of vehicles	Overview of line forms and network forms Line/area operation planning Network planning Determination and evaluation of effects	Design procedure for network planning: intuitive procedure or optimization procedure. Optimization of the network with regard to “economic efficiency of operation”, “improvement of the service”, and “improvement of the operational quality”.	Line network with individual lines Lines with all stops Timetable kilometres Timetable hours Timetable efficiency Estimation of the number of vehicles Fare revenue Other fixed and variable line costs Fuel consumption/costs Greenhouse gas emissions Infrastructure costs Average transfer frequency Number of direct riders Average travel time

Input data, work steps, methods, and results that are specific to the automated bus are marked in green.

shows an overview of the line network planning.

4.3.2. Capacity Planning

Within the capacity planning, the number and type of offered seats is calculated, appropriate to the demand in the service area [71]. The capacity is variable due to the variable trip intervals (temporal) and the vehicles used (spatial) [71,74]. Since capacity results from the product of trip frequency and the size of the vehicles used, the goal of capacity planning is to minimize the use of operating resources so that energy consumption and greenhouse gas emissions are reduced and fewer personnel (operating costs) are tied up [48,71,74]. This is achieved by deploying vehicles according to demand [49,71,74]. Other goals are to maximize passenger comfort, in terms of trip frequency, seat availability, minimizing standing times, and increasing safety [49,70,71].

The individual steps of this phase are shown in

Table 7. Overview capacity planning following [48,49,70,71,74,75,105].

Input Data	Work Steps	Methods	Results
Planning goals Traffic demand (results of traffic survey) Acceptance of automated buses Traffic route network Route network Line sections Vehicle types (size and number of seats and standing places)	Determination of the line/area load Identification of relevant cross-sections Selection of trip distance Selection of vessel size Verification of transport capacity/quality	Evaluation and further processing of demand and acceptance data Calculation of transport capacity and quality on the basis of the objectives of operating resources, passenger comfort, and safety Verification of quality criteria on the basis of regulations	Cycle times (frequencies) Number and type of vehicles Size of vehicles Cross-sectional load Stop load

Input data, work steps, methods, and results that are specific to the automated bus are marked in green.

. Essential for capacity planning are passenger numbers related to the day and the line or area [49,71]. As part of capacity planning, demand data are further processed to show the load for different seasons, days of the week, traffic times, lines, and routes [49,70,71]. The data are summarized in the cross-sectional load, which can be calculated for lines or routes and is given in the dimension passengers per time unit [70,71]. Another important objective is the stop load, from which the boarding and deboarding behavior can be derived and on the basis of which the equipment at the stop (information and weather protection facilities) is also determined [71]. On the basis of these results, the decisive cross-section (with the heaviest load) is identified [71].

Next, the trip distance and the vessel size are selected and the above-mentioned key figures are calculated on this basis [71,74]. Furthermore, the occupancy rate and the maximum vehicle occupancy can be calculated [71]. Finally, the transport capacity and quality are checked by comparing the results with the specifications from guidelines as well as local targets [71,105]. If the capacity and quality are sufficient, the timetabling process continues [49,71]. Otherwise, the trip distance and vessel size must be adjusted again and the quality criteria recalculated [49,71]. Timetable planning is carried out in parallel with capacity planning, which is why there is feedback between these two work steps [71,75].

Considering automated buses in capacity planning does not entail any change in the five steps. However, acceptance has an impact on the demand data, which is used in step 1 to calculate the line load. If automated buses are used, operational resources will be reduced as no drivers are used and the buses are electrically powered. Furthermore, safety is increased by this new technology, which influences the evaluation. Low operating costs would make it possible to increase the frequency of the vehicles and improve passenger comfort.

4.3.3. Timetable Planning

Transportation companies are obliged to prepare timetables [71]. Timetables contain the route with all stops as well as departure and arrival times [48,71,73,74]. The objectives of timetable planning are:

• Minimization of travel times (higher speeds, less waiting time, reduction of transfer frequency)
• Maximization of schedule stability (reliability of operations, buffer times)
• Optimization of number of vehicles
• Maximization of capacity
• Optimization of energy management (range increase, reduction of charging times)
• Memorability of the timetable (constant travel times) [49,71,72,74,75]

The number of vehicles is the result of vehicle scheduling, which is why the schedule is taken into account as an input variable in vehicle scheduling and has an impact on economic efficiency [49,71,72,74,75]. In the following explanations, only timetable planning for road-based means of transport is discussed (see

Table 8. Overview timetable planning following [48,49,71,72,73,74,75,105].

Input Data	Work Steps	Methods	Results
Planning goals Traffic route network (also for automated buses) Route network Operational efficiency (number of vehicles) Traffic demand Cross-section load Bus stop loading Time shares in the timetable Vehicle performance/driving characteristics of conventional and automated buses (e.g., speed, fuel consumption) Experience with the use of automated buses (e.g., travel times) Track layout of the route Permissible speed	Definition of the time components of the timetable Determination of travel time per section and line (target travel time) Timetable generation per line Connection planning Adjustment of individual timetables Determination and evaluation of effects	Derivation of time components from rules and regulations Determination of travel time on the basis of calculation or operational performance with subsequent evaluation Timetable generation based on quality objectives (travel time, timetable stability, number of vehicles, capacity maximization, energy management, memorability of the timetable) using various methods Connection scheduling: comparison of the schedules of two lines and calculation of a target offset time at transfer stops Evaluation on the basis of the quality objectives “operational constraints”, “reliability of operation”, “total travel time” and “transfer distances”	Timetable for line network Timetable per line Bus arrival and departure times per stop

Input data, work steps, methods, and results that are specific to the automated bus are marked in green.

).

The basis for timetable planning is the line network [73,74]. First, the time components of the timetable, e.g., transport time, turnaround time, or empty running time, are defined on the basis of rules and regulations [48,49,71]. This is followed by determination of the travel time. The travel time is measured during operation or during a test run and statistically evaluated [71,105]. The target travel time is determined on the basis of a frequency distribution [49,71,105]. Together with capacity planning and the resulting cycle times, the results are summarized in a timetable [49,71,75,105].

The timetable is determined based on the target travel times and with regard to the above-mentioned objectives [48,71]. Since the objectives are divergent, the solution is calculated using optimization methods [49,75,105]. The timetable is based on different target concepts, such as interval timetable, rendezvous timetable, or integrated interval timetable [71,73,75]. Three methods have proven successful in determining the timetable. These are “synchronization of individual trips,” the “quadratic semi-assignment problem”, and the “periodic event scheduling problem” [74,75]. The system cycle from the capacity planning is used to define the timetable times and to increase the memorability for the passenger [74,75].

In the fourth step, the connections between the lines are planned [71,74]. The aim of connection planning is to take into account operational constraints (length of stops), to increase the reliability of operations, and to minimize the additional extension of the total travel time and transfer distance [49,71,74]. Since at least two lines have to be coordinated with each other when transferring, an offset time is determined to ensure connection reliability [49,71,72]. Finally, the effects are determined and evaluated by comparing the results with the design and optimization objectives [105].

If a flexible mode of operation is used for an automated bus, it must be decided whether a scheduled or a non-scheduled mode of operation is selected [114]. In the case of timetable-based services, the procedure is the same as described above, although the timetables and travel times cannot be calculated exactly, as stops are only called at on demand [114]. This must be communicated to customers in advance. For a non-scheduled form of service, no timetable is created, but the operating period is defined. In this case, the procedure for timetable planning would change fundamentally and be significantly shorter.

Assuming that a timetable needs to be created, timetable planning for automated buses does not add any new work steps, but some work steps are affected. Regarding step 2, it is also recommended for the automated bus to measure travel times if they are already available from the county or transit agency, since obstacle detection by the sensors leads to more frequent obstructions and unscheduled stops. If no vehicle is available, the travel time must be calculated based on the experience of the manufacturer and other operations with automated buses. In addition, new input data (traffic route network and vehicle performance) must be included in the evaluation due to the use of automated buses.

An example of the “periodic event scheduling problem” is provided by Schmidt and Schöbel, who are developing a procedure that allows route choice to be considered in an integrated manner when designing the timetable [115]. The total travel time (c_TTF (π)) is the sum of the travel times of a passenger (π_j − π_i, where π is a timetable and _i and _j are activities) multiplied by the weight of each activity (w_(i,j)) [115].

$${\mathrm{c}}_{\mathrm{TTF}} \left(\mathsf{\pi }\right)=\min {{\displaystyle \sum }}_{\left(\mathrm{i}, \mathrm{j}\right)\in { \mathrm{A}}_{\mathrm{op}} }{\mathrm{w}}_{(\mathrm{i}, \mathrm{j})} \times ({\mathsf{\pi }}_{\mathrm{j}} - {\mathsf{\pi }}_{\mathrm{i}})$$

(3)

The use of automated buses has a variety of effects on Equation (3). Due to low operating costs (see Section 4.3.2), it is possible to use more vehicles and thus reduce the frequency of trips. Lower waiting times would reduce the total travel time. However, the total travel time is increased if the automated buses travel at lower speeds than conventional buses. Furthermore, the use of automated buses is independent of personnel, which is why the buses can also be used at higher frequencies during off-peak times (e.g., at night). These changed conditions due to the use of automated buses must be taken into account in the timetable.

4.3.4. Vehicle Scheduling

On the basis of the timetable, it is the task of vehicle scheduling to link individual journeys to vehicle rotations (see

Table 9. Overview of vehicle scheduling following [48,49,71,73,75,88,105].

Input Data	Work Steps	Methods	Results
Timetable with line assignment Time buffers for delays Requirements for the scheduling method Time components in the timetable Technical constraints of conventional buses Deployability of automated buses, effort for deployment, and infrastructure costs per street (evaluation model) Cost structure	Qualitative planning of vehicle deployment Quantitative planning of vehicle deployment Quantitative planning of vehicle reserves	Determination of the technical boundary conditions of the vehicles Linking of individual trips as well as calculation of the number of vehicles and deployment curve Quantitative determination of vehicle deployment on the basis of the quality objectives “economic optimization” and “stability of operations” with the use of optimization methods Calculation of the operating and maintenance reserve on the basis of rules and regulations	Wagon schedule Number of vehicles per operating day Operating curve per day Operating and workshop reserve Number of transfer runs, service times, empty times, service kilometers, empty kilometers

Input data, work steps, methods, and results that are specific to the automated bus are marked in green.

) [49,71,74]. In public transportation, a vehicle rotation describes the connection of individual trips of a vehicle from the beginning to the end of the operating day [48,71]. In vehicle turnaround planning, a distinction is made between line-specific (final stop = starting stop) and non-line-specific (transfer to another line) [73,75]. The transfer increases the flexibility of the transportation company, as the scheduling of personnel can be better planned [71,75]. However, the empty runs generate additional costs, which has a negative impact [75]. Therefore, the goal of vehicle scheduling is to serve the schedule with the lowest possible number of vehicles with the following objectives:

• Economic optimization (total number of vehicles deployed, maximization of vehicle turnaround times, minimization of empty runs, and minimization of infrastructure costs when using automated buses)
• Stability of operations (repeatability of operational processes, attention to operational reserves, and reduction of disruptions) [48,49,71,73]
• Both objectives are to be optimized under consideration of operational and technical framework conditions [49,71]. This includes assigning suitable means of transport to the transport operations (capacity), adapting the length of the rotations to the refuelling/loading operations, taking into account additional refuelling or loading times, and considering the capacity of depots with regard to parking spaces [71]. Furthermore, it must be taken into account in the planning that the timetable must be adhered to, that different vehicle types must be used, and that maintenance intervals must be observed [48,49,105].

The first step is the qualitative planning of the vehicle deployment, in which the technical boundary conditions of the vehicles are assigned to deployment profiles [71]. Curve running characteristics, the clearance gauge, the vehicle width, the permissible total weight, the length, as well as the loading cycles and range of the buses have an impact on which roads can be used by conventional buses [71].

The second step involves quantitative planning of vehicle deployment [71]. Here, individual trips are connected via link elements and a vehicle schedule is created that contains several courses [71]. A course describes all parts of a vehicle’s journey from leaving the depot until its return [71]. The number of vehicles on a line is determined on the basis of the round-trip time of a line and the frequency of the service, which are then added together for the entire line network [71]. At this point, optimization methods (e.g., by Bunte and Kliewer (2009) or approaches from the field of operation research) are used to best achieve the opposing goals of economic efficiency and operational stability while taking into account the operational and technical constraints [49,75,116,117]. In order to ensure reliability in the operational process, additional vehicles must be kept in reserve, which are then planned [49,71]. For economic reasons, only as many vehicles as necessary should be purchased as reserves [71].

The use of automated buses does not affect the general process of vehicle scheduling, but it does affect the design of the individual work steps. The qualitative evaluation of the vehicle deployment of automated buses has already been determined in the evaluation model (traffic route planning). Therefore, the results can be used directly here, which is why no new qualitative assessment needs to be carried out. In step 2 of this planning phase, the use of automated buses has a major impact, since no flexibility has to be taken into account with regard to driver deployment. As a result, automated buses can be used to plan a completely line-specific schedule. This means that no empty runs have to be taken into account, which simplifies the entire process and reduces costs. This influences the mathematical optimization. Bunte & Kliewer show the assignment model that minimizes the sum of operating costs (c_ij) per connection (x_ij) between trip _i and trip _j [116].

$$\min {{\displaystyle \sum }}_{\mathrm{i}=1}^{\mathrm{n}}{{\displaystyle \sum }}_{\mathrm{j}=1}^{\mathrm{n}}{\mathrm{c}}_{\mathrm{ij}}{ \times \mathrm{x}}_{\mathrm{ij}}$$

(4)

The use of automated buses minimizes operational operating costs, which means that more vehicles could be used.

If the automated bus is used in an area operation without a timetable, no wagon plans can be defined. Here, only the number of vehicles is used on the basis of capacity planning. A change of drivers must be taken into account with conventional buses, whereas this is not the case with automated buses. However, empty runs can be minimized through vehicle scheduling and the possibility of trip bundling [88].

4.3.5. Crew Scheduling

After the vehicles have been allocated to the timetable, the personnel are assigned to the timetable in this phase and the duty schedule and duty rosters are created [48,71,74]. In the context of public transportation planning, a distinction must be made in personnel scheduling between the driver and stationary personnel such as control center or maintenance shop employees [49,71]. The goal of crew scheduling is to fulfil the timetable with a minimum number of drivers [71,105]. For this purpose, the following objectives are included in the evaluation:

• Efficiency of personnel deployment
• Minimization of disruptions to the operational process
• Assignment of the appropriate personnel to the rosters
• Creation of a suitable working environment
• Compliance with legal framework conditions [48,49,71].

The efficiency of personnel deployment is measured in terms of personnel costs [48,71]. These can be minimized by reducing the number of employees or labor costs and by making working hours more flexible [49,71]. Dividing an employee among different services does have advantages in terms of flexibility [71]. However, the changeover points are prone to disruptions and must be minimized [71]. For an employee to be assigned to a vehicle on a line, general expertise in the transit system, expertise in the vehicle, and route knowledge of the line are necessary [71,73]. When creating the working environment, occupational health target criteria and social influencing factors must be taken into account [71]. Legal framework conditions such as the Occupational Health and Safety Act or the Working Hours Act must be observed when planning personnel deployment [48,49,71,73].

The planning of personnel deployment for passenger transportation is divided into three parts (see

Table 10. Overview crew scheduling following [48,49,71,73,74,75,105].

Input Data	Work Steps	Methods	Results
Timetables/wagon timetable Legal framework for working, driving, and break times Requirements for interchanges and breaks Cost structure Estimation of the workload for monitoring automated buses	Formation of individual duties Duty sequence planning Adjustment of personnel presence to the circulation mass Crew scheduling for maintenance (also for automated buses) Crew scheduling for the control center (also for automated buses)	Determination of the duty schedule based on the quality objectives by using algorithms For maintenance: calculation of the effort and derivation of a resource and schedule plan to determine the qualitative and quantitative personnel deployment For the control center: three-stage procedure as for the driving personnel with regard to the effort and the operating time period.	Duty schedule for driving personnel Duty rosters for driving personnel Duty schedule for stationary personnel Duty rosters for stationary personnel

Input data, work steps, methods, and results that are specific to the automated bus are marked in green.

) [71]. First, individual services are formed by determining the place and time of the start and end of the shift [49,71,75]. The legal framework conditions regarding daily and weekly working times as well as rest and break times are taken into account [48,71]. The duty schedule, which contains anonymous services, is optimized with respect to the target variables mentioned above [49,71,73,75]. A solution can be found by algorithmic approaches, e.g., by Borndörfer et al. [49,75,118]. At the end of the duty schedule design, different objectives such as duty schedule time, vehicle turnaround time, wage hours, or the duty schedule efficiency can be evaluated [71]. The crew scheduling problem can be thought of as a set-partitioning problem, where the cost of a service (c_j) is multiplied by the performance of a service (x_j) and summed over all services j [118].

$$\min {{\displaystyle \sum }}_{j \in J}{c}_j\times x_j$$

(5)

Since the transport company usually provides an operating service seven days a week, but the employees are only allowed to work five days a week, a duty sequence with personal weekly duties is planned in the second step so that each employee works the same number of days on the weekend and during the week [49,71,73,74]. There are three types of duty rosters: fixed duty rosters, elective duty rosters, and request duty rosters [71].

In the third step, the staff presence per shift is adjusted to the demand of the population [71]. This is possible by varying the duty roster in terms of working days (e.g., fewer staff on weekends, more during the week) or in terms of duty time (e.g., more staff in the morning than at noon) to accommodate peaks in work traffic [49,71].

During the fourth step, which is described in the literature, stationary personnel for maintenance are determined [49,71]. To calculate personnel deployment, the effort required is determined on the basis of monitoring or benchmarks and then transferred to a resource and schedule plan [49,69]. From this, the qualitative and quantitative personnel requirements are calculated [49,71].

In terms of driver scheduling, there is a big difference between automated and conventional buses. The first three steps are used exclusively for conventional buses, since automated buses do not require a driver. Thus, no duties and no duty sequences have to be planned and changed. Equation (5) again shows that there are no costs for drivers and that the mathematical optimization is automatically minimized.

With regard to scheduling for maintenance, it cannot be assumed that repairs and maintenance measures will have to be carried out more frequently when automated buses are used. However, repairs and maintenance measures may affect the on-board computer or the sensors, which is why specially trained personnel are needed (change in qualitative personnel requirements).

As a new fifth step, the deployment planning of stationary employees for the OCC is recommended, since more personnel capacities are needed there for the operation of automated buses [49,71]. For the calculation of stationary personnel in the OCC, the same three-stage procedure is recommended as for the driving personnel. However, duty scheduling is not carried out in relation to individual driving cycles, but in relation to the workload over the entire operating period. This must be covered by the control center personnel so that the vehicles have a contact person at all times. This becomes particularly interesting when using automated buses, as these can be used all day. The workload is either collected during operation or estimated. Optimization is based on the five quality objectives. According to the literature, the workload will change and increase with the use of automated buses, which must be taken into account in duty scheduling [119,120]. Duty rostering and the adjustment of personnel presence to the circulating mass is carried out analogously to the driving personnel. This step is completely new and almost exclusively due to the use of automated buses.

4.4. Consideration and Decision

In the fourth phase, the object of planning is first classified. Public transportation planning is part of municipal planning and is based on the local transport plan as well as the regional transport plan and the state transport plan. In the following consideration, the extent to which the objectives of local public transport planning correspond with the higher levels is also examined [62].

The previously prepared action concepts and variant evaluations are set in relation to each other within the framework of the consideration. This means that the individual quality objectives of the action concepts are compared with each other, and it is checked whether alternatives have an absolute or relative advantage over the others. Furthermore, it is examined whether minimum standards such as the accessibility quality or the travel time correspond to the specifications, and whether legal basic conditions (among other things working time) have been considered. Political institutions also involve the public in this phase to ensure transparency about the effects and evaluation. Subsequently, a decision is made on the implementation of the public transport planning, although rejection or readjustment can also be a result of the decision-making process. The decision is ultimately reviewed by the higher-level approval authority and the line concession is awarded [62]. Consideration of automated buses does not affect the approach to decision making for this phase.

4.5. Implementation of the Concepts

If the implementation of one or more lines with conventional but, in this context, primarily automated buses is approved, the implementation will take place. The implementation of projects for the introduction of automated buses has already been briefly described in Section 2.3. First, planning is initiated by putting together a project team and establishing project goals and a time/resource plan. Then, the vehicles are procured by means of a public tender, if this has not already been carried out. In parallel, the route and infrastructure are prepared for the use of automated buses on the basis of the feasibility assessment. At the same time, approval and authorization are obtained for the operation of the buses in public areas. After delivery of the vehicle, the route is programmed into the automated bus. This step ends with several test drives, during which the driving time can also be measured. In addition, training of the employees takes place so that the buses can be controlled remotely or on site and the employees are prepared for emergency situations. The operating period and frequency are determined and the timetable is drawn up. As part of operational planning, vehicle deployment and driver deployment (in terms of stationary personnel in the control center) are updated and operations can begin. Compared with the conventional bus, the automated bus requires significantly more effort for route setup, licensing, and employee training.

4.6. Tasks across All Phases

Parallel to all phases, the public is informed and, if possible, involved. This ensures that the concerns of the public are included. Furthermore, acceptance by the population is present, so the public transportation planning is not hindered during implementation. Continuous monitoring and evaluation are important not only during the study of measures but throughout all phases. This ensures that the procedures and schedule of the transport planning process are adhered to. To this end, conditions must be recorded and appropriate measures defined. Ultimately, quality management ensures that goals, procedures, and responsibilities are clearly defined, transparently communicated, and adhered to throughout the entire transportation planning process.

5. Discussion

In considering the first research question, the elaboration in Section 2.1 and Section 2.2 shows that a number of approaches exist for transportation planning and public transportation planning. Therefore, the basic approach to transportation planning is not questioned by the authors. From the literature research on the public transportation planning process, it is also noticeable that many sources deal with mathematical modelling but only a few sources deal with the work steps and contents of the individual phases [48,49,70,71,72,74,75]. However, the approaches are designed to be general. Specific topics such as on-demand transportation, electric buses, or automated buses are not considered in detail. On the other hand, the results in Section 2.3 show that so far there have been only project reports but no detailed procedure models for the mentioned topics. The question arises whether this is because the issues are new and not researched, or not yet as widespread, or do not require a proprietary approach. From the results of this publication, it is clear that no fundamentally new approach needs to be developed. Nevertheless, the individual topics significantly change the process of public transport planning, which is why a separate approach is needed. Therefore, the authors are of the opinion that the topics are currently not yet widely used or not frequently applied, which is why no procedural models exist.

With regard to the second research question, pilot operations and publications were analyzed in Section 3.1 and Section 3.2 with regard to the use of automated buses [28,33,94,95]. From this, the requirements for the process model were derived. The conceptual matrices in Section 3 clearly show that the current planning processes of transportation planning and public transportation planning cannot be used for the introduction of automated buses without adaptations. Due to the current state of the art of automated buses, the vehicles cannot be used on every route and are dependent on the infrastructure [28,33,94,95]. In addition to technical feasibility, the legal framework must also be considered [28,33,94,95]. However, the examples from the areas of on-demand transport and electric buses show that no new process model needs to be developed [87,88,93]. For this reason, the authors decided to integrate the requirements of automated buses into the current planning processes. This answers the third research question. Certainly, technological advances may change the operating conditions of automated buses. However, the process model has been developed in such a way that future changes can be taken into account.

The procedure for the introduction of automated buses developed in Section 4 is based on the transportation planning and public transportation planning processes. The requirements of automated buses are implemented in every phase of public transportation planning. Therefore, it is legitimate to ask the question: does the consideration of automated buses significantly change the overall process? Figure 4 summarizes the changes that result from the consideration of automated buses in the planning process. While in the original traffic route planning almost all traffic roads are suitable for conventional buses according to [71], a more elaborate assessment of the infrastructure is needed for automated buses. Not only do new input values (vehicle technology, infrastructure costs, or legal framework conditions) have to be included, but the work steps and methods also change. Due to the infrastructural requirements of the buses, new target values must also be taken into account. New input parameters are also taken into account in stop planning, and an additional check is made as to which form of service is to be selected for the vehicles (new work step). In line network planning, the individual work steps do not change, but new input parameters are used, which is why new target values must be taken into account. According to this, the inclusion of automated buses changes not only the process but also the results significantly compared with the original methods of Schöbel, Schnieder, Liu et al., and Guihaire and Hao [70,71,72,74]. Accordingly, the impact of automated buses on network planning can be said to be large. The work steps of capacity and timetable planning do not change as a result of the automated buses. Only new input values are included, which is why only existing target values already used by Ceder and Wilson, Desaulniers and Hickman, and Schnieder are influenced [48,49,71]. For this reason, the changes in these two phases are evaluated as small. As a result of the fact that no driving personnel have to be taken into account in vehicle scheduling, the quantitative vehicle planning changes significantly there. In addition, the infrastructure costs that now have to be considered change the objectives. Overall, the authors assess the influence of the automated bus on this phase as medium. Automated buses do not require driver scheduling, unlike the original Ceder and Wilson and Desaulniers and Hickman methods [48,49]. For this reason, the scope of work of the OCC changes significantly, which is why this must be recalculated and taken into account in the target figures. Accordingly, the influence of automated buses on this phase is also rated as high. In summary, it can be stated that the criteria of automated buses can be well integrated into the procedure of public transport planning and that this has a significant influence on the procedure and the result. For this reason, the authors are convinced that the developed process model for the introduction of automated buses makes an important contribution to public transport planning. Since this procedure is also based on the fact that automated buses are dependent on the state of technological development, the procedure has been developed in such a way that the future development of the vehicles is incorporated. Basically, the procedure is only valid until the buses are fully autonomous (Level 5). Since this can take until 2040 or later according to the literature, this procedure is important for the further use of automated buses [43,44].

Compared with Gao et al. who include range and charging conditions in the selection of bus routes, the method developed here considers many more parameters (e.g., infrastructure, vehicle technology, acceptance) in network planning [52]. Hatzenbühler et al. consider infrastructure costs in network planning in the same way as this approach [54]. However, they refer exclusively to network planning and not to the entire process as in this paper [54]. In contrast to this method, Poinsignon et al. also focus only on network planning, where vehicle performance (including speed) is taken into account. Basically, the previous publications in the literature focus on mathematical modeling. This publication, in comparison, focuses on the procedural approach. The authors are convinced that both types of consideration are important. Furthermore, it is believed to be helpful if the structure and procedure of the individual phases are clearly defined, because this also facilitates mathematical modelling. This is shown by the descriptions in Section 4.

In the developed planning approach in Section 4, on-demand transport and electric propulsion of buses are considered, but the integration could have been elaborated in more detail. This was not attempted in the context of this publication, as the focus is on the changes brought about by the use of automated buses. There is a need for future research at this point.

The evaluation of variants in Section 4.3 is the central component of the public transport approach and the basis on which implementation is decided. With respect to automated buses, the operational capability or route suitability is determined based on the evaluation model that is developed during traffic route planning (Section 4.3.1). Further research is needed to determine the structure, parameters, and effects of the evaluation model. Furthermore, the result of the evaluation model will be further used in the context of public transportation planning. The extent to which the result of the evaluation model can be related to the other quality objectives also requires further research.

6. Real-World Applications

To explain the practical application of the process model, it is important to understand how public transportation is planned. In Europe, each country prepares its own transport infrastructure plan based on the trans-European networks in the EU [121,122]. Depending on the subdivision of the respective country, a transport plan for a state or region is derived from it [121,122]. The lowest level is formed by municipal transportation planning [121,122]. In this context, the district develops the requirements for the public transportation service, which are recorded in the local transportation plan [121,123]. Accordingly, the local transportation plan forms the basis for municipal public transportation planning (Figure 5) [121,123]. On this basis, it is the task of the respective transportation companies to design the public transportation service in accordance with the set requirements [123]. Depending on the planning horizon, local public transport planning can be divided into three areas. Within the framework of strategic planning (planning horizon of three to five years), network planning is carried out by planning the lines and line combinations [71,72]. At the lower level, capacity and timetable planning are partly updated annually and thus assigned to tactical planning [71,72]. The vehicle and crew scheduling of operational planning is takes place within a planning period of months or weeks [71,72].

Accordingly, the procedure can be used by transportation planners of municipalities or transportation companies in the context of local public transportation planning. Municipalities can use the planning procedure presented in Section 4, for example, to draw up the local transportation plan and define the requirements for the public transportation service. For transport companies, it is possible to plan various aspects of the public transport system in a city or district using the method developed here. The process model can be used in all planning phases, from the route network to driver scheduling. The advantage of the approach developed in this paper is that not only conventional buses are considered, but the use of automated buses in an existing or new route network can also be evaluated. In addition, the two modes of transportation can also be compared with each other. The identification of suitable routes for automated buses, which is possible with this method, is the first step to integrate automated buses sensibly into a route network and to solve the problem of the last mile. Based on the results of public transport planning with automated buses, it is then possible to prepare routes accordingly and implement an operation.

The results serve not only local authorities and transportation companies but also political decision makers, because it is possible based on the results of this process model to prepare political decisions regarding the implementation of new or modification of existing (automated) bus routes. Consequently, the process is an important element in implementing a public transportation service in accordance with the requirements of the local transportation plan and the citizenship.

Last but not least, the results also offer added value for the manufacturers of automated buses because it becomes apparent on which routes the buses can be used. Conversely, it can be deduced which future development steps are necessary to increase the service area and cover more use cases.

7. Conclusions

Automated buses that are operated electrically can contribute to CO₂ reduction in the transport sector. At present, however, the technological development is not complete and the buses are mostly used with an operator on board. In order to promote the use of automated buses in the future, technological development is an important variable. Furthermore, however, the planning of operational areas for these vehicles can make a contribution. For this reason, this article focuses on an approach for the introduction of automated buses.

Since the literature shows that automated buses have special requirements for infrastructure and the routes on which they are used, the idea is developed that these special features must be taken into account in the planning process for automated buses. Therefore, approaches from transportation planning and public transportation planning were researched and examined with regard to the criteria for the use of automated buses. Since no planning procedure currently covers all the criteria for automated buses, a separate planning procedure for the introduction of automated buses was developed.

In summary, this research work establishes, for the first time, a framework for transportation planning that comprehensively considers the contextual factors associated with automated buses. These factors encompass legal regulations, the present state of technological advancement, and the intricacies of the transportation network infrastructure. In general, this approach to transportation planning is innovative because it breaks with traditional approaches, incorporates new technologies, and takes a holistic perspective that considers legal, technical, and infrastructure aspects simultaneously. It is a forward-looking and adaptive approach to address evolving transportation challenges and opportunities. The experimental validation was derived from a comprehensive assessment, encompassing the scrutiny of published literature, insights gathered from pilot projects, rigorous comparative analyses, and invaluable stakeholder feedback sourced from a specific project.

The focus of this work is not on the design of the overall process, which is why the phases of transportation planning and public transport planning are taken from the literature. Rather, the innovation of this work is to integrate the criteria and requirements of automated buses into the existing process. This mainly refers to new input parameters that bring automated buses into the process. Another contribution of this paper is to elaborate and describe the resulting changes to the work steps and methods of the individual phases of public transport planning.

The results of this work show that there is little change in the steps in most phases of the public transportation planning process. However, the new input parameters of automated buses change the evaluation of the quality objectives. This has a large impact on the outcome of each phase. For example, automated buses reduce operating and fuel costs. Yet, the low speed of the buses also increases the travel time. Accordingly, this has an impact on the route network, frequency, and timetable. The biggest changes can be observed in vehicle scheduling and crew scheduling. Since automated buses are not staffed, fewer empty runs have to be scheduled and driver scheduling is no longer necessary for these buses.

With the help of this adapted procedure, it is possible to consider and evaluate the use of not only conventional but also automated buses in a route network. In this way, the two means of transport can be compared with each other. Future research is needed on the integration of on-demand transport and electric vehicles into the planning procedure. In addition, an evaluation model that assesses the operational capability of automated buses on a route needs to be designed in detail.