On-Site Demonstrations of C-ITS Architecture and Autonomous Vehicle Operation in the City of Trikala, Greece

On-site demonstrations of C-ITS architecture and autonomous vehicle operation in the city of Trikala, Greece Christos Ballis1, Anna Antonakopoulou2, Evangelos Tsougiannis3, Tasos Theodoridis4, Alexandros Liazos5, Odisseas Raptis6, Loukas Vavitsas7, Elena Patatouka8, Angelos Amditis9 1 ICCS, [email protected], Greece, 2 ICCS, [email protected], Greece, 3 ICCS, [email protected], Greece, 4 ICCS, [email protected], Greece, 5 ICCS, [email protected], Greece, 6 e-Trikala, o[email protected], Greece, 7 e-Trikala, [email protected], Greece, 8 e-Trikala, [email protected], Greece 9 ICCS, [email protected], Greece Abstract. This paper delves into the Cooperative Intelligent Transport Systems (C-ITS) architecture of the on-site pilot conducted in the city of Trikala, Greece, as part of the EU-funded project, IN2CCAM. The pilot tests and demonstrates innovative Cooperative, Connected, and Automated Mobility (CCAM) services through the deployment of autonomous vehicles (AVs) and their integration with a C-ITS platform. Featuring a fleet of autonomous electric minivans on a predefined route equipped with advanced smart digital infrastructure, this initiative aligns with the city's vision to tackle congestion, promote the use of shared transport services, and enhance the equity of the local transport system. Furthermore, a Mobility as a Service (MaaS) mobile application enables the public to plan multimodal trips, integrating on-demand shared passenger transport services with AVs, public transport services, shared micromobility solutions, and active mobility, such as walking or cycling. The proposed C-ITS architecture includes: a) AV fleet management and monitoring, b) a Green Light Optimal Speed Advisory (GLOSA) functionality to achieve smoother and more fuel-efficient journeys, c) a traffic-based green wave system to contribute to high traffic efficiency by adjusting traffic signals according to real-time congestion levels, and d) C-ITS messages between the infrastructure and the AVs to alert about uncontrolled crossings of Vulnerable Road Users (VRUs). This paper aims to outline the architectural components of this pilot and discuss future work, which will include evaluating the feasibility, effectiveness, and societal impact of the proposed innovations in the local ecosystem of Trikala. Keywords: CCAM, C-ITS, AV, MaaS, GLOSA, VRU, pilot. 1 Introduction Cooperative, connected, and automated mobility (CCAM) prioritizes moving people and goods along the road networks in a safe, quick, cost-effective, comfortable, and 2 environmentally friendly manner. Cooperative Intelligent Transport Systems (C-ITS) technologies, automated vehicles (AVs), advanced traffic management systems, and Mobility-as-a-Service (MaaS) platforms, among others, are utilised to this aim. The scope of the proposed work is to implement a reference C-ITS architectural schema with the scope to integrate an autonomous passenger transportation service into the local transportation ecosystem of the city. Within the framework of the EU IN2CCAM project, the city of Trikala, Greece, serves as a hub for the on-site demonstrations of C-ITS solutions discussed in this paper. The Trikala pilot features a fleet of two electric, retrofitted, and automated minivans which have been granted permission to travel without a driver on a predefined route equipped with advanced smart digital infrastructure. The 9.6 km long route runs between the city centre and the Science and Sports University and has been selected as it serves the railway station, the city’s thematic park, sports facilities, and major suburbs and villages that are underserved from the current public transport line. The autonomous service mainly targets university students and residents of areas on the outskirts of the city. It offers on-demand trips which can be booked via a mobile application that acts as a multimodal journey planner for the city. Users can use the mobile application to book trips between fixed locations along the route. In addition, to ensure seamless integration with the local public transport system, the autonomous service stops have been selected to align with existing underserved bus stop locations. Fig. 1. Screenshot of the AV route in Trikala. 3 In summary, the solutions implemented in this pilot project aim to improve the safety and efficiency of the autonomous service, as well as to improve its interaction with mixed traffic flows along safety-critical and congested sections of the route. Namely, the following advancements will be developed and deployed: a) Advanced traffic management services to enhance the management and monitoring of the AV fleet in mixed traffic flow along the selected route. b) A traffic-based green wave system which contributes to higher traffic efficiency by adjusting traffic signals according to real-time traffic volume. c) A Green Light Optimal Speed Advisory (GLOSA) functionality to achieve smoother and more fuel-efficient journeys. d) C-ITS messages between the infrastructure and the AVs to alert about uncontrolled crossings of Vulnerable Road Users (VRUs). e) A Mobility-as-a-Service (MaaS) mobile application to enable the public to plan multimodal trips. This app integrates the on-demand shared passenger transport service with AVs, public transport options, shared micromobility solutions, and active modes like walking or cycling. This paper is structured as follows: Section 2 outlines the architecture of the C-ITS solutions included in this pilot, Section 3 discusses the a priori evaluation of the innovations, Section 4 presents the future work, and lastly, Section 5 concludes this paper. 2 Architecture 2.1 Overview The architecture supporting the on-site demonstrations of the autonomous service in Trikala, Greece, is designed to seamlessly integrate various components, all aimed at enhancing urban mobility and safety. An outline of the architecture is presented below. 4 Fig. 2. Architecture of the C-ITS solutions implemented in the Trikala pilot study. The four main components of this architecture are: a) The physical and digital infrastructure, which include Road-Side Units (RSUs), installed at signalized intersections, On-Board Units (OBUs), mounted on the roofs of the AVs, traffic lights equipped with traffic light controllers, traffic cameras with object detection capabilities, and a MaaS application. b) The ITS backend, which includes a set of backend processes along with a local database to store the recorded data. c) The Traffic Management Centre (TMC), which incorporates fleet and operational management capabilities related to the AV service, along with additional TMC capabilities. d) The AVs, with their software and hardware subcomponents. 2.2 Traffic-based green wave system Traffic congestion is among the most challenging problems in urban management. To this aim, green wave strategies aim to optimise traffic flow by coordinating signal timings along corridors to reduce congestion and emissions. Such strategies play an essential role in intelligent transportation systems (Zheng et al., 2020). The outcome of a green wave system would be less stressed motorists, more reliable transportation, punctual bus services, decreased noise pollution, and fewer pollutants (SWARCO, 2024). On the contrary, green wave policies could lead to enhanced attractiveness of privately owned cars, and, thus, increased gas emissions, especially on some primary arterials (Bloder and Jäger, 2021). Regarding studies that examine the impacts on traffic after applying green wave strategies, one could refer to Tseng, Chang and Kuo (2017), who simulated the utili- 5 sation of vehicular ad hoc networks (VANETs) to apply virtual traffic light strategies. The results implied that traffic flow and average speed are improved by this tactic, while energy is also saved. In a different study, Shi et al. (2020) proposed a highefficiency multi-intersection coordination algorithm based on Vehicle-to-Everything (V2X) communications, simulated its application, and showcased, likewise, that average speed increased, whereas volume-to-capacity percentage and number of stops decreased. In this project, this solution involves the use of stationary artificial intelligence (AI) traffic cameras installed at one of the main intersections along the route in Trikala where the electric, retrofitted, and automated minivans are permitted to travel autonomously during the Trikala demonstrations. The cameras provide a continuous insight into the real-time traffic levels of the area and, consequently, allow for adjustments to the relevant traffic signal programs. Prerequisite for this automated process is the storage and analysis of historic traffic data, which include classified traffic counts and speeds. Fifteen-minute intervals are used, and typical traffic levels are determined for each. The process involves analysing the current level of traffic, and by comparing it against the typical situation, it applies incremental changes to the signal phases. In other words, the duration of the green time allocated to the main route adjusts, within a pre-selected range, according to the current traffic volume. This is achieved with the development of Application Programming Interfaces (APIs) which interact with Traffic Light Control and Management software and apply the relevant signalling programs. 2.3 Green Light Optimal Speed Advisory (GLOSA) The GLOSA application provides precise and timely information about traffic signal timings and traffic light positions via infrastructure-to-vehicle (I2V) communication, which allows drivers to cruise with more constant speeds and with less stops at traffic lights (Katsaros et al., 2011). In this pilot, the GLOSA functionality is being tested and demonstrated with one of the two automated minivans. Signal Phase and Timing Extended Messages (SPATEM) and MAP (topology) Extended Messages (MAPEM) are broadcast by an RSU located on Pylis street and received by the OBU of the AV. The former includes real-time information about the traffic light signal phase and timings, while the latter provides information about the topology and geometry of the signalized junction. A web application has been developed to display this information in a clean, intuitive, and easy-to-comprehend way to the safety driver. The output of the process is shown on a custom Human-Machine Interface (HMI) on a tablet installed inside the minivan. Furthermore, the Automated Driving System (ADS) of the vehicle utilises the GLOSA-recommended speeds to adjust the vehicle’s speed accordingly, if that is possible given the traffic conditions and the presence of obstacles. A total of five signalised intersections have been equipped with RSUs which can exchange messages via ITS-G5, cellular network (e.g., 4G/5G), and Wi-Fi/Ethernet, among others. For this GLOSA application, the following processes are implemented. First, a backend process has been developed to access the SPATEM, request the vehicle’s location and speed information, and then combine them so that they are used as 6 input in the GLOSA algorithm. The GLOSA calculations take place on the fly and the suggested speeds are displayed to the safety drivers on a web application. In this demonstration, the C-ITS messages are received from the OBU when the minivan is within approximately 250m to the signalised junction. The GLOSA algorithm involves the main following steps. a) Starts with checking whether there is a relevant traffic light in proximity. b) If yes, the distance to the traffic light and the time required to reach it (denoted as Ttl) are calculated. c) Determines the colour of the traffic light at Ttl. Depending on the colour of the traffic light, it follows different processes to calculate a recommended speed. d) The process loops back to check for relevant traffic lights in proximity, continuously adjusting the speed as needed. The flow chart of the GLOSA algorithm is shown in Fig. 3. Fig. 3. Flow chart of the GLOSA algorithm demonstrated in Trikala. 7 The range of recommended speeds has been set so that they are within the range that the AV is allowed to cruise by law, in other words a maximum speed of 30kph. Lastly, very low recommended speeds are excluded to avoid creating safety hazards due to very slow movement in traffic. Fig. 4. Screenshot of the web application that displays the GLOSA-recommended speeds. 2.4 VRU detection and warning to approaching AVs VRUs include a wide range of road users, such as pedestrians, cyclists and powered two wheelers (Scholliers, van Sambeek and Moerman, 2017). They have been identified by policymakers as a group of traffic participants that deserves specific attention in efforts to reduce the number of fatalities and serious injuries (European Commission, Directorate-General for Mobility and Transport, 2020). VRUs are more susceptible to road traffic accidents and fatalities than other road users. Improving their safety stands out as a primary concern for many municipalities and governments, and various measure are taken to mitigate accident risks and increase road safety (Hejazi and Bokor, 2024). Previous VRU-safety projects aimed to improve safety by utilizing on-board sensors, such as cameras or radars, systems which are particularly useful in several use cases, however they are not effective when there is no line-of-sight connectivity between vehicle and VRU (Scholliers, van Sambeek and Moerman, 2017). Soto et al. (2022) note that there are strong incentives to provide vehicles with communication-based applications to increase the safety of their occupants and other road users. Furthermore, a VRU protection cross-working group of the 5G Automotive Association (5GAA) released a white paper in 2020 where they focused on the following use cases (5GAA Automotive Association, 2020): a) VRU high risk zones: i.e., static or dynamic messages to drivers or vehicles when they enter high risk areas where the presence of many VRUs is likely. 8 b) Interactive communications between VRU and vehicles: i.e., negotiations between the VRU’s device and a vehicle (e.g., a VRU asks the vehicle for permission to cross the road). c) VRU safety messages and algorithms: i.e., the most common use case, where risk assessment is continuously performed by the most appropriate unit (e.g. infrastructure or vehicle) and warnings are issued to avoid collisions. The VRU detection and warning system, implemented in Trikala, belongs to the ‘VRU safety messages and algorithms’ category. It aims to improve the safety of both VRUs and passengers that make use of the automated service. Stationary AI traffic cameras, mounted on traffic poles, have been configured to detect VRUs that cross the road in an uncontrolled way (i.e., without using designated crosswalks or traffic signals). When such an event is detected, a backend process is initiated to prompt the transmission of C-ITS messages from the RSU installed at the crossing. The messages are received by the OBU of the AV, if the vehicle is nearby the intersection, and, subsequently, they are displayed on a tablet device (i.e., the same web application that displays the GLOSA output), which is installed in front of the safety driver to increase their awareness. Fig. 5. Screenshot of the AI-based video analytics detecting uncontrolled crossings on Pylis street in Trikala, Greece. 9 Fig. 6. Screenshot of the web application that displays VRU-warnings to the safety driver (the VRU-warning functionality is integrated in the web application utilised for GLOSA). 2.5 Mobility-as-a-Service (MaaS) application A Mobility-as-a-Service (MaaS) is a relatively new concept, often credited to Sampo Hietanen for introducing it to the Finnish Transportation Ministry in 2006 (AriasMolinares and García-Palomares, 2020). Among the several definitions given in the previous years, Kamargianni and Matyas (2017) defined MaaS as ‘a user-centric, multimodal, sustainable and intelligent mobility management and distribution system, in which a MaaS Provider brings together offerings of multiple mobility service providers (public and private) and provides end-users access to them through a digital interface, allowing them to seamlessly plan and pay for mobility.’ As Arias-Molinares and García-Palomares (2020) point out in their review paper, transport authorities and operators, either private or public, should cooperate to achieve optimal implementation of MaaS, while a data driven approach could prove beneficial for transport planning and policy making. Polydoropoulou et al. (2020) shaped the key elements of the MaaS business eco-system via data collection in Manchester, Luxembourg, and Budapest, and highlighted mobility providers, public transport, and regional authorities as the key stakeholders that should build trust among one another. The gaps regarding standardization and legal framework, as well as the lack of open programming interfaces were stated as barriers of extensive MaaS implementation. Giesecke, Surakka and Hakonen (2016), whilst agreeing that MaaS is the next paradigm change in the transport sector, had already provided concerns regarding its conceptual issues, such as interoperability with ITS, actual sustainability, and end-user acceptance. Later, the study of Butler, Yigitcanlar and Paz (2021) refers to possible lack of cooperation and trust between public and private cooperations, as well as to possible difficulty appealing to elderly people and frequent users of privately-owned vehicles. 16 Scholliers, J., van Sambeek, M. and Moerman, K. (2017) ‘Integration of vulnerable road users in cooperative ITS systems’, European Transport Research Review, 9(2), p. 15. Available at: https://doi.org/10.1007/s12544-017-0230-3. Shen, Y., Zhang, H. and Zhao, J. (2018) ‘Integrating shared autonomous vehicle in public transportation system: A supply-side simulation of the first-mile service in Singapore’, Transportation Research Part A: Policy and Practice, 113, pp. 125–136. Available at: https://doi.org/10.1016/j.tra.2018.04.004. Shi, Y. et al. (2020) ‘A Coordination Algorithm for Signalized Multi-Intersection to Maximize Green Wave Band in V2X Network’, IEEE Access, 8, pp. 213706–213717. Available at: https://doi.org/10.1109/ACCESS.2020.3039263. Soto, I. et al. (2022) ‘A survey on road safety and traffic efficiency vehicular applications based on C-V2X technologies’, Vehicular Communications, 33, p. 100428. Available at: https://doi.org/10.1016/j.vehcom.2021.100428. SWARCO (2024) Green Wave in Traffic: Challenges. Available at: https://www.swarco.com/mobility-future/intelligent-transportation-systems/green-wave-traffic (Accessed: 25 April 2024). Tseng, C.-W., Chang, Y.-J. and Kuo, Y.-W. (2017) ‘Green wave-based virtual traffic light management scheme with VANETs “Green wave-based virtual traffic light management scheme with VANETs”’, Int. J. Ad Hoc and Ubiquitous Computing. Available at: http://ecomove-project.eu. Venkatesh, V. et al. (2003) ‘User Acceptance of Information Technology: Toward a Unified View’, MIS Quarterly, 27(3), pp. 425–478. Available at: https://doi.org/10.2307/30036540. Venkatesh, V., Thong, J.Y.L. and Xu, X. (2012) ‘Consumer Acceptance and Use of Information Technology: Extending the Unified Theory of Acceptance and Use of Technology’, MIS Quarterly, 36(1), pp. 157–178. Available at: https://doi.org/10.2307/41410412. Zheng, Y. et al. (2020) ‘A Novel Approach to Coordinating Green Wave System With Adaptation Evolutionary Strategy’, IEEE Access, 8, pp. 214115–214127. Available at: https://doi.org/10.1109/ACCESS.2020.3037129.