Internationales Verkehrswesen (74) 4 | 2022 69 Wissenschaft TECHNOLOGIE Towards Vision Zero V2X Communication for Active Vulnerable Road User Protection V2X Communication, Vulnerable Road User, ITS-G5, C-V2X, LTE Almost half of the casualties on European roads can be accounted to the group of the so-called Vulnerable Road Users (VRUs). The introduction of V2X Communication makes it possible to extend the awareness horizon of automated and autonomous vehicles, in order to avoid accidents with VRUs. In this paper we learn how V2X communication will protect VRUs, what requirements have to be met, what the key performance metrics are and how different communication and localization technologies perform. The Vision Zero goal is to reduce the road traffic casualties including VRUs to zero. Fabian de Ponte Müller, Estefania Munoz Diaz, Stephan Sand, Clarissa Böker, Lukas Merk I n urban road traffic, motorized vehicles coexist and share space with lessor non-motorized road users, as for instance pedestrian and cyclists. Passenger cars and trucks have a high mass and already at low speed carry great amount of kinetic energy that can cause severe damages in an accident. Pedestrians, cyclists and motorbikes are especially exposed, since they are not surrounded by a steel cage that could protect them in case of an impact. Therefore, pedestrians, cyclists and motorbikers are also known as Vulnerable Road Users (VRU). In Europe, according to the Annual Accident Report of the European Commission 22 %, 10 % and 19 % of all road fatalities correspond to pedestrians, cyclists and motorbikers, respectively [1]. Vulnerable Road User Protection Consequently, VRU protection has gained a lot of attention over the last years. Not only the legislator has pushed towards deploying new safety measures to protect VRUs, but also the automotive industry has gradually incorporated new VRU safety features into vehicles. Passive safety systems aim at minimizing the consequences of an accident. Helmets (worn by cyclists and motorbike) drivers are a good example of a passive safety measure. Some motorbikes are equipped with airbag systems that unfold in case of an accident to absorb the energy of the biker’s body hitting the front part of the motorbike. Passenger cars have also experienced changes in their design to minimize the injuries when hitting against pedestrians and cyclists, e. g. the wipers have been hidden behind the hood. Some passenger cars also feature a system that automatically lifts the hood to absorb part of the energy when an impact with a VRU is detected. Active safety systems aim at avoiding the occurrence of accidents. There has been a strong trend in the past two decades in equipping passenger cars with advanced driver assistance systems (ADAS) and safety systems, as for instance, the electronic Stability Program (ESP), pre-collision warning and automatic emergency brakes (AEB). Since 2016, the European New Car Assessment Programme (Euro NCAP) extended the AEB test criteria to include also the detection of pedestrians. Though, car models to obtain the highest 5-star distinction require to be equipped with pedestrian detection systems. Some truck manufacturers have also put a right-turning assistant for detecting a bicycle in the driver’s dead-angle as an eligible safety system and the European Commission is considering making such systems mandatory across Europe. Active safety systems in vehicles rely on a sensor system that detects the presence of a VRU and enables the vehicle to predict a potential hazardous situation. Usually, cameras, radars or laser scanners are used to this end. These systems work well, especially when combined in a sensor fusion approach. However, they have in common that they can only detect VRUs in their unobstructed field of view. With the introduction of V2X communication, automated vehicles are able to exchange information with other road users, the road infrastructure and back-end servers. Hence, the automated vehicle evolves into a connected and cooperative vehicle, which opens new possibilities for protecting vulnerable road users. By incorporating communication capabilities, the limitations of on-board perception sensors of automated vehicles can partly be compensated. This can be accomplished in two different ways: cooperative VRU awareness and infrastructure-side awareness. Both paradigms are further explained next. Cooperative VRU Awareness In this approach, VRUs equipped with a dedicated electronic device can actively make themselves aware to approaching vehicles. Especially the smartphone or wearable devices, such as smart-watches or smart-glasses, make the possibility of exchanging information between vehicles and pedestrians practical. Today’s smartphones incorporate wireless communication technologies, such as Wi-Fi, cellular communication and Bluetooth and are equipped with multiple sensors, like global navigation satellite system (GNSS) receivers and motion sensors, that can help localizing accurately the device. Internationales Verkehrswesen (74) 4 | 2022 70 TECHNOLOGIE Wissenschaft Smartphones feature a high-level of computational power, include a power supply and come along different human-machine interfaces, such as the display and the haptic and acoustic interfaces. The internal software architecture of these devices makes it relatively easy to deploy new software “Apps”, that make use of all these possibilities. The limiting factor for cooperative VRU protection is however still the size and weight of the devices, the prize and the battery consumption. Although, technological advancement will on the long run make these factors less heavy, a solution is required in the short term. In this regard, the strong increase in electric-bikes, pedelecs and e-scooters in the last years makes the option of cooperative VRU protection increasingly interesting. A relatively large power supply and the ability to transport more weight are key advantages to deploy active localization systems and communication technologies into these means of transport. Cooperative VRU Awareness requires two subsystems: high accuracy localization and reliable, lowlatency communication. VRU Localization For collision avoidance purposes it is necessary to pinpoint all road users on a common map. VRUs can compute their position using the GNSS-receiver embedded in their smartphone. However, a high positioning accuracy is required for collision avoidance purposes and it is difficult to achieve using only GNSS, especially in urban environments. Particularly, a position uncertainty (1σ) of less than half a meter is required to obtain a high detection rate keeping the false positive rate under 10 %. Figure 1 shows the receiver operating curves (ROC) in dependence of the vehicle and VRU position uncertainty at the collision instant. The position accuracy deduced with the information provided by the GNSS receiver embedded in commercial smartphones can be increased using the inertial sensors available in the smartphone as well or maps, among others. Our study shows that the position accuracy reached by GNSS aided by inertial sensors is still not enough for reaching the half a meter accuracy required for collision avoidance purposes. The blue curve in Figure 2-left shows our combined GNSS and inertial sensor solution with a 2D position accuracy of 2,7 m (1σ), while the red curve uses only the GNSS position provided by the smartphone. We have used the smartphone Samsung Galaxy S20 in suburban environment. Likewise, Figure 2-right shows the positioning results for bicycles. The blue curve represents the position accuracy for the combined GNSS and inertial sensors solution, while the red curve represents the GNSS-only position accuracy. For these experiments the Figure 1: Collision detection performance for three different positioning uncertainties (σ = [1m, 0.5m, 0.2m]) All figures by the authors unless otherwise noted Figure 2: CDF positioning error curves for pedestrian localization (left) and bicycle localization (right) using a smartphone in a suburban environment Internationales Verkehrswesen (74) 4 | 2022 71 Wissenschaft TECHNOLOGIE same smartphone was attached to the handlebar of the bicycle in semi-urban environment. For VRU protection and robust collision avoidance, not only the position is important, but also speed and direction of movement are needed. Since these parameters are not known at infinite precision, also a measure of the accuracy is required. Further, since the aim is to avoid collision that are about to happen, reliable prediction of the movement is required. For a speed of around 40- km/ h, 1.5 seconds are required to get a vehicle to standstill. Hence, this is also the prediction time for the trajectory of the vehicle and the VRU. Whereas vehicle’s future trajectory might be well predicted in this timeframe due to its high mass, it will be more challenging for the case of a bicycle or even a pedestrian. Here dedicated movement and intention models for pedestrians and cyclists are the key to success. V2X Communication Mainly two communication technologies are regarded for exchanging information between VRUs and vehicles: ad-hoc wireless communication or cellular communication. Ad-hoc communication is a broadcast communication without centralized coordination and link establishment. Example of ad-hoc wireless communication are ITS-G5 (or DSRC), which are based on IEEE 802.11p/ bd, or Cellular-V2X over the LTE-PC5 interface. In principle, also traditional cellular communication over a base-station (Uu interface in 3GPP-LTE nomenclature) can be used to transmit the information from VRUs to vehicles. This, however, requires a central server to disseminate and geographically filter the information amongst nodes. The relevant performance metrics for the communication are the update rate, the end-to-end latency, the packet error rate (PER) and the communication range. The highest update rate foreseen for V2X is 10 Hz, although an adjusted update rate in dependence of the node dynamics is used. The end-to-end latency for the ad-hoc communication technologies is usually below 50 ms, as our link-level tests with ITS-G5 and C-V2X on an apron showed [2] (see Figure 3-left). The maximum communication range is dependent on the transmit power and possible obstructions and blockage due to vehicles and buildings. Here we can expect slightly better performance for C-V2X than for ITS-G5 due to the physical layer with improved coding mechanisms and hybrid automatic repeat request (HARQ). Our direct performance tests for a V2V link showed that CV2X had indeed a better performance in terms of a lower PER compared to ITS-G5 when driving around buildings in an in an urban environment (see Figure 3-right). However, packet errors will occur not only due to low signal strength but also due to mutual interference with other nodes. Here, network simulations with different technologies will answer, which technology is mostly suited for crowded urban environments. To keep mutual interference at an adequate level, one approach is to lower the update rate or the TX power. These techniques are as well in the interest of a thrifty power consumption at the VRU’s wearable device. A vehicle driving at 50 km/ h (14 m/ s) will travel 22 m in 1.5 seconds. Hence, a 30 to 50 m transmission range seems appropriately to warn an approaching vehicle on time. We tested a bicycle-vehicle scenario in an urban canyon using ITS-G5 communication and using a TX power of 23dBm. Figure 4 shows the RX power, the range and the PER over time. It can be seen how 80 to 100 m range were achieved with a PER of less than 50 % with this setup. By extrapolation and visual inspection, it can be stated that even with a TX power between 0 to 5 dBm a coverage up to 50 m could be achieved. The line between localization and communication vanishes, when communication technologies are used for localization. This is possible with technologies as for instance IEEE 802.11bd Next Generation V2X, IEEE 802.15.4 Ultrawide Band (UWB) and IEEE 802.11az. The latter two use large radiofrequency bandwidths to per- Figure 3 left: E2E-delay for V2V link using ITS-G5 (blue) and C-V2X (red) for different message lengths. Right: PER for a V2V link in urban environment using ITS-G5 (blue) and C-V2X (red) Figure 4: Pedestrian-vehicle collision scenario in urban canyon. Distance between a pedestrian and a vehicle, PER for a direct ITS-G5 communication link and Rx power at the vehicle over time. Internationales Verkehrswesen (74) 4 | 2022 72 TECHNOLOGIE Wissenschaft form precise round-trip delay measurements and estimate the distance between a moving node and several fixed anchor nodes. Out of several of these ranging measurements the position can be estimated [2]. Infrastructure-side awareness Cooperative VRU Awareness requires the VRUs to carry a dedicated electronic device and, thus, creating a hurdle to a broad system deployment and to a measurable impact to accident statistics. Infrastructure-side awareness represents a paradigm shift in VRU protection. It is not the VRU who actively makes other vehicles aware of its presence, but the road infrastructure. By using a suited perception system, as for instance a camera installed on a nearby pole, road users, including VRUs, are detected, localized and tracked. This information can be encoded into a Collaborative Perception Message (CPM) and transmitted over V2X communication. This approach has been tested at the test field for cooperative and connected driving in Düsseldorf in the frame of the KoMoDnext project. A camera system was placed above a pedestrian crossing and was able to detect crossing pedestrians and generating CPMs, which were broadcasted over ITS-G5, C-V2X and LTE [4]. Not only cameras are suited to detect VRUs at urban intersections. In the frame of the German-funded project VIDETEC, DLR and IMST have tested radio-based detection systems for VRU protection [5]. IMST develops 24 GHz and 77 GHz radar chips that enable anonymous road user classification by processing micro-Doppler signatures with the help of machine learning techniques. In this context, we evaluated a very promising technology based on processing radio signals termed Joint Communication and Sensing, which is foreseen to be a key technology for the future 6th Generation cellular communication. Every node of a distributed array of antennas around the intersection transmits periodic beacons that are received by all other nodes. By intelligent signal processing of the distorted incoming signal, the location and speed of road objects can be obtained. Independent of the perception means, the information about the presence, the movement, the direction and the type of VRU needs to be delivered into the automated vehicle to be fused with the on-board perception sensors, incorporated into its local dynamic map and considered by the vehicle’s automation [6]. To this end, three competing technologies for V2X have been tested in the Düsseldorf field test. At a roadside intersection three radio devices for ITS-G5, C-V2X and LTE were placed on a pole at a height of four meters. The coverage of all three technologies was tested by driving around the south-east block. It could be observed, that CPMs were received over the direct links before having visual connection to the intersection. Figure 5-left shows that a minimum coverage of 114 m could be measured. This represents more than 8 seconds when driving at a maximum speed of 50 km/ h and seems to be sufficient for warning the vehicle about crossing pedestrians. With LTE the highest coverage was obtained. However, a centralized approach will need to decide which information to forward to which vehicle according to distance or route. On the right Figure 5, the E2E delay for all three technologies can be compared to each other. Again, C- V2X has slightly higher delay than ITS-G5, while LTE doubles the latency reaching values above 350ms. VRUs, unlike vehicles, are not restricted to move on roads and can very suddenly change their curse. For sudden events, as for instance a pedestrian making a “last-second” turn onto the street, this timeframe might be already too large to timely react. Especially, if further time for processing and actuation has to be considered. Towards Vision Zero The Vision Zero goal is to reduce the road traffic casualties to zero. V2X enabled connectivity and cooperation will support the highly automated vehicle to drive more efficient in urban environments, while maintaining a high level of safety. We have presented in this article that ad-hoc V2X communication, either through ITS-G5 or C-V2X, is able to convey information about the presence of VRUs in less than 100 ms making it possible for a vehicle to safely maneuver or come to standstill. Therefore, the V2X communication has a great potential to contribute to the Vision Zero. However, in real environments two key aspects of V2X communication have to be guaranteed, namely the robustness of the communication against mutual interference in highly congested environments and semantic “trustworthiness”. These two aspects are key to the vehicle automation that is necessary to avoid hazardous situations between vehicles and VRUs. Furthermore, we have discussed that the VRUs can contribute to their own safety by sharing their position computed e.g. with their own smartphone. However, the Road- Side Unit Figure 5 Left: Coverage of CPM transmissions from RSU at a vehicle during in Düsseldorf. Blue with ITS-G5, Red with C-V2X and Yellow with LTE. Right: E2E delay with all three technologies. Satellite picture provided by Google Earth. Internationales Verkehrswesen (74) 4 | 2022 73 Wissenschaft TECHNOLOGIE position accuracy required for this safety critical application is difficult to meet in urban environments with the current technology. Last but not least, we have seen that RF-based perception system, as for instance Joint Communication and Sensing, are promising candidates to support and enhance camera-based systems for infrastructure based VRU perception. ■ REFERENCES [1] European Commission (2020): Annual statistical report on road safety in the EU 2021, Directorate General for Transport, May 2022 [2] De Ponte Müller, F.; Rashdan, I.; Schmidhammer, M.; Sand, S. (2021): ITS-G5 and C-V2X Link Level Performance Measurements. ITS World Congress 2021, 10.-11. Okt. 2021, Hamburg, Deutschland. [3] Ponte Müller, F.; Munoz Diaz, E.; Perul, J.; Renaudin, V.(2020): Urban Vulnerable Road User Localization using GNSS, Inertial Sensors and Ultra-Wideband Ranging. In: 31st IEEE Intelligent Vehicles Symposium, IV 2020. 2020 IEEE Intelligent Vehicles Symposium, 20.- 23. Okt. 2020, Las Vegas, USA. ISBN 978-1-7281-6673-5. ISSN 2642-7214. [4] Böker, C.; Leich, A.; Andert, F.; Steuernagel, S. (2021): Investigations on Collective Perception for Left Turning Manoeuvres. ITS World Congress 2021, 11.-15. Sept. 2021, Hamburg, Deutschland. [5] Kulke, R.; Hägelen, M.; Jetten, R.; Schmidhammer, M.; De Ponte Müller, F.; Rashdan, I. (2021): Increased traffic safety by means of intelligent detection and localization technologies. EuRAD 2021 18 th European Radar Conference, 16.-18. Okt. 2021, London, UK. [6] Lapoehn, S.; Heß, D.; Böker, C.; Böhme, H.; Schindler, J. (2021): Infrastructure-aided Automated Driving in Highly Dynamic Urban Environments. ITS World Congress 2021, 11.-15. Sept. 2021, Hamburg, Deutschland. Clarissa Böker Institute of Transportation Systems, German Aerospace Center (DLR), Braunschweig clarissa.boeker@dlr.de Lukas Merk Institute of Transportation Systems, German Aerospace Center (DLR), Braunschweig lukas.merk@dlr.de Stephan Sand, Dr. Team Leader Vehicular Applications, Institute of Communications and Navigation, German Aerospace Center (DLR), Weßling stephan.sand@dlr.de Estefania Munoz Diaz, Dr. Team Leader Multimodal Personal Navigation, Institute of Communications and Navigation, German Aerospace Center (DLR), Weßling estefania.munoz@dlr.de Fabian de Ponte Müller, Dr. Senior Researcher, Institute of Communications and Navigation, German Aerospace Center (DLR), Weßling fabian.pontemueller@dlr.de FACING THE CHALLENGES OF MOBILITY Founded in 1949 - bound forward to face the challenges of tomorrow‘s mobility: With an editorial board of renowned scientists and an advisory board of directors, CEOs and managers from all transport industry areas, »Internationales Verkehrswesen« and »International Transportation« - the worldwide distributed English-language edition - rank as leading cross-system transport journals in Europe for both academic research and practical application. Rail and road, air transport and waterway traffic — »International Transportation« and »Internationales Verkehrswesen« stimulate a worldwide interdisciplinary discussion of the numerous defiances in mobility, transport, and logistics. The magazines are targeted at planners and decision makers in municipalities, communities, public authorities and transportation companies, at engineers, scientists and students. With peer-reviewed scientific articles and technical contributions the magazines keep readers abreast of background conditions, current trends and future prospects - such as digitalization, automation, and the increasing challenges of urban traffic. Read more about the magazines and the subscription conditions: www.internationales-verkehrswesen.de www.international-transportation.com INTERNATIONALES VERKEHRSWESEN AND INTERNATIONAL TR ANSPORTATION »Internationales Verkehrswesen« and »International Transportation« are published by Trialog Publishers Verlagsgesellschaft, D-Baiersbronn IV_Image_halb_quer.indd 1 IV_Image_halb_quer.indd 1 15.04.2018 19: 50: 55 15.04.2018 19: 50: 55