Vulnerable Road User Intention Detection for Urban Locations (VRUIDFUL)

VRUIDFUL项目旨在将重点明确地放在扩大OTH Amberg-weden之前在移动性领域的研究工作上，以包括智能基础设施(例如，监控传感器、先进的红绿灯系统)。具体地说，将分几个阶段研究它们在真实城市试验场的设置以及它们与脆弱道路使用者(VRU)集成的安全系统：在本应用范围内，安贝里脆弱道路使用者(VRU)特别频繁的区域(大学、KITA和幼儿园周围)将配备智能基础设施单元(IISU)。这些将被用来收集和评估广泛的数据。IISU的传感器技术基于雷达、激光雷达、热成像和立体摄像机，各单元的智能旨在对目标进行检测和分类，特别是VRU和VRU组。获取的抽象数据(3D立方体)通过移动无线电传输到中央服务器，ITS-G5(部分存储LORA)，可用于进一步分析。由于只将抽象的数据传输到服务器并存储，因此确保了数据保护的合规性。同时，试验区选定的红绿灯系统将配备V2X通信单元并连接到交通控制计算机，以提前发出红绿灯相位。这些信息将进行时间同步并存储在服务器上。其他数据，如天气状况，也被记录下来，以便有可能对红绿灯切换和天气状况对VRUs行为的影响做出有充分依据的评估。基于研究数据，将在第二阶段与合作伙伴(例如来自巴伐利亚人工智能移动网络AIMotion)一起启动关于脆弱道路用户及其行为的意图识别的研究项目。研究结果将被用来为车辆制定警告策略和建议行动，以便及早发现潜在危险的情况(例如行人或骑自行车的人希望在较长的“绿灯”阶段过后“快速”过马路，而行人灯变成“红灯”，或行人疏忽地跟随一群其他VRU)，从而避免事故发生。要评估这些努力和好处。在第三阶段，这些警告策略和建议的行动应该与相关的利益相关者讨论，即交通规划者、车辆制造商、市民等。例如，最初，智能手机应用程序可以帮助提高行人意识。然而，总体目标是在与车辆制造商和交通规划者的联合项目中实现综合系统，即不仅将行人纳入警告概念，还将车辆纳入警告概念。

The VRUIDFUL project is intended to place a clear focus on expanding the previous research work of OTH Amberg-Weiden in the field of mobility to include intelligent infrastructure (e.g., monitoring sensors, advanced traffic light systems). Specifically, their setup in a real urban test field and their integration into an integral safety system for vulnerable road users (VRU) will be researched in several phases: Within the scope of this application, an area in Amberg that is particularly heavily frequented by vulnerable road users (VRU) (around the university, KiTa and kindergarten) is to be equipped with intelligent infrastructure units (IISU). These will be used to collect and evaluate extensive data. The sensor technology in the IISU is based on radar, lidar, thermal imaging and stereo camera, the intelligence in the units aims at the detection and classification of objects, especially VRU and groups of VRU.The obtained abstracted data (3D cubes) are transmitted to a central server by means of mobile radio, ITS-G5 (and partly LoRA), stored and can be used for further analysis. Since only abstracted data is transmitted to the server and stored, data protection compliance is ensured.In parallel, selected traffic light systems in the test area will be equipped with V2x communication units and a connection to the traffic control computer to send out traffic light phases in advance. This information will be time-synchronized and stored on the server. Other data, such as weather conditions, are also recorded so that it is possible to make a well-founded evaluation of the influences of traffic light switching and weather conditions on the behavior of VRUs.Based on the research data, research projects on intention recognition of vulnerable road users and their behavior are to be initiated together with partners (e.g. from the Bavarian AI mobility network "AImotion") in a second phase. The resulting findings are to be used to derive warning strategies and recommended actions for vehicles in order to detect potentially dangerous scenarios (such as pedestrians or cyclists who want to "quickly" cross the street after a longer "green" phase when the pedestrian light turns "red", or pedestrians are inattentively following a crowd of other VRUs) at an early stage and avoid accidents. Effort and benefit are to be evaluated.In a third phase, these warning strategies and recommended actions should be discussed with the stakeholders involved, i.e., traffic planners, vehicle manufacturers, citizens, etc. Initially, a smartphone app, for example, could help increase pedestrian awareness. However, the overall goal would be to make the integral system a reality in joint projects with vehicle manufacturers and traffic planners, i.e., to integrate not only pedestrians but also vehicles into the warning concept.