Connected Vehicle : Virtual validation of vehicle’s cooperative perception

Autonomous and connected vehicle or also could be termed as CASE (Connected cars, Autonomous / Automated driving, Shared and Electric vehicle) is becoming quite normal in academic ecosystem. The vision behind this terminology is to make mobility safer, much more reliable, and emission-free. With the advancement in Artificial Intelligence capabilities and communications technology, the trends are moving at a higher pace. Where 5G, C-V2X are Key harnessing pillar's of AI capabilities with cloud computing in autonomous and connected driving. The V2X in it's all version (V2V, V2I, V2N..etc) could be seen as the third eye of the vehicle, which give access to the information of other vehicles and their surroundings that are not possible otherwise. This thesis work is part of an EU project called AI@EDGE, from which one of the test cases is being considered for developing this thesis work. The report first discusses the key technological enabler's of AI@EDGE platform based on AI exploitation for various connected vehicle services and functions. The use case related to 'virtual validation of vehicle cooperative perception' would be the main focus of this project work. Given the complexity and costs to support many vehicles in the real world, the use case adopts 5G emulation and virtual traffic simulation environments. Considering different possible specific traffic scenarios, the roundabout is a particularly challenging situation where fluidity and safety are important and will be implemented under the Use case. To test it in a virtual environment a traffic simulator (SUMO) is being used and the roundabout scenario is built. With the help of TraCI library of SUMO the simulation data are being subscribed and communicated to the edge server. For the purpose of sending and receiving CAM/DENM messages a AMQP client is also integrated with the TraCI based program. To deploy the program along with AMQP client on a TBM, on-board device, the communication specific testing is carried out on a 5G emulation bench. Various test cases such as cell handover, dual-connectivity, carrier aggregation, and MIMO etc; related to cellular network LTE/5G are being performed on a network based emulation HIL environment. In addition, the network throughput testing is also carried out for the DUT (Device Under Test), which is equipped with C-V2X wave stack. Finally, the feasibility of a virtual validation platform for vehicular perception is being assessed based on the developed capabilities and results from the HIL testing.