Method and process for co-simulation with virtual testing of real environments with pedestrian interaction

摘要

Virtual testing of interaction between vehicles and road users (pedestrians) comprising creating virtual reality (VR) environment, based on a real world environment or a new virtual environment, with vehicles and road users (pedestrians), sending details about VR scene to a test person and a vehicle, measuring actions of test person in VR environment using a motion capture system, and predicting pedestrian movement and predicting vehicle control actuator signals based on test person actions that are sent to a visualisation processing unit. The invention may test the safety algorithms of autonomous vehicles to improve collision avoidance software between driverless cars and pedestrians. Evasive manoeuvre secure motion planning and uncertainty quantification may be used and the method may be operated in a non-sequential multi-threading manner. The test person may be immersed in a virtual reality cage with VR glasses or a virtual reality cave. The movement, gestures and actions of the user may be captured by the MoCap system and analysed in different real-world traffic situations in VR environments based on 3D models of real cities and with daily typical road behaviour and realistic human behaviour. Outdoor virtual testing with augmented reality (AR) glasses in real city environments is also disclosed.