| Autonomous vehicles,intelligent connected vehicles,and new energy vehicles usually require comprehensive technical tests including safety,functionality and system before driving on the actual road.It is also an important test method for autonomous driving research and automotive development institutions today.At this stage,mixed reality and digital twins are gradually becoming technical hotspots for innovation and development in various fields,and present a very broad prospects for development.Because the technical methods of mixed reality and digital twins are efficient,easy to use,low-cost,and repeatable.Spatial visualization and other characteristics,so this paper proposes a mixed reality vehicle digital twin test environment in view of the various degrees of functional unilateralism,test limitations,and poor operability that are common in several current simulation test methods for automatic driving testing.The construction method is to construct virtual mapping objects and digital twins corresponding to real vehicles,roads and roadside facilities in virtual three-dimensional scenes,and introduce virtual vehicles and pedestrians as traffic participants to achieve a mixed reality system form the test environment specifically includes the following aspects.1.Constructed the vehicles digital twin and test road environment,including the digital three-dimensional modeling,collision detection and spatial coordinate conversion process of the digital twin prototype vehicle in the virtual space,and realized the spatial and geographic information characteristics through GIS spatial correction.Open Street Map road mapping in the virtual environment,and the use of Easy Road3 D to construct the road environment in the threedimensional scene.2.Propose the construction method of the mixed reality system,including the structure of the system and the realization method and process of the function.The real unit includes actual vehicles,roads,and roadside facilities;the virtual unit includes the corresponding digital twins and virtual traffic participants in the virtual space.The communication system includes the external communication process of OBU intelligent vehicle equipment and roadside equipment.And the network transmission process between all data and the host computer.3.The HMM map matching algorithm corrects the position deviation of the vehicle during the driving process,and improves the accuracy of the matching between the vehicle and the map road.The result shows that the positioning deviation caused by the poor GPS satellite signal can be well eliminated,and then solved the problem that the vehicle deviates from the driving road in the virtual scene.4.Based on the LSTM vehicle trajectory prediction algorithm,a multi-step trajectory prediction method oriented to real-time system characteristics is proposed.The experiment verifies that both prediction methods can well eliminate the time delay generated in the communication and network transmission process,and the results show that multi-step trajectory prediction has high accuracy and error control capabilities.The method for building a vehicle digital twin test environment based on mixed reality proposed in this paper can construct a mixed reality test environment well and conduct vehiclevehicle and vehicle-human crash tests.The performance of the constructed mixed reality system has good performance.It can carry out diversified tests under different conditions,and guarantee the reliability of test results. |