Implementation Of Virtual Collaboration And Augmented Reality Training System For Electric Vehicles Based On Unity3D

With the gradual maturity of computer,image forming,Internet,mobile touch-screen terminal and other technologies,virtual reality and augmented reality technology have developed rapidly,and have been successfully applied to teaching,sports,entertainment,medical,construction and other fields.The traditional electric vehicle training power system is a high-pressure component,there is high-pressure danger in the training process,remote collaborative training cannot be realized,the training site and time are limited,and the training cost is high.To solve these problems,the Unity3 D is taken as a development platform,and virtual collaborative and augmented reality training systems for electric vehicles are designed and developed respectively by using virtual reality and augmented reality technologies.Among them,the virtual collaborative training system is used in the PC platform,which realizes system functions such as user scene roaming,learning mode,assessment mode,multi-person collaborative work,individual independent work,data management,etc.Augmented reality training system is used in touch-screen mobile terminals and realizes the functions of recognizing part structure through rotation and scaling and training through disassembly and assembly.The key technologies of system implementation are studied as follows:The virtual collaborative training mechanism was studied.This paper studies the communication mechanism between the client and the server based on the TCP protocol,and proposes to solve the problem of operation conflict by using the lock mode of component operation authority setting.The distributed collision detection technology was studied.In view of the large number,different shapes and complex structure of virtual collaborative training system models,a distributed collision detection algorithm is proposed,in which OBB bounding box is added to the object client for preliminary collision detection,while Mesh Collider is added to the server for more accurate collision detection,which solves the conflict between the two constraints of accuracy and real-time collision detection of bounding box in large virtual training scenes.Aiming at the limited performance of touch screen mobile terminals in augmented realitytraining system,the collision detection algorithm with OBB bounding box is proposed to improve the accuracy and real-time of collision detection.The model optimization technology was studied.In view of the large number of mesh patches in the 3D model of the virtual scene and the limited real-time rendering capability of the system,the dynamic LOD technology was adopted to reduce the number of rendering patches,thus improving the real-time rendering performance of the system.As the number of virtual scene rendering models leads to system sticking,cone volume culling technology was used to cull out-of-view models and occlusion culling technology was used to cull occluded models to reduce the number of system rendering models so as to improve the system rendering speed.