By taking the GEAC edge computing controller (Orin), researched and developed by TZTEK independently, as the AI computing platform by the roadside, with access to cameras, millimeter wave radar, laser radar and other sensors, and through multi-sensor fusion perception algorithm, it is able to provide ITS with accurate, real-time and stable perception data by performing perception calculations in terms of real-time recognition, tracking, positioning, speed, heading, etc., of road traffic participants. Then real-time lane level intelligent perception of high-precision of traffic targets at road sections and intersections can be realized, completing early warning and analysis of traffic incidents, thus providing information input for perception and decision making of vehicle-road collaboration of intelligent connected vehicles. The vehicle terminal uses GEAC edge computing module (Orin) as the AI computing platform, with access to the camera, laser radar and other sensors at the vehicle terminal to identify the objects and obstacles around the vehicle,and transmits the road side sensing data through RSU and OBU link to the vehicle terminal as the other input. Combined with the vehicle-road fusion sensing algorithm, purposes of warning and alerting at the vehicle terminal, ensuring traffic safety and improving traffic efficiency are finally realized.