Hash partitioning and spatial compensation in perception fusion algorithm for achieving cooperative vehicle-road interaction

With the rapid development of information technology, intelligent transportation technology is widely applied in the transportation. To enhance the environmental information perception ability of the vehicle road collaborative system, the perception fusion of the vehicle road collaborative system is the research core. A perception fusion algorithm for vehicle road collaborative systems is designed based on spatial position compensation and hash partitioning technology. The experimental results showed that the designed perceptual fusion algorithm performed well in accuracy and recall. The accuracy value was 90%, and the recall rate was 75.47%. The average absolute value error and root mean square error indicators converged to 0.547 and 0.452 respectively, with a faster convergence speed. The accuracy of multi target tracking ultimately converged to 90.28% as the number of iterations increased. The real-time tracking performance fluctuated less with the increase of sample size, ranging from 80.00%-90.00%. In the perception effect simulation experiment, the designed perception algorithm was subjected to secondary hash partitioning, and the fusion rate was significantly improved. Compared to the Detect2Rank method, the fusion time significantly decreased. The perception results obtained through spatial position compensation were more accurate. The value of the intersection ratio significantly increased, and the increase had more than doubled. This study contributes to improving the efficiency, safety, and sustainable development of transportation systems.