Application of vehicle trajectory big data analysis based on DPCS algorithm

The collection and analysis of vehicle trajectory data has become increasingly important not only to provide real-time information on traffic flow and patterns, but also because it has significant implications for optimising traffic management, reducing congestion and improving road safety. This study proposes a vehicle trajectory big data analysis method based on density peak clustering algorithm and conducts simulation experiments to verify it. The results showed that the dual population cuckoo search algorithm performed well in optimising the Rasterikin, Rosenbrock and Ackley functions, finding near optimal solutions of -1.8560, 0.0001 and 0.0002 within 55, 105 and 85 iterations, respectively. In the 20 to 100 dimensional space, the algorithm converged quickly, the solution was stable, the average solution was about 123.54 and the standard deviation was 2.36. In the NuScenes data set experiment, the DPCS algorithm maintained a stable accuracy of 0.9 to 1.0 after 12 rounds of training, outperforming the comparison algorithms and demonstrating its reliability in object detection. The analysis method is of great significance for improving the operational efficiency and safety of the transportation system.