Intelligent Car Autonomous Driving Tracking Technology Based on Fuzzy Information and Multi-sensor Fusion

Abstract

Car autonomous driving technology is a key development direction in the future, but road environment data is relatively complex, and single sensor driven autonomous driving algorithms have the problem of insufficient trajectory prediction accuracy. In response to this, this study uses fuzzy information to construct a fuzzy neural programming network for data processing, and uses a fuzzy sample dataset for training. The multi-sensor autonomous driving tracking algorithm is used to fuse and calculate the data collected by the sensors. The data is transformed into a Cartesian coordinate system to calculate the predicted lateral and longitudinal distance, velocity, and acceleration of the tracking algorithm. The results demonstrated that the distance and speed predicted by the driving tracking algorithm were basically consistent with the true values, and the longitudinal distance was 0.34 m and 0.28 m higher than the square root Kalman filtering algorithm and the extended Kalman filtering algorithm, respectively. The longitudinal velocity denoising accuracy of the algorithm has been improved by 18.65% and 31.27% compared to other algorithms. Therefore, the autonomous driving tracking algorithm was better able to maintain a safe distance than other algorithms, and its acceleration changes were smoother, improving the ride comfort and safety of the vehicle. The tracking algorithm has a stronger ability to remove noise from sensors and can adapt to more diverse environments. The designed fuzzy information and multi-sensor fusion tracking algorithm for car autonomous driving provides a reference for subsequent research on automotive autonomous driving technology.