Design of Intelligent Driving Control Model Combining BP-PID and Deep Attention Mechanism

XiaoJiang QIN

Abstract

With the acceleration of urbanization, intelligent driving technology has become a key approach for addressing traffic congestion and the increasing number of traffic accidents. It is therefore necessary to explore this technology in depth. However, the existing intelligent driving control methods are still facing limitations with regard to control accuracy, real-time performance, and safety. In order to address these issues, this study proposes an intelligent driving control model that integrates a Convolutional Block Attention Module, a Back Propagation Neural Network, Proportional-Integral-Derivative Control, and multiple optimization algorithms. The experimental results showed that the proposed model achieved a trajectory tracking deviation of 0.18 m, an average steering angle change rate of 3.13°/s during uphill driving, and a slope speed error of 0.46 km/h. Also, the acceleration and deceleration lag time and response time were 1.23 s and 7.2 s, respectively. In addition, the braking trigger delay during cornering was 28 ms, the acceleration fluctuation rate during lane-changing on straight roads was 0.87 m/s3, and the obstacle-avoidance path offset and lateral acceleration were 0.536 m/s2 and 0.335 m/s2, respectively. All these results show that the proposed model outperforms the other employed models, indicating that it effectively improves the overall intelligent driving control performance and provides a valuable reference for the related technologies.

Keywords

Back propagation neural network, Proportional-integral-derivative control, Attention mechanism, Intelligent driving, motion control, Model optimization.

View full article