Multimodal Fusion Lateral Stability Control Strategy for Autonomous Vehicles

Abstract

In recent years, with the rapid development of autonomous driving technology, vehicle lateral stability control has that put higher requirements. Although traditional control strategies usually depend on monocular information, resulting fewer robustness and adaptability in driving with complex situations. From the above-mentioned challenges, this paper develops a multimodal fusion lateral stability control strategy for autonomous vehicles. The framework combines various data from vision, radar to vehicle dynamics in order to generate a complete perception of the surrounding environment and state of the vehicle. The proposed method utilizes a gradual fusion of heterogeneous information via patient selection based on closeness to the instability state in its designed fusion algorithm for more accurate identification of risk resilience. A beacon-based coordinated control module fuses the decision-making information and implements steering as well as braking interventions in real-time to preserve longitudinal and lateral stability. Experimental validations reveal that this control strategy improves the vehicle tracking performance against lateral disturbances by a significant margin, thereby achieving a more precise trajectory tracking and better handling robustness on different road conditions. This research offers a realistic and efficient approach to the lateral stability control of autopilots, which can promote driving safety for future increasing autonomous vehicles and autonomic system reliability. We will focus on increasing the computational efficiency of the fusion algorithm in future works, as well as broadening traffic scenarios to more extreme and unpredictable merging environment.