Real-time Obstacle Avoidance Control and Path Planning with Verification for Autonomous Vehicles Using Sensor Measurements

Abstract

In recent years, autonomous technology has advanced rapidly, with the real-time obstacle avoidance capability of autonomous vehicles on the road being a primary concern for ensuring safety. Therefore, we propose an innovative path planning and control methodology to enhance the real-time obstacle avoidance performance of autonomous vehicles. In terms of path planning, we propose a global path planning approach based on an optimized A* algorithm, incorporating a dynamic weighting method and cubic-spline curve smoothing. This improves path search efficiency by 20% and enhances path smoothness by 3%. Additionally, we introduce a local path planning method based on fifth-degree polynomial interpolation curves. Regarding the control approach, we design a position-velocity double-layer proportional-integral-derivative controller for longitudinal control and a feedforward linear quadratic regulator controller for lateral control. Finally, we validate the proposed methodology through control system design, driving path simulation analysis, and real-world vehicle tests using sensor measurements. The application of this innovative control method in a campus environment demonstrates a clear improvement in the real-time driving capability of autonomous vehicles, increasing the obstacle avoidance success rate by 5%. Furthermore, it enables autonomous vehicles to achieve precise control and navigation on the road, reducing the vehicle’s lateral control error and shortening the longitudinal control response time by 0.1 s. These findings provide a valuable technical reference for path planning and control in autonomous vehicle systems.