Obstacle Avoidance Path Planning Algorithm Based on Model Predictive Control

摘要

In recent years, as image processing and control technology have been studied extensively, autonomous vehicle becomes an active research area. For autonomous driving, it is essential to generate a safe obstacle avoidance path considering the surrounding environments. In this paper, an algorithm based on real-time output constraints model predictive control (RCMPC) is devised for obstacle avoidance path planning in the high-speed driving situations. Four simulations were conducted to compare with the normal model predictive control (NMPC) algorithm. The MPC computation times were also compared to verify robustness of the algorithm in the high-speed driving situations. The ISO 2631-1 comfort level standard was used to quantify driver's comfort and to compare with the results. The results of the RCMPC resulted in faster computation times than that of the NMPC and showed a high comfort level scores.