Optimal Control-Based Approach for Extreme Emergency Obstacle Avoidance in Autonomous Vehicles

The obstacle avoidance challenge for autonomous vehicles is typically tackled using a hierarchical control structure that involves path planning and tracking. However, existing scenarios often lack the urgency required for true emergency conditions, leaving the vehicle’s obstacle avoidance capabilities largely unexplored. This paper addresses the challenge of extreme emergency obstacle avoidance for autonomous vehicles by formulating it as an optimal control problem. The vehicle is modeled as a particle, and the control inputs are chosen to be the direction of acceleration, simplifying the problem significantly. The optimal control problem is thoroughly analyzed and solved analytically, providing real-time insights into the vehicle’s obstacle avoidance capability. The resulting optimal direction of acceleration, is utilized in constructing a controller that employs a PI controller. The resolution of longitudinal and lateral slips, necessary for achieving maximum tire force in the desired direction, is performed offline through an optimal problem based on the magic formula. These slips are realized using a 4WS (four-wheel steering) and brake systems. Simulations demonstrate superior obstacle avoidance distances compared to a commonly adopted hierarchical method at both medium and high speeds. The g-g diagram illustrates the proposed method’s effective utilization of tire forces, keeping them close to the optimal direction.