Collision avoidance control with combined steering and braking based on driving risk assessment

To improve the active safety obstacle avoidance capability of intelligent vehicles in dynamic traffic environments, this study proposes a collision avoidance control framework with combined steering and braking based on driving risk assessment. First, the motivations for lane changing and braking during obstacle avoidance are analyzed, and a safety distance zone model considering driver characteristics is established to determine the decision boundaries for warning, emergency braking, and lane-changing maneuvers. Then, a predictive collision risk assessment method is developed by integrating the motion states and relative position relationships of the host vehicle and surrounding traffic vehicles, enabling the vehicle to evaluate potential collision risks within the prediction horizon. On this basis, an active safety obstacle avoidance controller is designed using model predictive control, in which steering and longitudinal braking actions are coordinated through a deviation prediction model to optimize the front wheel angle and longitudinal acceleration simultaneously. In addition, a quintic polynomial-based path planning method is adopted to generate feasible obstacle avoidance trajectories under safety and stability constraints. The proposed strategy is validated through CarSim/Simulink co-simulation and real-vehicle experiments. The results demonstrate that the proposed method can adaptively adjust vehicle control actions according to the assessed driving risk, effectively accomplish obstacle avoidance in complex scenarios, and enhance driving safety and control stability.