In response to the decline in vehicle stability and the resulting safety risks caused by inappropriate driver operations during high-speed emergency obstacle avoidance, a human–machine cooperative control strategy based on driver operation recognition is proposed. The strategy establishes a vehicle controllability boundary by integrating real-time driver inputs with tire adhesion limits, enabling dynamic evaluation of the influence of operations on system controllability and identification of potential inappropriate operations. On this basis, a control authority allocation mechanism is developed, capable of adaptively adjusting to vehicle states and driver operations. By combining road boundary constraints with vehicle stability envelope constraints, the strategy dynamically regulates the steering angle, ensuring vehicle stability while retaining the driver’s effective intentions as much as possible. Unlike conventional path-tracking or single-envelope control approaches, the proposed method achieves early identification and proactive mitigation of instability risks induced by inappropriate driver operations, thereby reducing associated safety hazards. To validate the effectiveness of the strategy, two representative scenarios, double lane change and curve avoidance, were designed. Simulation and driver-in-the-loop experiments demonstrate superior performance in terms of vehicle stability, human–machine cooperation, and safety, achieving a higher level of coordinated control and performance balance. The findings provide new insights into the design of human–machine cooperative control strategies under extreme conditions, contributing to enhanced fault tolerance of intelligent driving systems against inappropriate driver operations and improved driving safety.
Liu et al. (Mon,) studied this question.