Human–Machine Cooperative Steering Control Based on Non-cooperative Nash Game

To address the conflict of inconsistent driving intentions between the driver and driving assistance system in cooperative steering, a human–machine cooperative steering control framework based on non-cooperative Nash game is proposed. The competitive relationship between the driver and driving assistance system with different control objectives is modeled as a non-cooperative Nash game. The cooperative steering under human–machine Nash game condition is transformed into a model predictive control optimization problem to achieve coordination through Nash equilibrium. A flexible driving authority allocation strategy based on time-to-collision is then proposed for vehicle lateral control, with dynamic driving authority weights fed into the human–machine cooperative steering model in real time. The effectiveness of the non-cooperative Nash game human–machine cooperative steering strategy and flexible driving authority allocation strategy is validated through simulations and hardware-in-the-loop experiments. Results show the proposed strategy can improve vehicle safety while retaining certain driver freedom when human–machine control targets conflict, by flexibly transferring authority from the driver to the assistance system as collision risk increases.