基于前车失稳工况下紧急避撞控制策略研究OA

To avoid collisions for self-driving vehicles with a front vehicle experiencing sideslip in a non-connected environment,this paper proposes a control framework integrating state identification,trajectory prediction,trajectory planning and tracking.First,a sideslip identification strategy is developed based on lane departure features and centrifugal acceleration extremes.Then,the constant rate of rotation and acceleration(CTRA)model is employed to predict the trajectory of sideslip vehicles.The unscented Kalman filter(UKF)is used to deal with the model uncertainty and perceptual noise.An estimation model is built covering the possible future positions of the preceding vehicles.Next,a multi-objective optimization function is developed by segmented parameterized driving process.The constrained nonlinear optimization solver is used to generate the collision avoidance trajectory with comprehensive consideration of safety,comfort and driving recovery ability.Finally,the model predictive controller(MPC)is designed based on the three-degree-of-freedom dynamics model.The trajectory is accurately tracked through rolling optimization.Simulation experiments show the proposed method effectively identifies sideslips and successfully produces collision-avoidance trajectories.