Abstract: To address the trajectory deviation and stability decline of autonomous vehicles in complex scenarios due to the coupling of lateral and longitudinal motions, this paper proposes a laterallongitudinal cooperative control algorithm based on dual proportional-integral-derivative (PID) and model predictive control (MPC). By establishing the kinematic and dynamic models of the vehicle, a hierarchical control strategy is designed with longitudinal speed as the coupling point. Specifically, the longitudinal control uses a position-velocity dual PID to generate acceleration commands, while the lateral control employs MPC to optimize the front wheel angle in real time. A cooperative architecture is constructed to achieve dynamic decoupling. In a joint simulation, experiments on a composite path consisting of an S-shaped curve and a right-angle turn show that under the cooperative control, the peak lateral tracking error is only 0.4 m, the yaw rate fluctuation range is reduced by 30%, the speed tracking accuracy error is less than 5%, and the adjustment amplitude of the front wheel angle is reduced by 30%. The results demonstrate that this algorithm effectively coordinates the conflicts between lateral and longitudinal controls, significantly improving the path tracking accuracy and driving stability, providing theoretical support for the practical application of autonomous driving.

Key words: autonomous driving; model predictive control; lateral-longitudinal cooperative control

摘要： 针对自动驾驶汽车在复杂场景下因横纵向运动耦合导致的轨迹偏差与稳定性下降的问 题，文章提出一种基于双比例-积分-微分（PID）与模型预测控制（MPC）的横纵向协同控制 算法。通过建立车辆运动学与动力学模型，以纵向车速为耦合点设计分层控制策略，即纵向 控制采用位置-速度双 PID 生成加速度指令，横向控制通过模型预测控制实时优化前轮转角， 并构建协同架构实现动态解耦。在联合仿真中，针对 S 形弯道与直角转弯复合路径的实验表 明，协同控制下横向跟踪误差峰值仅为 0.4 m，横摆角速度波动范围降低 30%，速度跟踪精度 误差小于 5%，且前轮转角调整幅度减少 30%。结果表明，该算法能有效协调横纵向控制冲突， 显著提升路径跟踪精度与行驶稳定性，为自动驾驶实际应用提供了理论支持。

关键词: 自动驾驶；模型预测控制；横纵向协同控制