基于单神经元的汽车方向自适应PID控制

针对汽车方向动力学控制存在的非线性和参数时变不确定性问题，提出了一种新的基于单神经元的汽车方向自适应PID控制算法。该算法利用了神经网络的自学习和自适应能力，实现了方向PID控制器的参数在线自整定，从而避免了传统的自适应PID控制必须在线辨识被控系统的参考模型参数而带来的计算工作量大的问题。仿真计算和场地试验验证表明该控制算法可有效地控制汽车按照预期给定的轨迹行驶，且保证了汽车方向闭环控制系统具有较强的适应性和鲁棒性。

Vehicle Direction Adaptive PID Control Based on Single Neural Cell

In view of the nonlinearity and parameter time-varying uncertainty of vehicle dynamics,a novel algorithm,i.e.single neural adaptive PID control strategy,is propsed for vehicle direction control.Based on self-learning and adaptive ability of neural network,the on-line self-tuning for parameters of direction PID controller is realized and the problem of long computation time for typical adaptive PID control can be avoided,in which the parameters of reference model of the controlled system must be identified on-line.The results of simulation and field test show that the algorithm can effectively control vehicle to follow the preset trajectory,and ensure the vehicle direction closed-loop control system have good robustness and adaptability.