受限状态下的高速列车迭代学习控制方法研究

针对受限状态下的高速列车自动驾驶系统的跟踪控制问题,基于列车动力学模型,提出一种带饱和函数的迭代学习控制算法。根据Lyapunov稳定性原理,利用列车运行过程中的状态偏差,推导出基于迭代学习控制的列车自动运行控制律。建立类Lyapunov的复合能量函数,通过在迭代域的差分,证明了其差分负定性和有界性,所设计的算法能够控制列车在迭代域对期望运行轨迹达到渐近收敛。采用本文提出的迭代学习控制算法对列车的跟踪性能进行验证,并与PID控制和D型迭代学习控制算法进行比较,结果表明:相较于其他两种算法,本文提出的算法在第3次迭代中就能控制列车精确跟踪期望轨迹,说明算法具有较快的收敛速度和较高的跟踪精度,且能够将控制输入约束在允许范围内。

State-constrained Iterative Learning Control Algorithm for High-speed Train Operation

To study the automatic operation tracking problem of high-speed train under state-constrained situation, this paper proposes a novel iterative learning control algorithm with saturation function based on the train dynamic model. According to Lyapunov stability principle, we deduce the iterative learning control law of train operation by applying the operation state deviation. Then the differential negative definiteness and boundedness of the Lyapunov-based composite energy function are verified. The proposed algorithm can achieve asymptotic convergence to the desired trajectory along the iteration domain. The proposed iterative learning control algorithm is applied to verify the tracking performance and compared with PID control algorithm and D-type iterative learning control algorithm. The results show that the proposed algorithm can control precisely the desired profile of train tracking at the third iteration time, which demonstrates that the proposed algorithm possesses better performance on convergence speed and tracking accuracy and can restrain the control input within an allowed range.