基于RBFNN的跨座式单轨车辆自动驾驶滑模控制研究

Sliding Mode Control for Automatic Driving of Straddle Monorail Vehicles Based on RBFNN

针对跨座式单轨全自动驾驶控制问题,在分析跨座式单轨运行场景的基础上,建立了基于RBFNN的跨座式单轨动力学模型,明确了控制目标,提出了基于基于径向基神经网络(RBFNN)的跨座式单轨固定时间滑模控制方法,设计了基于RBFNN的跨座式单轨固定时间滑模控制器,并基于Lyapunov稳定性理论证明了控制器的稳定性。通过数值仿真结果表明:设计的控制算法能使车辆的位置和速度在15 s内均跟踪到指令曲线,而有限时间控制器在24 s才能跟踪上指令曲线;有RBFNN相较于无RBFNN,在位置跟踪上更加接近指令位置约0.1 m,在速度跟踪上更加接近指令速度约0.01 m/s;仿真结果表明笔者设计算法能有效提升跨座式单轨车辆的运行效率,可为跨座式单轨全自动驾驶的实施提供控制算法借鉴。

Aiming at the problem of full-automatic driving control of straddle monorail,based on the analysis of the operation scene of straddle monorail,the dynamic model of straddle monorail based on RBFNN was established,and the control target was clearly defined.The fixed time sliding mode control method of straddle monorail based on RBFNN was proposed,and the fixed time sliding mode controller of straddle monorail based on RBFNN was designed.The stability of the proposed controller was proved based on Lyapunov stability theory.The numerical simulation results show that the designed control algorithm can make the vehicle position and speed track the command curve in 15s,while the finite time controller can track the command curve in 24s.Compared with the controller without RBFNN,the deigned controller with RBFNN is closer to the command position about 0.1m in position tracking and 0.01 m/s in speed tracking.The simulation results show that the designed control algorithm can effectively improve the operation efficiency of straddle monorail vehicles,which can provide control algorithm reference for the implementation of straddle monorail full-automatic driving.