双轮躯干-臂式机器人非确定性等价自适应容错控制方法

摘要:

在常规依靠PID控制器实现自适应容错控制时,双轮躯干-臂式机器人易受外部环境扰动使机器人在运动时难以按预设轨迹精确运动,导致运动跟踪误差较大,影响机器人整体性能。因此,提出双轮躯干-臂式机器人非确定性等价自适应容错控制方法。构建反映运动特性的动力学模型,结合扩展卡尔曼滤波器的改进自适应网络,设计故障在线诊断算法,通过学习机器人输出参数检测运行故障状态。针对运行环境中的不确定干扰因素,构造干扰观测器,求出目标时间段环境干扰估计值。引入滑模控制思想,构造基于RBF神经网络的自适应容错控制方案,结合故障诊断和干扰估计结果,实现对故障的有效补偿。实验结果表明:该方法进行控制处理后,机器人在故障状态下纵荡轨迹跟踪误差保持在0.2m以内,偏航角度跟踪误差保持在0.1rad以内,极大提升了机器人运行稳定性。

Abstract:

When relying on PID controllers to achieve adaptive fault-tolerant control, dual wheel torso arm robots are susceptible to external environmental disturbances, making it difficult for the robot to move accurately along the preset trajectory, resulting in significant motion tracking errors and affecting the overall performance of the robot. Therefore, a non deterministic equivalent adaptive fault-tolerant control method for a dual wheel torso arm robot is proposed. Construct a dynamic model that reflects the motion characteristics, and design an online fault diagnosis algorithm by combining an improved adaptive network with an extended Kalman filter. By learning the output parameters of the robot, detect the operating fault state. Construct an interference observer to estimate the environmental interference during the target time period in response to uncertain interference factors in the operating environment. Introducing the concept of sliding mode control, constructing an adaptive fault-tolerant control scheme based on RBF neural network, and combining fault diagnosis and disturbance estimation results to achieve effective compensation for faults. The experimental results show that after control processing using this method, the longitudinal trajectory tracking error of the robot remains within 0.2m and the yaw angle tracking error remains within 0.1rad under fault conditions, greatly improving the stability of the robot operation.