首页|期刊导航|机械与电子|考虑质心偏置的水射流清洗机器人行走部RBF自适应滑模控制

考虑质心偏置的水射流清洗机器人行走部RBF自适应滑模控制OA

RBF Adaptive Sliding Mode Control for the Locomotion Unit of a Water-jet Cleaning Robot Considering Centroid Offset

中文摘要英文摘要

针对水射流清洗机器人在高盐雾、强湍流及顽固污垢层等非结构化工况下,因质心与几何中心不重合引入非完整约束,导致动力学参数强摄动、突变负载与水射流反冲激振相互耦合,进而引发运动学位姿漂移及轨迹跟踪稳态误差过大的难题,提出一种考虑质心偏置的 RBF 自适应滑模控制方法.首先,基于质心偏置下的非完整约束特性建立行走部动力学模型;其次,设计基础滑模控制器并用 Lyapunov 函数证明其渐近稳定性;最后,引入 RBF 神经网络对系统集总不确定项进行逼近,并融合最小参数学习法将权值调整简化为单参数自适应调节,构建改进型自适应滑模控制器,在降低计算负担的同时增强对非完整约束及强扰动工况的鲁棒性.实验结果表明,所提方法对期望位姿的控制最大误差小于 0.6 cm,轨迹跟踪更贴近最佳清洗路径,位姿超调与振荡得到显著抑制,有效提升了机器人的轨迹跟踪精度与驱动稳定性.

This paper addresses the challenge of kinematic pose drift and excessive steady-state trajec-tory tracking error in water-jet cleaning robots operating under unstructured chemical conditions such as high salt spray,strong turbulence,and stubborn dirt layers.These problems arise because the misalignment between the center of mass and geometric center introduces nonholonomic constraints,causing strong per-turbations in dynamic parameters,coupling between abrupt load variations and water-jet recoil-induced vibration.To solve this,an RBF adaptive sliding mode control method considering centroid offset is pro-posed.First,a dynamic model of the locomotion unit is established based on the characteristics of the non-holonomic constraints under centroid offset.Second,a baseline sliding mode controller is designed,and its asymptotic stability is proven using a Lyapunov function.Finally,a radial basis function(RBF)neural net-work is introduced to approximate the lumped system uncertainties,and a minimal parameter learning method is incorporated to simplify the weight update law into a single-parameter adaptive regulation,thus constructing an improved adaptive sliding mode controller.This approach reduces the computational bur-den while enhancing robustness against nonholonomic constraints and severe operating disturbances.Ex-perimental results demonstrate that the maximum control error for the desired pose is less than 0.6 cm,and the trajectory tracking adheres more closely to the optimal cleaning path.The pose overshoot and oscil-lation are significantly suppressed,effectively improving the trajectory tracking accuracy and driving stabil-ity of the robot.

朱赛伟

河南送变电建设有限公司,河南 郑州 450051

信息技术与安全科学

水射流清洗机器人行走部驱动控制RBF神经网络运动学模型自适应滑模控制器

water-jet cleaning robotdriving control of locomotion unitRBF neural networkkinemat-ic modeladaptive sliding mode controller

《机械与电子》 2026 (8)

65-72,8

河南送变电建设有限公司科技项目(SGTYHT/23-JS-003)

评论