%0 Journal Article
%T 煤矸石分拣机械臂的轨迹跟踪控制研究<br>Research on Trajectory Tracking Control for Coal Gangue Sorting Robot Arm
%A 王长浩
%A 秦强
%J Artificial Intelligence and Robotics Research
%P 437-447
%@ 2326-3423
%D 2022
%I Hans Publishing
%R 10.12677/AIRR.2022.114045
%X 煤矸石分拣机械臂在捡取过程中，带式输送机上煤矸石抖振等外部因素对机械臂末端形成冲击，导致机械臂控制系统出现不确定性问题，造成机械臂抓取煤矸石时控制性能下降、轨迹跟踪误差偏大；为此，在传统滑模控制的基础上，改进出一种RBF神经网络的切换增益调节滑模控制方法。根据拉格朗日理论推导出煤矸石分拣机械臂动力学数学模型，针对模型中不确定部分，结合滑模控制算法与RBF神经网络，利用RBF神经网络来调节滑模控制中的切换增益，使得切换增益能够随着不确定部分改变而改变，同时减弱传统滑膜控制中的抖振问题。采用MATLAB软件，进行所述控制器仿真实验，分析机械臂关节角度跟踪效果及切换增益调节响应时间。仿真结果表明：本文所述控制方法可以更快地贴近期望轨迹值，同时跟踪误差也得到降低。搭建实验平台，在煤矸石分拣机械臂上展开实验，模拟分拣煤矸石几字形路径。根据激光标定仪反馈数据得出：该控制方法可以保证机械臂运动的合理性和安全性，能够实现较短时间内收敛达到煤矸石分拣过程中的期望轨迹，符合煤矸石分拣的实际工况需求。<br />
During the picking process of the coal gangue sorting robotic arm, external factors such as buffeting of coal gangue on the belt conveyor impact the end of the manipulator, which leads to the uncertainty of the control system of the manipulator, and causes the control system when the manipulator grabs the coal gangue. The performance is degraded and the trajectory tracking error is too large; for this reason, based on the traditional sliding mode control, an RBF neural network switching gain adjustment sliding mode control method is improved. According to the Lagrangian theory, the dynamic mathematical model of the coal gangue sorting manipulator is deduced. For the uncertain part of the model, the sliding mode control algorithm and the RBF neural network are combined, and the RBF neural network is used to adjust the switching gain in the sliding mode control. The switching gain can be changed with the change of the uncertain part, while reducing the chattering problem in the traditional synovial control. Using MATLAB software, the controller simulation experiment was carried out, and the tracking effect of the joint angle of the manipulator and the response time of the switching gain adjustment were analyzed. The simulation results show that the control method described in this paper can approach the desired trajectory value faster, and the tracking error is also reduced. Build an experimental platform, carry out experiments on the coal gangue sorting robotic arm, and simulate the zigzag path of sorting coal gangue. According to the feedback data of the laser calibrator, it is concluded that the control method can ensure the rationality and safety of the movement of the manipulator, and can achieve the desired trajectory in the coal gangue sorting process by convergence in a short time, which is in line with condition needs of the actual work of coal gangue sorting.
%K 机械臂，轨迹跟踪，RBF神经网络，滑膜控制<br>Robotic Arm
%K Trajectory Tracking
%K RBF Neural Network
%K Sliding Mode Control
%U http://www.hanspub.org/journal/PaperInformation.aspx?PaperID=58033