EI、Scopus 收录
中文核心期刊

6-PSU并联机器人动力学模型辩识及其驱动力矩估计新方法

A NEW METHOD FOR DYNAMIC MODEL IDENTIFICATION AND DRIVING TORQUE ESTIMATION FOR 6-PSU PARALLEL ROBOTS

  • 摘要: 并联机器人运动奇异常常会导致电机驱动力矩过载甚至损坏.为降低驱动电机的故障率,本文以6-PSU型并联机器人为研究对象,提出了基于动力学建模及其参数辨识的解决方案,能够根据预设轨迹提前精确估计驱动力矩以便用于安全性校核.首先,结合凯恩方法和虚功原理,建立了包含滚珠丝杠模组摩擦、伺服电机摩擦和惯性的逆动力学精细化模型;利用QR分解获得了最小可辨识参数集,在此基础上考虑了连杆截面的对称性特点,发现了可忽略的连杆惯量,得到了简化的线性化模型;利用统计的方法,验证了激励轨迹优化时考虑摩擦的必要性;利用Simulink/Multibody搭建机器人物理仿真模型进行了参数辨识与驱动力矩估计仿真实验,针对最小可辨识参数集模型和简化模型分别定量分析了两者的参数辨识精度、驱动力估计精度以及抗噪能力;最后,在6-PSU型并联机器人平台上开展了实验验证,结果表明,本文提出的简化模型在保证相同驱动力矩估计精度的前提下,有效简化了模型复杂度,具有明确的工程价值.

     

    Abstract: Parallel robot kinematic singularities often lead to excessive motor driving torque and even damage. To reduce the failure rate of drive motors, this paper takes a 6-PSU type parallel robot as the research object and proposes a solution based on dynamic modeling and parameter identification, which can accurately estimate the driving torque in advance according to the preset trajectory for safety verification. First, combining the Kane’s method and the principle of virtual work, an inverse dynamic refined model that includes the friction of ball screw modules, servo motor friction, and inertia is established; QR decomposition is used to obtain the minimum identifiable parameter set, and based on this, considering the symmetry of the cross-section of the linkage, the negligible linkage inertia is discovered, resulting in a simplified linearized model; statistical methods are used to verify the necessity of considering friction when optimizing the excitation trajectory; a physical simulation model of the robot is built using Simulink/Multibody for parameter identification and driving torque estimation simulation experiments, and the parameter identification accuracy, driving torque estimation accuracy, and noise resistance of both the minimum identifiable parameter set model and the simplified model are quantitatively analyzed; finally, experiments are conducted on the 6-PSU type parallel robot platform, and the results show that the simplified model proposed in this paper effectively simplifies the model complexity while ensuring the same driving torque estimation accuracy, and has clear engineering value.

     

/

返回文章
返回