摘要
针对自由漂浮空间服务机器人在轨捕获时与服务对象的碰撞问题,提出一种考虑模型不确定情况下的无力传感器阻抗控制方法。通过将不确定模型观测的接触力和神经网络对模型不确定部分辨识得到的接触力,共同作为末端执行器与服务对象抓捕点的估计接触力和阻抗控制器的反馈,使空间机器人在模型不确定和无力传感器的情况下仍能实现与服务对象的柔性接触。数值仿真结果表明,该方法可有效降低对空间机器人动力学建模精度的要求,具有较强的工程可行性。
A force sensorless impedance control approach is proposed for the collision between the free-floating space robot and the captured object by taking the dynamic parameters uncertainty into account.The contact force identified by the neural network and the other contact force observed by the inaccurate dynamic model are both employed as the estimated contact force between the end effector and the target object as well as the feedback of the impedance controller.The method enables the space robot to flexibly capture the service object even if the dynamic parameters are uncertain and the force sensor is not assembled.The numerical simulation results show that this method can effectively reduce the requirement for the precision of space robot dynamics modeling and is practical in engineering.
作者
陶东
张强
赵良玉
TAO Dong;ZHANG Qiang;ZHAO Liang-yu(School of Aerospace Engineering,Beijing Institute of Technology,Beijing 100081,China;Beijing Institute of Control Engineering,Beijing 100094,China)
出处
《宇航学报》
EI
CAS
CSCD
北大核心
2021年第6期766-774,共9页
Journal of Astronautics
基金
国家重点研发计划(2018AAA0103004)
国家自然科学基金(51805025,11532002)。
