机器人LFT变增益∞控制

针对平面两关节直接驱动机器人, 基于LMI技术提出一种设计能保证在整个运动范围内始终具有很好动态性能的LFT变增益H_∞ 控制器的新方法. 将机器人系统转化为以两关节夹角余弦值为变参数的LPV模型并表示为关于变参数的LFT结构, 利用变参数的测量值设计具有相同LFT结构的LPVH_∞ 控制器, 将此控制器的设计等价为以变参数为不确定项的LTI鲁棒控制器的设计并给出控制器可解的LMIs条件, 然后归纳出获得控制器的求解方法. 此控制器既克服了传统变增益控制器的缺陷,

LFT gain scheduled H-infinity control for a robotic manipulator

A new approach to design an LFT gain scheduled H-infinity controller guaranteeing good dynamic performance in the whole operation range for a planar two-joint direct drive robotic manipulator was presented. The robot system was translated into an LPV (linear parameter varying) model taking the cosine of joint between two joints as the varying parameter. This LPV model was expressed as LFT structure about the varying parameter. Using the measured value of the varying parameter, the LPV H-infinity controller was designed with the same LFT structure. The controller design was equivalent to the robust LTI controller design taking the varying parameter as uncertainty term. The solvable LMIs (linear matrix inequalities) condition of the designed controller was given and the solved method to obtain the controller was induced. The designed controller not only overcomes the limitation of the conventional gain scheduled controller, but also reduces the conservation via the measured varying parameter. Experiment results prove the effectiveness and advancement of the designed controller.