中文摘要：

建立基于干扰估计的机器人非线性反馈控制系统并证明其稳定性，在此基础上提出一种适用于机器人跟踪控制的新型自抗扰控制器。该控制器不需实时计算复杂的机器人动态模型，由两个跟踪微分器（TD）构成：一个用于安排系统的过渡过程；另一个用来估计速度和加速度，TD的滤波特性使其对量测噪声具有抑制作用。由被控对象的控制量与所估计加速度的反馈构成的“扩张状态”来自动检测系统模型和外扰的实时作用并实时进行动态补偿。除了和以往的自抗扰控制器一样具有很好的适应能力和很强的鲁棒性外，它还具有需整定参数少的特点。仿真结果表明，该控制器是有效的且具有很强的鲁棒性，而且系统响应快且超调小。

英文摘要：

A disturbance estimation-based nonlinear feedback control system for robots is presented, and its stability is demonstrated. On the basis of result, a novel active disturbance rejection controller （ARDC） is proposed, which avoids the burdensome online computation of robot inverse dynamics and implement the trajectory tracking of robots. The controller comprises two tracking differentiators （TDs） , one is used to arrange the transient process of system, and the other is used to estimate velocity and acceleration. TD can attenuate measurement noise due to its filtering characteristic. ＂Extended state＂, comprising the control input of controlled object and the feedback of estimated acceleration, is used to detect the real time effects of system model and the external disturbances, and then dynamically compensate their real-time effects. The proposed ARDC has the same good adaptability and expected robustness as previous ADRC, and it also has the characteristic of less parameters adjustment. Numerical results show that the control system has less overshoot, rapid response, and robustness.