In this article, we propose a robust hierarchical inverse dynamics control scheme for redundant manipulators, which guarantees both predefined-time convergence of regulation tasks and robustness against model uncertainties/disturbances. Predefined-time convergence of a robotic task means that the task error is regulated to the origin at a desired preset time, independently of the initial state of the robot. The strict hierarchy of tasks is ensured by the so-called dynamic consistency, which avoids conflicts between decoupled task constraints. Thus, it enables the robot with more dexterity for reaching targets in its workspace while overcoming obstacles and joint limits. Predefined-time convergence facilitates the design of a time schedule to execute sequential and simultaneous tasks. All these features are supported by a stability analysis and experiments with two torque-controlled mobile manipulators in pick-and-place applications.

中文翻译：

---

转矩控制冗余机械臂上的预定义时间鲁棒分层逆动力学

在本文中，我们为冗余机械手提出了一种鲁棒的分层逆动力学控制方案，它既保证了调节任务的预定义时间收敛，又保证了对模型不确定性/干扰的鲁棒性。机器人任务的预定义时间收敛意味着任务误差在所需的预设时间被调节到原点，与机器人的初始状态无关。所谓的动态一致性保证了任务的严格层次结构，避免了解耦任务约束之间的冲突。因此，它使机器人能够更灵活地到达其工作空间中的目标，同时克服障碍和关节限制。预定义时间收敛有助于设计时间表以执行顺序和同时任务。