Position-force control of mobile manipulator – nonadaptive and adaptive case

Abstract

This article presents a control algorithm for nonholonomic mobile manipulators with a simple, geometric holonomic constraint imposed on the robot's arm. A mathematical model in generalized, auxiliary and linearized coordinates is presented, as well as the constrained dynamics of the robotic system. A position-force control law is proposed, both for the fully known robot's model, as well as for the model with parametric uncertainty in the dynamics. Theoretical considerations are supported by the results of computer simulations.

Robust Position/Force Control of Nonholonomic Mobile Manipulator for Constrained Motion on Surface in Task Space

“…Another difficulty for designing a controller for MMs arises from holonomic constraints (e.g., as the task constraints and the structure of the RM) and nonholonomic constraints restricting the motion of the MR [25]. Few studies have focused on designing a controller for MMs to simultaneously provide position and force tracking [24,[26][27][28][29]. In [24], a controller is developed to track the reference position/force trajectories for a MM with both nonholonomic and holonomic constraints.…”

“…Few studies have focused on designing a controller for MMs to simultaneously provide position and force tracking [24,[26][27][28][29]. In [24], a controller is developed to track the reference position/force trajectories for a MM with both nonholonomic and holonomic constraints. In [30],…”

“…an NN-based extension attached to a model-based controller is designed to track the reference position and force of an MM in the presence of uncertainties and disturbances. The control designs in [24,30] require the exact model of the dynamics of the MM. In [29], an adaptive controller is designed for MM with holonomic and nonholonomic constraints to ensure force/position tracking despite the uncertainties and disturbances in its dynamic.…”

“…It is assumed that the force measurement of the end-effector of MM on task space is available as in [34]. The designed controller does not require any linearization technique or the linearly parameterizable assumption for the MMs dynamic, unlike proposed controllers in [24,27,[29][30][31]. Furthermore, the controller in [24,30] needs the exact knowledge of the dynamic model of MM, however, in this paper, the robust controller is designed to track the position/force of MM without the requirements of an exact model of the mobile manipulator.…”

“…The designed controller does not require any linearization technique or the linearly parameterizable assumption for the MMs dynamic, unlike proposed controllers in [24,27,[29][30][31]. Furthermore, the controller in [24,30] needs the exact knowledge of the dynamic model of MM, however, in this paper, the robust controller is designed to track the position/force of MM without the requirements of an exact model of the mobile manipulator. The controller in [24] does not consider the disturbances in the dynamic; however, for real-time applications, most of the dynamics are subject to disturbances.…”