Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Impedance Control for Manipulators Handling Heavy Payloads

Farhad Aghili

TL;DR

The paper tackles impedance control for manipulators carrying heavy payloads, where inertial and gravitational forces contaminate wrist force/moment sensing and can prevent achieving the desired impedance. It proposes an inner/outer loop impedance controller that realizes a general target impedance without relying on velocity/acceleration measurements, with extensions to nonlinear models and robustness to parameter uncertainties; stability and bounded control input are proven under explicit conditions such as $\det(\mathbf{1}-\mathbf{M}_p\mathbf{M}_d^{-1}) \neq 0$ and $\|\Delta_m\| \le \kappa(\mathbf{G})$. The nonlinear impedance formulation uses $M_d \ddot{\bm x} + h_d(\dot{\bm x}, \bm x) = \bm f_{ext}$ and introduces an acceleration estimate $\ddot{\bm x}^{\star}$ to realize the target dynamics; a robust variant adds a corrective input $\bm w$ with boundary-layer smoothing to handle modeling uncertainty. Experimental validation with a $16$ kg payload confirms effective impedance shaping and shows practical performance metrics (e.g., RMSEs around a few percent and successful contact tasks), demonstrating the method’s applicability to space robotics and heavy-load manipulation without additional accelerometers or estimators.

Abstract

Attaching a heavy payload to the wrist force/moment (F/M) sensor of a manipulator can cause conventional impedance controllers to fail in establishing the desired impedance due to the presence of non-contact forces; namely, the inertial and gravitational forces of the payload. This paper presents an impedance control scheme designed to accurately shape the force-response of such a manipulator without requiring acceleration measurements. As a result, neither wrist accelerometers nor dynamic estimators for compensating inertial load forces are necessary. The proposed controller employs an inner-outer loop feedback structure, which not only addresses uncertainties in the robot's dynamics but also enables the specification of a general target impedance model, including nonlinear models. Stability and convergence of the controller are analytically proven, with results showing that the control input remains bounded as long as the desired inertia differs from the payload inertia. Experimental results confirm that the proposed impedance controller effectively shapes the impedance of a manipulator carrying a heavy load according to the desired impedance model.

Impedance Control for Manipulators Handling Heavy Payloads

TL;DR

The paper tackles impedance control for manipulators carrying heavy payloads, where inertial and gravitational forces contaminate wrist force/moment sensing and can prevent achieving the desired impedance. It proposes an inner/outer loop impedance controller that realizes a general target impedance without relying on velocity/acceleration measurements, with extensions to nonlinear models and robustness to parameter uncertainties; stability and bounded control input are proven under explicit conditions such as and . The nonlinear impedance formulation uses and introduces an acceleration estimate to realize the target dynamics; a robust variant adds a corrective input with boundary-layer smoothing to handle modeling uncertainty. Experimental validation with a kg payload confirms effective impedance shaping and shows practical performance metrics (e.g., RMSEs around a few percent and successful contact tasks), demonstrating the method’s applicability to space robotics and heavy-load manipulation without additional accelerometers or estimators.

Abstract

Attaching a heavy payload to the wrist force/moment (F/M) sensor of a manipulator can cause conventional impedance controllers to fail in establishing the desired impedance due to the presence of non-contact forces; namely, the inertial and gravitational forces of the payload. This paper presents an impedance control scheme designed to accurately shape the force-response of such a manipulator without requiring acceleration measurements. As a result, neither wrist accelerometers nor dynamic estimators for compensating inertial load forces are necessary. The proposed controller employs an inner-outer loop feedback structure, which not only addresses uncertainties in the robot's dynamics but also enables the specification of a general target impedance model, including nonlinear models. Stability and convergence of the controller are analytically proven, with results showing that the control input remains bounded as long as the desired inertia differs from the payload inertia. Experimental results confirm that the proposed impedance controller effectively shapes the impedance of a manipulator carrying a heavy load according to the desired impedance model.
Paper Structure (7 sections, 1 theorem, 64 equations, 6 figures, 1 table)

This paper contains 7 sections, 1 theorem, 64 equations, 6 figures, 1 table.

Table of Contents

  1. Introduction
  2. Impedance Control
  3. Impedance Control with Inner/Outer Loop
  4. Nonlinear Impedance Model
  5. Robust Impedance Control
  6. Experiment
  7. Conclusions

Key Result

Theorem 1

Assume that the desired inertia matrix is selected such that eq:cond_bounded is satisfied. Then, applying control law eq:control_law_fs to the manipulator system eq:manipulator_imp attached to a payload with generalized inertia $\bm M_p$ establishes the desired impedance eq:desired_imp.

Figures (6)

  • Figure 1: A manipulator carrying a heavy payload.
  • Figure 2: Dynamics force and moment.
  • Figure 3: Trajectories of the linear and angular velocities.
  • Figure 4: The $z$-axis position of the payload.
  • Figure 5: Contact force and moment.
  • ...and 1 more figures

Theorems & Definitions (2)

  • Theorem 1
  • Remark 1