Adaptive actuator failure compensation for cooperative robotic manipulators with parameter uncertainties

Abstract

This article develops a new framework of adaptive actuator failure compensation control for cooperative manipulator systems with parameter uncertainties in addition to actuator failures, and designs and analyzes effective actuator failure compensation schemes for such robotic systems. The new adaptive control design uses an integration of multiple individual failure compensators and direct adaptation to handle various types of uncertainties in such robotic systems. The design can also be used for concurrent actuator failure cases, to expand the capability of adaptive actuator failure compensation. With a complete proof and performance analysis, it is shown that the proposed control scheme guarantees the desired closed-loop stability and asymptotic output tracking, despite actuator failures whose patterns, time instants and values are all unknown. Simulation results of a benchmark cooperative manipulator system are presented to verify the desired control performance of the system with both typical constant and square-wave actuator failure signals.

Adaptive compensation control of nonlinear systems with unknown actuator failures

“…Remark The actuator failure model (2) can be used to closely approximate a large class of practical failures, by a proper selection of these basic functions $$$ {f}_{jlh}(t) $$$, while parametrized by $$$ {\overline{d}}_{jlh} $$$. This kind of parameterized failure model has been applied in many literature, such as References 11,12,34‐36. Compared with Reference 37, the failure model used in our article contains more fault information.…”

Cooperative Robot Manipulators Dynamical Modeling and Control: An Overview

“…The challenge of dealing with faulty dynamical systems and potential solutions are explored in [142,159]. In the former, four types of faults are addressed, including faulty joint mechanisms and actuators, locked joint mechanisms, incorrect joint mechanism position measurements, and incorrect joint mechanism velocity measurements.…”

Viability Under Degraded Control Authority

“…In the context of switching systems, Niu et al [ 5 ] considered the problem of active mode discrimination (AMD) with temporal logic-constrained switching, where a set of known switching modes was known a priori . Other work relies on application-specific projection-based adaptation routines where a set of switching adaptive controllers is used [ 6 ]; however, such an approach does not incorporate non-smooth failure modes that conditionally act depending on the control input provides, nor are regions of degradation robustly identified as is done in the present work. A major contribution is the fact that we do not demand complete knowledge of the underlying system structure, which is essential in Lyapunov analysis for adaptive methods.…”