Disturbance estimation for robotic systems using continuous integral sliding mode observer

Abstract

This article presents a novel force‐sensor‐less method for the estimation of external forces for a general class of second‐order robotic systems. The method is based on the integral sliding mode observer (ISMO) which serves as a second‐order differentiator for the position measurement of the system. As a result, the system states and disturbance are estimated without explicitly using force and velocity measurements. To apply the ISMO to the general second‐order systems, a proper assumption is proposed to address their nonlinearity and discontinuity. The boundary‐layer method is applied to ensure that the virtual inputs of the observer are continuous such that the chattering phenomenon is attenuated. A Lyapunov‐based method is used to analyze the influence of the boundary layers on the convergence of the state‐ and disturbance‐estimation errors. This influence, mainly determined by the boundary‐layer scalars, is given in analytical forms as a reference for parameter selection. The method is evaluated by numerical simulation on a robot manipulator system and compared with a conventional sliding mode observer (SMO). The validation of the performance of the continuous ISMO indicates its generalizability to general second‐order robotic systems. Also, the advantage of continuous ISMO over the conventional SMO is reflected by its small estimation errors and superior responsiveness. In general, the proposed method in this paper may interest those who are seeking solutions for haptic robotic tasks without using force sensors.

Practical preset time fault‐tolerant control of uncertain Euler–Lagrange systems with input saturation and guaranteed performance

“…Therefore, it is still a challenging topic to obtain a high precision trajectory tracking control strategy for robotic systems. During the past decade, numerous control strategies are proved to be valid for robotic systems, such as sliding mode control 1,2 , model predictive control (MPC), 3 adaptive control 4,5 and fuzzy control. 6,7 Noteworthily, the majority of the foregoing methods are primarily concerned with situations where the actuator can adequately execute control commands.…”

“…To dispose of the aforementioned problem, FTC has drawn much attention because of its ability to overcome the adverse effects caused by various faults. In general, the FTC is often classified into two groups: (1) active FTC (AFTC) [8][9][10] and (2) passive FTC (PFTC). With AFTC techniques, the satisfactory performance can be ensured since the controller is reconfigured in real-time in accordance with the signals collected from the fault detection and isolation (FDI) unit.…”