Environment estimation for enhanced impedance control

Abstract

Impedance control is a popular method of controlling the dynamic response a' robot has to external forces. The advantage of such a control paradigm is the ability to manipulate in unconstrained, as well as constrained, environments. Unlike hybrid control methods that attempt to control forces and motions in orthogonal directions, impedance control consists of a single control scheme that accommodates both unconstrained and constrained maneuvers. One problem associated with such a control scheme is the selection of the robot's target impedance. This paper will illustrate, through analysis and experimentation, the shortcoming of improper target impedance selection. A new approach to adapting the target impedance, based upon an on-line estimate of the environment impedance, is proposed and demonstrated.

Optimal Critic Learning for Robot Control in Time-Varying Environments

“…To design an impedance control for an unknown environment, estimation of environment parameters has been an option and was extensively studied in the literature [9], [10]. A desired impedance model can be constructed if the stiffness and damping parameters of the environment can be precisely estimated.…”

“…A desired impedance model can be constructed if the stiffness and damping parameters of the environment can be precisely estimated. In [9], a Recursive Least Square (RLS) scheme has been implemented to estimate the environment parameters. Unfortunately, as discussed in [11], identification is usually time-consuming because it requires the procedures of model design, parameter estimation, and model validation at each step of the iterations.…”

Previewed Impedance Adaptation to Coordinate Upper-Limb Trajectory Tracking and Postural Balance in Disturbed Conditions

“…Among existing contributions in this domain, control methods such as impedance controllers 3 allows to define the reactive behavior of a manipulator during interaction tasks of which parameters can be defined in an optimal way. 4 This resulting behavior of the upper-limbs induces perturbations 5 on the lower-bodies of the robot. The control of these two systems having coupled dynamics and independent objectives can be handled in an anticipative way using a distributed model predictive control (DMPC) method 6 dynamically tuning reactive control parameters.…”

Online Contact Impedance Identification for Robotic Systems

“…For instance, environment parameter estimation has been used in the design of indirect adaptive impedance controllers [2], [3] or model reference adaptive controllers [4]. Online identified environment impedance has also been used for transparency in teleoperation systems [1], [5] or the identified range of the environment can be used to analyze robust stability of teleoperation systems [6].…”