Human Stabilization of Delay-Induced Instability of Haptic Rendering in a Stiffness Discrimination Task

Towards developing a coupled stability theory for haptic systems, we study the interaction of operators with time-delayed force feedback. In this work, we analyzed and validated experimentally the stability boundaries of an uncoupled system – without considering the human. We then designed an experiment in which the participants used a haptic device to interact with virtual elastic force fields in a stiffness discrimination task. We compared the performance and kinematics of users in uncoupled-unstable and uncoupled-stable conditions and characterized the stabilizing contribution of the users. We found that the users were able to perform the task regardless of the uncoupled-stability conditions. In addition, in uncoupled-unstable conditions, users maintained movement characteristics that were important for exploratory mediation, such as depth and duration of the movement, whereas other characteristics were not preserved. The results were reproduced in a simulation of the human controller that combined an inverse model and an optimal feedback controller. Adequate performance under the uncoupled-unstable yet coupled-stable conditions supports the potential benefit of designing for coupled stability of haptic systems. This could lead to the use of less conservative controllers than state-of-the-art solutions in haptic and teleoperation systems, and advance the fidelity of haptic feedback.

I. Introduction

Haptic systems enable human operators to sense a virtual or remote environment by rendering forces from the environment, and have a wide range of applications, such as surgical training [1], rehabilitation [2] and teleoperation [3], [4]. Adding force feedback enables the operator to feel the impedance of a remote or simulated environment, increases the feeling of direct manipulation (also known as transparency [5]) and facilitates the execution of complex tasks [6], [7], [8]. However, in state-of-the-art teleoperation systems, rendering impedance accurately to the operator is limited by the need to ensure stability [9]. Instability may arise from various factors, such as time discretization [10], [11], quantization [12], and, especially, delays [4], [13], [14], and it results in unwanted disturbances in the movements of the manipulators. A safe teleoperation system must be stable, especially in applications such as teleoperated surgery, where any unwanted movement may cause severe damage.