Online Stability in Human-Robot Cooperation with Admittance Control

Abstract

In the design of a compliant admittance controller for physical human-robot interaction, it is necessary to ensure stable and effective cooperation. The stability of the admittance controller is mainly threatened by a stiff environment. Many methods that guarantee stability in arbitrary environments, impose conservative control gains that limit the effectiveness of the cooperation. Inspired by previous work in frequency domain stability observers, a method is proposed in this paper to detect unstable behavior and stabilize the robot with online adaptation of the admittance control gains. The introduced instability index is based on frequency domain analysis, which very quickly detects unstable behavior by monitoring high frequency oscillation in the force signal. To treat the instability, an adaptation scheme of the admittance parameters is proposed, that relaxes conservative gains and improves the cooperation by considering the effect of variable admittance on the operators' effort. We investigate two human-robot co-manipulation tasks; cooperation within a zero stiffness environment and cooperation in contact with a stiff double-wall virtual environment. The proposed methods are validated experimentally with a number of subjects in cooperation with an LWR manipulator.

Progress and prospects of the human–robot collaboration

“…NAC and complementary stability were combined in [Lacevic and Rocco, 2011], but, as observed in [Dimeas and Aspragathos, 2016], their conservative gains were still hindering a smooth collaboration without efforts for the human operator. In [Dimeas and Aspragathos, 2016], the authors proposed online adaptation of an admittance controller gains on the basis of an instability index, computed in the frequency domain, that detects unstable behaviours of the human-robot coupled system by monitoring high-frequency oscillations in the force signal.…”

“…Dimeas and Aspragathos observed in [Dimeas and Aspragathos, 2016] that the bandwidth of voluntary motion in humans is relatively low and below 2Hz [de Vlugt et al, 2003], hence during physical interaction with the robot it is possible to discriminate the human operator's intent from the unstable motions thanks to frequency analysis. External oscillatory excitations of the arm up can be mitigated by the central nervous system (CNS) by changing the arm impedance: indeed, it has been shown by [E. Burdet, 2001] that the CNS can learn to control the magnitude, shape and orientation of the endpoint stiffness in a predictive way that is independent of the force needed to compensate for the imposed dynamics.…”

Fractional order admittance control for physical human-robot interaction

“…In such a condition, mass and damping values in human arm model would not change significantly, whereas arm stiffness may change drastically due to the cocontraction of muscles. Hence, we are investigating robustness of the proposed controller to stiffness changes since it has been reported that changes in human arm stiffness significantly contributes to the instabilities observed in pHRI [1], [15]. The stability analysis is conducted for a range of human arm stiffness values, and typical values of b h = 15 Ns/m and m h = 1.5 kg are used for human arm damping and mass, respectively, as reported in [6], [7].…”

Autonomous Self-Explanation of Behavior for Interactive Reinforcement Learning Agents

“…Human Robot Cooperation (HRC), in which people and robots work on the same task together in a shared environment, is an effective concept for both industrial and domestic robots [1,6]. By working in a complementary manner, both robots and people overcome the disadvantages of each other in order to achieve difficult tasks that cannot be achieved by either of them.…”