SCIENCE CHINA Information Sciences, Volume 63 , Issue 7 : 170201(2020) https://doi.org/10.1007/s11432-019-2748-x

Bio-inspired robotic impedance adaptation for human-robot collaborative tasks

More info
  • ReceivedJun 18, 2019
  • AcceptedNov 29, 2019
  • PublishedMay 26, 2020


To improve the robotic flexibility and dexterity in a human-robot collaboration task, it is important to adapt the robot impedance in a real-time manner to its partner's behavior. However, it is often quite challenging to achieve this goal and has not been well addressed yet. In this paper, we propose a bio-inspired approach as a possible solution, which enables the online adaptation of robotic impedance in the unknown and dynamic environment. Specifically, the bio-inspired mechanism is derived from the human motor learning, and it can automatically adapt the robotic impedance and feedforward torque along the motion trajectory. It can enable the learning of compliant robotic behaviors to meet the dynamic requirements of the interactions. In order to validate the proposed approach, an experiment containing an anti-disturbance test and a human-robot collaborative sawing task has been conducted.


This work was supported by National Natural Science Foundation of China (Grant Nos. 61861136009, 61811530281).


[1] Peternel L, Petri? T, Oztop E. Teaching robots to cooperate with humans in dynamic manipulation tasks based on multi-modal human-in-the-loop approach. Auton Robot, 2014, 36: 123-136 CrossRef Google Scholar

[2] Li Y, Tee K P, Yan R. A framework of human-robot coordination based on game theory and policy iteration. IEEE Trans Robot, 2016, 32: 1408-1418 CrossRef Google Scholar

[3] Li Y, Tee K P, Yan R, et al. Adaptive optimal control for coordination in physical human-robot interaction. In: Proceedings of the 2015 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 2015. 20--25. Google Scholar

[4] Li Y, Ge S S. Human-robot collaboration based on motion intention estimation. IEEE/ASME Trans Mechatron, 2014, 19: 1007-1014 CrossRef Google Scholar

[5] Wu Y, Wang R, D'Haro L F, et al. Multi-modal robot apprenticeship: imitation learning using linearly decayed DMP+ in a human-robot dialogue system. In: Proceeding of the 2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 2018. 1--7. Google Scholar

[6] Wang R, Wu Y, Chan W L, et al. Dynamic movement primitives plus: for enhanced reproduction quality and efficient trajectory modification using truncated kernels and local biases. In: Proceedings of the 2016 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 2016. 3765--3771. Google Scholar

[7] Chen F, Sekiyama K, Cannella F. Optimal subtask allocation for human and robot collaboration within hybrid assembly system. IEEE Trans Automat Sci Eng, 2014, 11: 1065-1075 CrossRef Google Scholar

[8] Ko W K H, Wu Y, Tee K P, et al. Towards industrial robot learning from demonstration. In: Proceedings of the 3rd International Conference on Human-Agent Interaction. ACM, 2015. 235--238. Google Scholar

[9] He W, Dong Y, Sun C. Adaptive neural impedance control of a robotic manipulator with input saturation. IEEE Trans Syst Man Cybern Syst, 2016, 46: 334-344 CrossRef Google Scholar

[10] Denisa M, Gams A, Ude A. Learning compliant movement primitives through demonstration and statistical generalization. IEEE/ASME Trans Mechatron, 2016, 21: 2581-2594 CrossRef Google Scholar

[11] Yang C, Ganesh G, Haddadin S. Human-like adaptation of force and impedance in stable and unstable interactions. IEEE Trans Robot, 2011, 27: 918-930 CrossRef Google Scholar

[12] Burdet E, Ganesh G, Yang C, et al. Interaction force, impedance and trajectory adaptation: by humans, for robots. In: Experimental Robotics. Berlin, Heidelberg: Springer, 2014. 331--345. Google Scholar

[13] Ficuciello F, Villani L, Siciliano B. Variable impedance control of redundant manipulators for intuitive human-robot physical interaction. IEEE Trans Robot, 2015, 31: 850-863 CrossRef Google Scholar

[14] He W, Dong Y. Adaptive fuzzy neural network control for a constrained robot using impedance learning. IEEE Trans Neural Netw Learning Syst, 2018, 29: 1174-1186 CrossRef PubMed Google Scholar

[15] He W, Meng T, He X. Iterative learning control for a flapping wing micro aerial vehicle under distributed disturbances. IEEE Trans Cybern, 2019, 49: 1524-1535 CrossRef PubMed Google Scholar

[16] Liang X, Zhao H, Li X. Force tracking impedance control with unknown environment via an iterative learning algorithm. Sci China Inf Sci, 2019, 62: 050215 CrossRef Google Scholar

[17] Liang X, Zhao H, Li X, et al. Force tracking impedance control with unknown environment via an iterative learning algorithm. Sci China Inf Sci 2019, 62: 050215. Google Scholar

[18] Li Z, Huang Z, He W. Adaptive impedance control for an upper limb robotic exoskeleton using biological signals. IEEE Trans Ind Electron, 2017, 64: 1664-1674 CrossRef Google Scholar

[19] Boaventura T, Buchli J, Semini C. Model-based hydraulic impedance control for dynamic robots. IEEE Trans Robot, 2015, 31: 1324-1336 CrossRef Google Scholar

[20] Roveda L, Iannacci N, Vicentini F. Optimal impedance force-tracking control design with impact formulation for interaction tasks. IEEE Robot Autom Lett, 2016, 1: 130-136 CrossRef Google Scholar

[21] Yang C, Zeng C, Fang C. A DMPs-based framework for robot learning and generalization of humanlike variable impedance skills. IEEE/ASME Trans Mechatron, 2018, 23: 1193-1203 CrossRef Google Scholar

[22] Yang C, Zeng C, Cong Y. A learning framework of adaptive manipulative skills from human to robot. IEEE Trans Ind Inf, 2019, 15: 1153-1161 CrossRef Google Scholar

[23] Ajoudani A, Fang C, Tsagarakis N. Reduced-complexity representation of the human arm active endpoint stiffness for supervisory control of remote manipulation. Int J Robotics Res, 2018, 37: 155-167 CrossRef Google Scholar

[24] Buchli J, Stulp F, Theodorou E. Learning variable impedance control. Int J Robotics Res, 2011, 30: 820-833 CrossRef Google Scholar

[25] Calinon S, Kormushev P, Caldwell D G. Compliant skills acquisition and multi-optima policy search with EM-based reinforcement learning. Robotics Autonomous Syst, 2013, 61: 369-379 CrossRef Google Scholar

[26] Li Z, Zhao T, Chen F. Reinforcement learning of manipulation and grasping using dynamical movement primitives for a humanoidlike mobile manipulator. IEEE/ASME Trans Mechatron, 2018, 23: 121-131 CrossRef Google Scholar

[27] Rozo L, Silvério J, Calinon S. Learning controllers for reactive and proactive behaviors in human-robot collaboration. Front Robot AI, 2016, 3 CrossRef Google Scholar

[28] Duan J, Ou Y, Xu S, et al. Learning Compliant Manipulation Tasks from Force Demonstrations. In: Proceedings of the 2018 IEEE International Conference on Cyborg and Bionic Systems (CBS). IEEE, 2018. 449--454. Google Scholar

[29] Ganesh G, Albu-Sch??ffer A, Haruno M, et al. Biomimetic motor behavior for simultaneous adaptation of force, impedance and trajectory in interaction tasks. In: Proceedings of the 2010 IEEE International Conference on Robotics and Automation. IEEE, 2010. 2705--2711. Google Scholar

[30] Li Y, Ganesh G, Jarrasse N. Force, impedance, and trajectory learning for contact tooling and haptic identification. IEEE Trans Robot, 2018, 34: 1170-1182 CrossRef Google Scholar

  • Figure 2

    (Color online) The diagram of the biomimetic controller.

  • Figure 3

    (Color online) The set-up (a) and the results of the disturbance test of (b) joint S0, (c) joint S1, (d) joint E0, (e) joint E1, (f) joint W0, (g) joint W1, and (h) joint W2.

  • Figure 4

    (Color online) The experimental set-up for the sawing task. (a) The human partner actively pulls the saw. The impedance of the robotic arm increases gradually from a small value to some extent. (b) The robot actively pulls the saw back owing to the large impedance. The human partner relaxes his muscle strength in this phase.

  • Figure 5

    (Color online) The experimental results of the sawing task. (a) Joint S1; (b) joint E1; (c) joint W0; (d) joints S0, E0, W0, and W2.


    Algorithm 1 The online learning of the robotic compliant movement

    Require:The desired robot arm posture $(q_{0},~~\dot{q}_{0})$;

    Output: The computed joint torque command $~\tau_{c}^{t}~$ by (2) at each time step;

    Initialize the stiffness and damping matrix as $~K^{0}=\text{diag}\lbrace0,0,~\ldots,~0\rbrace~$ and $~D^{0}=\text{diag}\lbrace0,0,~\ldots,~0\rbrace~$;

    Initialize the feedforward vector as $~u^{0}=\text{diag}\lbrace0,0,~\ldots,~0\rbrace~$;

    Set the constant coefficients $\pi~$, $~\delta$, $a$, and $~b~$;

    Set the constant parametric vectors $~\alpha~$ and $~\beta~$;

    Set the stiffness range $~K^{\text{max}}~$ and $~K^{\text{min}}~$;

    for each time step $t~\in~[1,T]$

    Get the current robot joint states $~q~$ and $~\dot{q}~$;

    Compute the angle error and the velocity error according to (4);

    Compute the sliding error according to (8);

    Compute the vector $~\gamma~$ according to (16);

    Update the feedforward torque $~u^{t+1}=u^{t}~+~\Delta~u~$, using (15);

    Update the stiffness matrix $~K^{t+1}=K^{t}~+~\Delta~K~$, using (17);

    Adjust the stiffness values in a proper range based on (18);

    Compute the damping matrix $~D^{t+1}~$;

    Compute the impedance term $~v^{t+1}~$;

    Compute the joint torque $~\tau_{c}^{t+1}~$, using (2);

    Send the joint torque command to the robotic joint motors;

    end for

Copyright 2020 Science China Press Co., Ltd. 《中国科学》杂志社有限责任公司 版权所有

京ICP备17057255号       京公网安备11010102003388号