Interactive Dynamic Walking: Learning Gait Switching Policies with Generalization Guarantees

In this paper, we consider the problem of adapting a dynamically walking bipedal robot to follow a leading coworker while engaging in tasks that require physical interaction. Our approach relies on switching among a family of Dynamic Movement Primitives (DMPs) as governed by a supervisor. We train the supervisor to orchestrate the switching among the DMPs in order to adapt to the leader’s intentions, which are only implicitly available in the form of interaction forces. The primary contribution of our approach is its ability to furnish certificates of generalization to novel leader intentions for the trained supervisor. This is achieved by leveraging the Probably Approximately Correct (PAC)-Bayes bounds from generalization theory. We demonstrate the efficacy of our approach by training a neural-network supervisor to adapt the gait of a dynamically walking biped to a leading collaborator whose intended trajectory is not known explicitly.

[1]  Sushant Veer,et al.  Steering a 3D limit-cycle walker for collaboration with a leader , 2017, 2017 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS).

[2]  Frank L. Lewis,et al.  Intent aware adaptive admittance control for physical Human-Robot Interaction , 2015, 2015 IEEE International Conference on Robotics and Automation (ICRA).

[3]  Gintare Karolina Dziugaite,et al.  Computing Nonvacuous Generalization Bounds for Deep (Stochastic) Neural Networks with Many More Parameters than Training Data , 2017, UAI.

[4]  Tad McGeer,et al.  Passive Dynamic Walking , 1990, Int. J. Robotics Res..

[5]  Anirudha Majumdar,et al.  PAC-Bayes control: learning policies that provably generalize to novel environments , 2018, Int. J. Robotics Res..

[6]  Sushant Veer,et al.  Probably Approximately Correct Vision-Based Planning using Motion Primitives , 2020, ArXiv.

[7]  Shuzhi Sam Ge,et al.  Human–Robot Collaboration Based on Motion Intention Estimation , 2014, IEEE/ASME Transactions on Mechatronics.

[8]  Hikaru Inooka,et al.  Variable impedance control of a robot for cooperation with a human , 1995, Proceedings of 1995 IEEE International Conference on Robotics and Automation.

[9]  David A. McAllester.  Some PAC-Bayesian Theorems , 1998, COLT' 98.

[10]  Jessy W. Grizzle,et al.  Toward Safety-Aware Informative Motion Planning for Legged Robots , 2021, ArXiv.

[11]  Sushant Veer,et al.  Adaptation of limit-cycle walkers for collaborative tasks: A supervisory switching control approach , 2017, 2017 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS).

[12]  David Vogt,et al.  Inferring guidance information in cooperative human-robot tasks , 2013, 2013 13th IEEE-RAS International Conference on Humanoid Robots (Humanoids).

[13]  Ryojun Ikeura,et al.  Cooperative motion control of a robot and a human , 1994, Proceedings of 1994 3rd IEEE International Workshop on Robot and Human Communication.

[14]  Paul Evrard,et al.  Human-Humanoid Co-working in a Joint Table Transportation , 2012, ICSR.

[15]  Sushant Veer,et al.  Composing limit cycles for motion planning of 3D bipedal walkers , 2016, 2016 IEEE 55th Conference on Decision and Control (CDC).

[16]  Christine Chevallereau,et al.  From stable walking to steering of a 3D bipedal robot with passive point feet , 2012, Robotica.

[17]  Csaba Szepesvári,et al.  PAC-Bayes with Backprop , 2019, ArXiv.

[18]  François Keith,et al.  Human-humanoid haptic joint object transportation case study , 2012, 2012 IEEE/RSJ International Conference on Intelligent Robots and Systems.

[19]  Sushant Veer,et al.  Input-to-State Stability of Periodic Orbits of Systems With Impulse Effects via Poincaré Analysis , 2019, IEEE Transactions on Automatic Control.

[20]  Tom Schaul,et al.  Natural Evolution Strategies , 2008, 2008 IEEE Congress on Evolutionary Computation (IEEE World Congress on Computational Intelligence).

[21]  Kyungnam Kim,et al.  Variable Admittance Control With Virtual Stiffness Guidance for Human–Robot Collaboration , 2020, IEEE Access.

[22]  Arno H. A. Stienen,et al.  Admittance control for physical human–robot interaction , 2018, Int. J. Robotics Res..

[23]  Clément Gosselin,et al.  General Model of Human-Robot Cooperation Using a Novel Velocity Based Variable Impedance Control , 2007, Second Joint EuroHaptics Conference and Symposium on Haptic Interfaces for Virtual Environment and Teleoperator Systems (WHC'07).

[24]  Sushant Veer,et al.  Switched Systems With Multiple Equilibria Under Disturbances: Boundedness and Practical Stability , 2020, IEEE Transactions on Automatic Control.

[25]  Bowen Weng,et al.  Robust Feedback Motion Policy Design Using Reinforcement Learning on a 3D Digit Bipedal Robot , 2021, 2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS).

[26]  Sushant Veer,et al.  Safe Adaptive Switching among Dynamical Movement Primitives: Application to 3D Limit-Cycle Walkers , 2018, 2019 International Conference on Robotics and Automation (ICRA).

[27]  Sushant Veer,et al.  Almost driftless navigation of 3D limit-cycle walking bipeds , 2017, 2017 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS).

[28]  Aaron D. Ames,et al.  An Inverse Dynamics Approach to Control Lyapunov Functions , 2020, 2020 American Control Conference (ACC).

[29]  Jun Nakanishi,et al.  Dynamical Movement Primitives: Learning Attractor Models for Motor Behaviors , 2013, Neural Computation.

[30]  Aaron Ames,et al.  Global Position Control on Underactuated Bipedal Robots: Step-to-step Dynamics Approximation for Step Planning , 2020, 2021 IEEE International Conference on Robotics and Automation (ICRA).

[31]  E. Westervelt,et al.  Feedback Control of Dynamic Bipedal Robot Locomotion , 2007 .