Robust Quadrupedal Locomotion on Sloped Terrains: A Linear Policy Approach

Kartik Paigwar, Lokesh Krishna, sashank tirumala, naman khetan, aditya varma, ashish joglekar, Shalabh Bhatnagar, Ashitava Ghosal, Bharadwaj Amrutur, Shishir Kolathaya
Proceedings of the 2020 Conference on Robot Learning, PMLR 155:2257-2267, 2021.

Abstract

In this paper, with a view toward fast deployment of locomotion gaits in low-cost hardware, we use a linear policy for realizing end-foot trajectories in the quadruped robot, Stoch 2. In particular, the parameters of the end-foot trajectories are shaped via a linear feedback policy that takes the torso orientation and the terrain slope as inputs. The corresponding desired joint angles are obtained via an inverse kinematics solver and tracked via a PID control law. Augmented Random Search, a model-free and a gradient-free learning algorithm is used to train this linear policy. Simulation results show that the resulting walking is robust to terrain slope variations and external pushes. This methodology is not only computationally light-weight but also uses minimal sensing and actuation capabilities in the robot, thereby justifying the approach.

Cite this Paper

BibTeX
@InProceedings{pmlr-v155-paigwar21a, title = {Robust Quadrupedal Locomotion on Sloped Terrains: A Linear Policy Approach}, author = {Paigwar, Kartik and Krishna, Lokesh and tirumala, sashank and khetan, naman and varma, aditya and joglekar, ashish and Bhatnagar, Shalabh and Ghosal, Ashitava and Amrutur, Bharadwaj and Kolathaya, Shishir}, booktitle = {Proceedings of the 2020 Conference on Robot Learning}, pages = {2257--2267}, year = {2021}, editor = {Kober, Jens and Ramos, Fabio and Tomlin, Claire}, volume = {155}, series = {Proceedings of Machine Learning Research}, month = {16--18 Nov}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v155/paigwar21a/paigwar21a.pdf}, url = {https://proceedings.mlr.press/v155/paigwar21a.html}, abstract = {In this paper, with a view toward fast deployment of locomotion gaits in low-cost hardware, we use a linear policy for realizing end-foot trajectories in the quadruped robot, Stoch 2. In particular, the parameters of the end-foot trajectories are shaped via a linear feedback policy that takes the torso orientation and the terrain slope as inputs. The corresponding desired joint angles are obtained via an inverse kinematics solver and tracked via a PID control law. Augmented Random Search, a model-free and a gradient-free learning algorithm is used to train this linear policy. Simulation results show that the resulting walking is robust to terrain slope variations and external pushes. This methodology is not only computationally light-weight but also uses minimal sensing and actuation capabilities in the robot, thereby justifying the approach.} }
Endnote
%0 Conference Paper %T Robust Quadrupedal Locomotion on Sloped Terrains: A Linear Policy Approach %A Kartik Paigwar %A Lokesh Krishna %A sashank tirumala %A naman khetan %A aditya varma %A ashish joglekar %A Shalabh Bhatnagar %A Ashitava Ghosal %A Bharadwaj Amrutur %A Shishir Kolathaya %B Proceedings of the 2020 Conference on Robot Learning %C Proceedings of Machine Learning Research %D 2021 %E Jens Kober %E Fabio Ramos %E Claire Tomlin %F pmlr-v155-paigwar21a %I PMLR %P 2257--2267 %U https://proceedings.mlr.press/v155/paigwar21a.html %V 155 %X In this paper, with a view toward fast deployment of locomotion gaits in low-cost hardware, we use a linear policy for realizing end-foot trajectories in the quadruped robot, Stoch 2. In particular, the parameters of the end-foot trajectories are shaped via a linear feedback policy that takes the torso orientation and the terrain slope as inputs. The corresponding desired joint angles are obtained via an inverse kinematics solver and tracked via a PID control law. Augmented Random Search, a model-free and a gradient-free learning algorithm is used to train this linear policy. Simulation results show that the resulting walking is robust to terrain slope variations and external pushes. This methodology is not only computationally light-weight but also uses minimal sensing and actuation capabilities in the robot, thereby justifying the approach.
APA
Paigwar, K., Krishna, L., tirumala, s., khetan, n., varma, a., joglekar, a., Bhatnagar, S., Ghosal, A., Amrutur, B. & Kolathaya, S.. (2021). Robust Quadrupedal Locomotion on Sloped Terrains: A Linear Policy Approach. Proceedings of the 2020 Conference on Robot Learning, in Proceedings of Machine Learning Research 155:2257-2267 Available from https://proceedings.mlr.press/v155/paigwar21a.html.

Related Material