Robust Quadrupedal Locomotion on Sloped Terrains: A Linear Policy\n Approach

Abstract

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