Bayesian Optimization Using Domain Knowledge on the ATRIAS Biped @ICRA-cg8kk
Bayesian Optimization Using Domain Knowledge on the ATRIAS Biped  @ICRA-cg8kk
Uploaded May 2018 | Updated November 2026, 21 hours ago
ICRA 2018 Spotlight Video
Interactive Session Tue PM Pod H.4
Authors: Rai, Akshara; Antonova, Rika; Song, Seungmoon; Martin, William; Geyer, Hartmut; Atkeson, Christopher
Title: Bayesian Optimization Using Domain Knowledge on the ATRIAS Biped

Abstract:
Robotics controllers often consist of expert-designed heuristics, which can be hard to tune in higher dimensions. Simulation can aid in optimizing these controllers if parameters learned in simulation transfer to hardware. Unfortunately, this is often not the case in legged locomotion, necessitating learning directly on hardware. This motivates using data-efficient learning techniques like Bayesian Optimization (BO) to minimize collecting expensive data samples. BO is a black-box data-efficient optimization scheme, though its performance typically degrades in higher dimensions. We aim to overcome this problem by incorporating domain knowledge, with a focus on bipedal locomotion. In our previous work, we proposed a feature transformation that projected a 16-dimensional locomotion controller to a 1-dimensional space using knowledge of human walking. When optimizing a human-inspired neuromuscular controller in simulation, this feature transformation enhanced sample efficiency of BO over traditional BO with a Squared Exponential kernel. In this paper, we present a generalized feature transform applicable to non-humanoid robot morphologies and evaluate it on the ATRIAS bipedal robot, in both simulation and hardware. We present three different walking controllers and two are evaluated on the real robot. Our results show that this feature transform captures important aspects of walking and accelerates learning on hardware and simulation, as compared to traditional BO.
Bayesian Optimization Using Domain Knowledge on the ATRIAS BipedIncorporating Potential Contingency Tasks in Multi-Robot Mission PlanningModeling of Synthetic Fiber Ropes and Frequency Response of Long-distance Cable-Pulley SystemStability and Predictability in Dynamically Complex Physical InteractionsA Locomotion Recognition System Using Depth ImagesSAT-C: An Efficient Control Strategy for Assembly of Heterogeneous Stress-Engineered MEMS MicrorobotTalk Resource-Efficiently to Me: Optimal Communication Planning for Distributed Loop Closure DetectiScaling Inertial Forces to Alter Weight Perception in Virtual RealityIndoor Coverage Path Planning: Survey, Implementation, AnalysisUsing Information Invariants to Compare Swarm Algorithms and General Multi-Robot AlgorithmsArtistic Pen Drawing on an Arbitrary Surface Using an Impedance-Controlled RobotA Synchronization Scheme for Position Control of Multiple Rope-Climbing Robots
ICRA 2018 |

Bayesian Optimization Using Domain Knowledge on the ATRIAS Biped

SHARE TO X SHARE TO REDDIT SHARE TO FACEBOOK WALLPAPER