English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Zendy Plus

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Zendy Tools

Zendy Open

Model-free reinforcement learning (RL) for legged locomotion commonly relieson a physics simulator that can accurately predict the behaviors of everydegree of freedom of the robot. In contrast, approximate reduced-order modelsare commonly used for many model predictive control strategies. In this work weabandon the conventional use of high-fidelity dynamics models in RL and weinstead seek to understand what can be achieved when using RL with a muchsimpler centroidal model when applied to quadrupedal locomotion. We show thatRL-based control of the accelerations of a centroidal model is surprisinglyeffective, when combined with a quadratic program to realize the commandedactions via ground contact forces. It allows for a simple reward structure,reduced computational costs, and robust sim-to-real transfer. We show thegenerality of the method by demonstrating flat-terrain gaits, stepping-stonelocomotion, two-legged in-place balance, balance beam locomotion, and directsim-to-real transfer.

GLiDE: Generalizable Quadrupedal Locomotion in Diverse Environments with a Centroidal Model