Body Randomization Reduces the Sim-to-Real Gap for Compliant Quadruped Locomotion

Alexander Vandersompele, Gabriel Urbain, Hossain Mahmud, Francis Wyffels and Joni Dambre

Frontiers in Neurorobotics,

March 2019 · doi:10.3389/fnbot.2019.00009


Designing controllers for compliant, underactuated robots is challenging and usually requires a learning procedure. Learning robotic control in simulated environments can speed up the process whilst lowering risk of physical damage. Since perfect simulations are unfeasible, several techniques are used to improve transfer to the real world. Here, we investigate the impact of randomizing body parameters during learning of CPG controllers in simulation. The controllers are evaluated on our physical quadruped robot. We find that body randomization in simulation increases chances of finding gaits that function well on the real robot.

subject terms:robotics, human brain project, HBP, neurorobotics, neuromorphics, brain simulation, spiking neural networks, NRP, robot simulation