Custom quadruped robot

with L Square Labs · CAD · 3D printing · MuJoCo · RL · 2025 – present
Summary video of the project.
12
joints, 3D-printed robot
0.3 m/s
untethered walking
0
falls in the recorded trials

Overview

I designed, 3D-printed and assembled a 12-joint quadruped. The walking policy is trained with reinforcement learning in MuJoCo and transferred to the robot. The robot walks untethered at 0.3 m/s and did not fall in the recorded trials.

Videos and images

Walking on the real robot.
The robot in a simulated scene reconstructed with Gaussian splatting.
The robot balancing on the floor.
The robot balancing on the floor.
First test with a single leg and inverse kinematics.
First test with a single leg and inverse kinematics.
Policy rollout in simulation.
Policy rollout in simulation.

Details

Hardware

The robot has 12 joints, three for each leg. I designed the leg mechanism and the body in CAD and printed the parts on a 3D printer. The first tests were done with a single leg on a stand to check the inverse kinematics.

Modelling

To make the simulation match the robot I measured the actuators. I sent chirp signals to the joints and fitted an actuator model to the response. The error of the fitted model was 0.061 rad (RMSE). The IMU sometimes stopped sending data for 1.9 to 3.7 seconds, so I added the same kind of gaps to the simulation during training.

Training and transfer

The walking policy is trained with reinforcement learning in MuJoCo and then run on the robot without further training. The robot walks untethered at 0.3 m/s. It did not fall in four trials at 0.3 m/s and one trial at 0.5 m/s.

Simulation with scanned scenes

I also tested the robot in simulation inside scenes reconstructed with Gaussian splatting.

Other projects

Sim-to-real manipulation with Gaussian-splat simulationStereo-inertial SLAM and monocular VIOContinual learning for manipulationManipulation control and teleoperationHumanoid manipulationHospital delivery and inspection robotSimulation projects