Dynamics filtering

We use a sampling-based trajectory optimization method (Mujoco MPC) as a dynamics filter to generate simulated trajectories that are dynamically feasible and closely track the recorded data.

mjpc-cube-run-1.mp4

Solid: Filtered trajectory Shadow: Raw demonstration

airpod-mjpc-01.mp4

airpod-mjpc-02.mp4

airpod-mjpc-03.mp4

bulb-mjpc-01.mp4

bulb-mjpc-02.mp4

bulb-mjpc-03.mp4

cube-flip-mjpc-01.mp4

cube-flip-mjpc-02.mp4

cube-flip-mjpc-03.mp4

key-lock-mjpc-01.mp4

key-lock-mjpc-02.mp4

key-lock-mjpc-03.mp4

nut-mjpc-01.mp4

nut-mjpc-02.mp4

nut-mjpc-03.mp4

peg-insertion-mjpc-01.mp4

peg-insertion-mjpc-02.mp4

peg-insertion-mjpc-03.mp4

peg-stand-mjpc-01.mp4

peg-stand-mjpc-02.mp4

peg-stand-mjpc-03.mp4

Page updated

Google Sites

Report abuse