MJWarp Simulation Compares Robot Stability Across MLP Angles
Split-screen video posted on X shows two legged robots running deterministic policies in MJWarp simulation.
@baketnk_en posted a video on X showing a split-screen real-time MJWarp simulation of two identical robotic figures. The left agent runs a deterministic policy at an 8° MLP angle and achieves 99.22% survival over 510 routes. The right agent runs at a 10° MLP angle. YacineMTB, a research engineer focused on reinforcement learning, retweeted the post with the comment duckpole.
Combined views
3.1K
2 posts, first seen 5h ago
MJWarp Simulation Compares Robot Stability Across MLP Angles
Split-screen video posted on X shows two legged robots running deterministic policies in MJWarp simulation.
@baketnk_en posted a video on X showing a split-screen real-time MJWarp simulation of two identical robotic figures. The left agent runs a deterministic policy at an 8° MLP angle and achieves 99.22% survival over 510 routes. The right agent runs at a 10° MLP angle. YacineMTB, a research engineer focused on reinforcement learning, retweeted the post with the comment duckpole.
