CARTPOLE-V1 PHYSICAL AI LAB

Gymnasium [Robotics Dynamics]
PPO Neural Policy vs Optimal LQR Dynamics Suite | HWIHWA LAB
PHYSICAL ENGINE ONLINE
ACTIVE CONTROLLER: TRAINED PPO (20K)
BENCHMARK: 500.0 (SOLVED)
🤗 Model Hub GitHub
PHYSICAL AI · 60 FPS DYNAMICS
⚡ FORCE IMPULSE APPLIED!
ACTUATOR: ±10.0 N
EQUILIBRIUM: 100% | DT: 20ms (50Hz)
SCORE
0 / 500
⚡ DRAG / CLICK TO DISTURB AI
CONTROLLER ARENA & MISSION CONTROL
4 ARCHITECTURES
⚡
CART POS (X) [-2.40m ~ +2.40m]
+0.00 m
CART VEL (Vx) [-3.00m/s ~ +3.00m/s]
+0.00 m/s
POLE ANGLE (θ) [-12.0° ~ +12.0°]
+0.0°
ANGULAR VEL (θ̇) [-3.00r/s ~ +3.00r/s]
+0.00 r/s
⚖️ CONTROL LAW PPO MLP [4➔64➔64➔2]
⚡ ACTION DUTY 50% L / 50% R
📡 SWITCH FREQ 14.2 Hz
SIM-TO-REAL DYNAMICS & PERFORMANCE BENCHMARK
ROBOTICS TUNER
POLE LENGTH (L) 0.50 m
POLE MASS (Mp) 0.10 kg
CART MASS (Mc) 1.00 kg
GRAVITY (g) 9.81 m/s²
PLANETARY PRESET: 9.81 m/s² [🌍 EARTH]
PHASE ATTRACTOR [θ vs θ̇]
θ: +0.0° θ̇: +0.00r/s
STABLE SPIRAL CONVERGENCE
RADIUS: 0.02 rad
EPISODE REWARD & 10-MA
ONLINE
SCORE 0
BEST 500 ★
EQUILIBRIUM 100%
EPISODE 1