E2DRL: End-to-End Differentiable Reinforcement Learning for 3D Obstacle Avoidance in Quadruped Robots

Differentiable reinforcement learning shows strong promise for robot policy learning, but its application to legged locomotion has remained limited to blind walking or simulated 2D tasks. Extending it to end-to-end visual obstacle avoidance is challenging because hard collision constraints and early termination upon contact corrupt first-order gradients, trapping optimizers in local minima.

From the abstract. Our summary is in progress.