An implementation of "Bridging hamilton-jacobi safety analysis and reinforcement learning" by Jaime F Fisac, Neil F Lugovoy, Vicenç Rubies-Royo, Shromona Ghosh, Claire J Tomlin
For simplicity, states are just sampled uniformally.
Example:
python hj_dqn_brt.py --num-epochs 10_000 -batch-size 16_000 --exp-name dubins3d
- python=3.8
- jax
- optax
- flax
- wandb
- tqdm
- optimized_dp