(waiting for Update) Based on sumo develop an environemnt of vehicles avoidance pedestrians.
The code need install sumo https://sumo.dlr.de/docs/Downloads.php
cd /home/user/repos/CAV
run ppo2 with python naive_v6_1_1/run_sumo_new.py --env sumo_env-v6 --env_path /home/user/repos/CAV/envs/naive_v6_1_1/exp.sumocfg --reward_model 3 --seed 1 --num_workers 6 --gpu_index 0
run ppo2-safe-reward with python naive_v6_1_1/run_sumo_new.py --env sumo_env-v6 --env_path /home/user/repos/CAV/envs/naive_v6_1_1/exp.sumocfg --reward_model 4 --seed 1 --num_workers 6 --gpu_index 0
run cppo-pid with python naive_v6_1_1/run_sumo_new.py --env sumo_env-v6 --env_path /home/user/repos/CAV/envs/naive_v6_1_1/exp.sumocfg --reward_model 3 --seed 1 --algo cppo --num_workers 6 --gpu_index 0
run sac-safe-reward with python naive_v6_1_1/run_sumo_new.py --env sumo_env-v6 --env_path /home/user/repos/CAV/envs/naive_v6_1_1/exp.sumocfg --algo sac --reward_model 4
Use tensorboard to analyze the training process data.
Thanks for codes:
PPO: zhangchuheng123 https://github.com/zhangchuheng123/Reinforcement-Implementation/blob/master/code/ppo.py
SAC: ZenYiYan https://github.com/Yonv1943/ElegantRL/blob/master/AgentZoo.py
Code framework：ShangtongZhang https://github.com/ShangtongZhang/DeepRL/tree/master/deep_rl