Simple Self Driving Car using Reinforcement Learning

By Nishat Parveen

To understand how Reinforcement Learning is being used to make autonomous driving possible, it is important to implement the fundamentals in the most basic model of a self driving car.

Simple_road environment

A very basic scenario, but useful to apply and understand RL concepts. The environment is designed similarly to openai gym env

Dependencies

Create a conda environment: conda create --name rl-for-driving-car python=3.7

Install the packages: pip install -r requirements.txt

Example of settings

After training, env_configuration.json is generated to summarize the configuration.

{
  "min_velocity":0,
  "previous_action":null,
  "initial_state":[
    0,
    3,
    12
  ],
  "max_velocity_2":2,
  "state_ego_velocity":3,
  "obstacle1_coord":[
    12,
    2
  ],
  "actions_list":[
    "no_change",
    "speed_up",
    "speed_up_up",
    "slow_down",
    "slow_down_down"
  ],
  "goal_velocity":3,
  "goal_coord":[
    19,
    1
  ],
  "previous_state_position":0,
  "obstacle":null,
  "initial_position":[
    0,
    0
  ],
  "previous_state_velocity":3,
  "state_features":[
    "position",
    "velocity"
  ],
  "state_obstacle_position":12,
  "obstacle2_coord":[
    1,
    3
  ],
  "rewards_dict":{
    "goal_with_bad_velocity":-40,
    "negative_speed":-15,
    "under_speed":-15,
    "action_change":-2,
    "over_speed":-10,
    "over_speed_near_pedestrian":-40,
    "over_speed_2":-10,
    "per_step_cost":-3,
    "goal_with_good_velocity":40
  },
  "max_velocity_1":4,
  "max_velocity_pedestrian":2,
  "using_tkinter":false,
  "state_ego_position":0,
  "reward":0
}