Lightweight and minimalistic openai gym environment for differential drive robot with 2D lidar. This environment does not require any sophisticated simulator and is meant for running on cloud python notebook based servers like Google Colaboratory.
Number of lidar beams can be changed inside the environment file. All obstacles are circular and spawned randomly on each episode. Obstacles are not allowed to intersect, this is taken care of during spawning.
The goal point is always (0.0, 0.0)
Actions: array [ linear velocity (range: [0.0, 1.0]), angular velocity (range:[-2.0, 2.0]) ]
Vector of size = 2
Observations: array [ lidar ranges(all beams), straight line distance to goal, angular deviation between current heading and straight line towards goal ]
Vector of size = num of lidar beams + 2
Make sure your PYTHONPATH environment variable has the path to the directory of this package
In your code
import gym_diff_drive.envs.diff_drive_lidar16Use the following as environment id:
DiffDriveLidar16-v0