Fat-Bots in 2D

A gym like environment for fat-robots in a 2-D world.

---

Fatbot Swarm Gathering

https://github.com/Nelson-iitp/fatbot/blob/main/fig/fatbot.mp4

Sample World with 4 bots

https://github.com/Nelson-iitp/fatbot/blob/main/fig/vid_ppo_world_x4_1.mp4

Sample World with 5 bots

https://github.com/Nelson-iitp/fatbot/blob/main/fig/vid_ppo_world_x5_1.mp4

Sample World with 6 bots

https://github.com/Nelson-iitp/fatbot/blob/main/fig/vid_ppo_world_x6_1.mp4

World

Movement Model - Velocity and Speed

Sensor Data

Sensor can be a device like range-finder, distance-sensor, lidar, radar, sonar, depth-camera

• Sensor in fat-bot simulates 2 types of sensors:
  • X-ray : can detect center and boundary of all bots within its scan-distance (occlusion clearly detectable)
  • D-ray : (distance-ray) a simple range-finder device

Robot's View

Visibility Model

Sensor Data for Occluded bots

Red-Bot: all neighbours fully visible

Blue-Bot: Occlusion Detected

Green-Bot: Occlusion Detected

All Fully Visible