We had to program an Autonomous Ground Vehicle (AGV) to make it capable of autonomously traversing an indoor environment to assist moving objects from one place to another. The purpose of the bot could be effective as a bot in a warehouse, which could pick up specific objects from one place and put it at another specific location, thus reducing the amount of manpower required.
Making the AGV involved the following tasks:
- Understanding ROS (used ROS Melodic in Ubuntu 18.04)
- 2D Mapping for Autonomous Navigation (ROS Navigation Stack)
- 3D Mapping for Arm Manipulation (ROS Manipulation Stack)
- Recognizing specific Objects (ROS Perception Stack)
- Pick and Place of Objects
- Finally, working all of the above together in a reallife simulation!
Video link for simulation of the AGV in Gazebo: https://youtube.com/playlist?list=PLCh5SBkKMYlupPjIG5wdsTn_vT5Os9wWy
Here's a glimpse of how the AGV looked like:
