Multi-robot systems are becoming increasingly popular and are being used in more and more applications such as warehouses, agriculture, and transportation. The objective of this project is to control and coordinate a group of ground robots. Controlling a single ground robot requires planning the desired path and computing actuation inputs to maneuver the robot based on the feedback of its state. Given a set of interchangeable tasks for a multi-agent system, such as a group of robots, the first decision that needs to be made is to determine which agent does what, e.g., which robot goes to which location of interest.
The aim of this project is to assemble, interconnect, and program a group of JetBots such that every predefined task is completed by one robot. The tasks are given by goal locations that need to be visited. The robots can obtain information about their surroundings with an onboard camera and detect objects by processing the images with a neural network. The project will consist of setting up and calibrating hardware as well as implementing machine learning, control, and task assignment methods in software. It will also involve interfacing the robots with a central computer and a visual motion capture system. The project will culminate in experiments that demonstrate the implemented task assignment and robot control methods.
Requirements and key learnings:
The student is required to have good knowledge of control theory and basic programming skills (C++ or Python). Experience with combinatorial optimisation and ROS (Robot Operating System) is beneficial. Within this project, the student will learn how to build a robotic system, apply machine learning, and derive multi-agent control algorithms.
This project can be taken as a master project (PDM) by extending its scope to the investigation of cutting-edge collision avoidance algorithms.
Background reading:
->https://jetbot.org/master/,
->https://journals.sagepub.com/doi/full/10.1177/0278364913515307.
Interested students should contact Tony Wood.