Vision-based control of intelligent robot networks

Abstract

It is observed that multi-robot systems are cheaper and more effective when compared to a single robot. However, the effectiveness of multi-robot systems in uncertain environments appears to lag natural formation control counterparts such as a swarm of ants and a flock of birds. This is because mobile robots are usually programmed and tested in a controlled environment and rely heavily on global positioning and orientation data. Thus, in this project, formation control, obstacle avoidance and loss of vision correction algorithms are designed with the consideration of uncertain environment. The control algorithms mainly use vision-based data such as LIDAR sensor and cameras for control. MATLAB and Simulink software were used for the testing and simulation of the control algorithms designed.