Modeling, Simulation, and Design of a Time-Delayed Multi-Agent System of Autonomous Vehicles

With the advent of cheap wireless communication and advancement in cloud computing, controllers are continually being placed off the systems they stabilize. A physical distance between the controller and plant brings in a new issue for modeling and designing control environments. Time delays slow the exchange of information to non-negligible values and must be accounted for when implementing new control systems. When considering mobile robotic platforms, a delay can lead to the mobile agents falling off cliffs, driving in restricted areas, or lose flight stability. A solution to account for the time delays will be presented by accounting for a set update interval before the controller is updated in a discrete fashion.

Presented in this thesis is a thorough look at a single unmanned ground vehicle and the maximum bound on the update time of a controller before the system becomes unstable. The ground vehicle is controlled first by a state feedback controller and then with observer based feedback and each maximum bound is found and simulated. A new environment is set up to include a flying robot to send the ground vehicle points of interest through a wireless network to test the validity of the controller given another vehicle increasing the burden on the wireless network.