Abstract

This thesis deals with two control problems for unmanned vehicles, namely battery shortage prevention for electric unmanned vehicles and collision avoidance strategy for constrained multi unmanned vehicle systems.
In the battery shortage prevention problem, we design a novel control architecture, equipped with a battery manager module, capable of avoiding energy shortage by appropriately imposing time-varying upper bounds on the vehicle's maximum acceleration. Here, the dual-mode control paradigm known as set-theoretic model predictive control is applied to couple the reference tracking and the battery shortage problems. First, we offline design a conservative maximum acceleration profile capable of assuring that the electric unmanned vehicle will reach the desired target without incurring into a battery shortage along the given path. Then, online, by following a receding horizon approach, we show that the battery manager can enhance the performance by using the current battery's state-of-charge. Moreover, a simulation example is presented to clarify and show the proposed control framework's potential and features.
In the collision avoidance problem, we deal with vehicles moving in a shared environment where each UV follows a trajectory given by a local planner. We assume that the planners are uncoordinated and each vehicle is subject to different constraints and disturbances. In this context, we design a new centralized traffic manager that, in conjunction with ad-hoc designed local model predictive controller, can ensure the absence of collisions while minimizing the total vehicle's stops occurrences. In particular, in a receding horizon fashion, the traffic manager exploits available previews on the successive vehicle's waypoints to speed-up or speed-down the vehicles and minimize the chance of collisions. Moreover, by exploiting basic set-theoretic arguments, traffic manager can impose a vehicle to stop and safely prevent collisions whenever necessary. Finally, two different simulation examples are presented to better illustrate the capability of the proposed solution.