An Extended Potential Field Controller for use on Aerial Robots
AuthorWoods, Alexander Cromwell
AdvisorLa, Hung M
StatisticsView Usage Statistics
This thesis focuses on the design and implementation of an extended potential field controller (ePFC) which enables a quadcopter aerial robot to track a dynamic target while simultaneously avoiding obstacles in the environment. The design of the ePFC extends the foundational concepts of a traditional potential field controller (PFC), which uses attractive and repulsive potential fields to navigate toward a target and avoid obstacles. A traditional PFC is a function of only the relative positions of the drone to the target and obstacles, respectively, and has shortcomings for aerial robots which are much harder to control than ground robots. The proposed ePFC takes into account the relative velocities of the drone to the target and obstacles, respectively, in addition to the relative positions which enhances the controller’s ability and improves performance. The proposed controller is simulated using Matlab’s Simulink tool, and the simulation results show that the ePFC reduces the overshoot of the robot’s location in response to a step input by 19% and the settling time by nearly 17% when compared to a traditional PFC. The proposed controller is implemented on an experimental platform, the ARDrone 2.0, and the obtained results show that the drone is able to track both static and dynamic targets, moving in either set or arbitrary patterns, all while avoiding obstacles in the test space. Compared to the simulation, the experimental results show an overshoot 2% higher, and a settling time only 0.5 sec slower.