The deployment of autonomous vehicles in several domains requires efficient planning and tracking of trajectories that can satisfy constraints pertaining to the operating environment and the vehicle’s kinematics, while minimizing a suitable performance index. Catering to it, in this talk, we will discuss optimal planning and tracking of trajectories with the primary focus on applications like autonomous delivery systems, surveillance, and collision avoidance.
We will begin with the characterization and computation of an exact time-optimal trajectory between two oriented points via a moving circle for a non-linear Dubins vehicle model. Here, the moving circle may represent a moving obstacle or a neighborhood of interest. The choice of Dubins model is owing to the fact that it is a simple yet sufficiently realistic model to depict a simplified kinematics of fixed-wing unmanned aerial vehicles, cars, and underwater vehicles.
Thereafter, considering the scenarios with multiple vehicles, we will present a cooperative guidance law for collision avoidance among a heterogeneous team of vehicles in an n-on-n planar engagement scenario. For computational tractability, in this case, we will discuss the use of a linearized engagement model and minimization of team effort of the vehicles as the suitable performance index.
Finally, for tracking a planned trajectory we will present the application of game-theoretic approaches for precise and robust tracking of aggressive trajectories by multirotor platforms. The applicability of this approach for Dubins vehicle model will also be demonstrated.
The talk will also include results from experimental validation of the proposed control methods.
