A novel approach for trajectory tracking of UAVs

• Published in: Aircraft Engineering 2014-01, Vol.86 (3), p.198-206
• Main Authors: De Filippis, Luca, Guglieri, Giorgio, B. Quagliotti, Fulvia
• Format: Article
• Language: English
• Subjects: Aerospace engineering; Aerospace industry; Aircraft; Aircraft accidents; Aircraft control; Aircraft guidance; Algorithms; Collision avoidance; Collision dynamics; Commands; Engineering; Flight control systems; Genetic algorithms; Graphs; Horizon; Kinematics; Logic; Mathematical models; Nonlinear control; Obstacle avoidance; Optimal control; Optimization; Planning; Predictive control; Search algorithms; Studies; Task complexity; Tracking; Tracking errors; Tracking systems; Tracks (paths); Trajectories; Unmanned aerial vehicles; Unmanned aircraft
• ISSN: 1748-8842; 0002-2667; 1758-4213
• DOI: 10.1108/AEAT-01-2013-0016

Purpose – The purpose of this paper is to present a novel approach for trajectory tracking of UAVS. Research on unmanned aircraft is constantly improving the autonomous flight capabilities of these vehicles to provide performance needed to use them in even more complex tasks. The UAV path planner (PP) plans the best path to perform the mission. This is a waypoint sequence that is uploaded on the flight management system providing reference to the aircraft guidance, navigation and control system (GNCS). The UAV GNCS converts the waypoint sequence in guidance references for the flight control system (FCS) that, in turn, generates the command sequence needed to track the optimum path. Design/methodology/approach – A new guidance system (GS) is presented in this paper, based on the graph search algorithm kinematic A* (KA*). The GS is linked to a nonlinear model predictive control (NMPC) system that tracks the reference path, solving online (i.e. at each sampling time) a finite horizon (state horizon) open loop optimal control problem with genetic algorithm (GA). The GA finds the command sequence that minimizes the tracking error with respect to the reference path, driving the aircraft toward the desired trajectory. The same approach is also used to demonstrate the ability of the guidance laws to avoid the collision with static and dynamic obstacles. Findings – The tracking system proposed reflects the merits of NMPC, successfully accomplishing the task. As a matter of fact, good tracking performance is evidenced, and effective control actions provide smooth and safe paths, both in nominal and off-nominal conditions. Originality value – The GNCS presented in this paper reflects merits of the algorithms implemented in the GS and FCS. As a matter of fact, these two units work efficiently together providing fast and effective control to avoid obstacles in flight and go back to the desired path. KA* was developed from graph search algorithms. Maintaining their simplicity, but improving their search logics, it represents an interesting solution for online replanning. The results show that the GS uploads the collision avoidance path continuously during flight, and it obtains straightforward the reference variables for the FCS, thanks to the KA* model.