Predictable Motion in Unpredictable Domains: The Spotlight Tracking Problem
Kimberly H. Chung, Venkatesh G. Rao, Raffaello D’Andrea · AIAA Guidance, Navigation, and Control Conference and Exhibit · 2006
We introduce the spotlight tracking problem, the problem of having a vehicle stay within a moving tolerance constraint boundary along a nominal path while avoiding obstacles. The formulation models situations where multivehicle coordination must be achieved based on pre-mission commitments, due to communication and observation constraints. The problem is analytically intractable, and we present and evaluate flve heuristic solutions that incorporate varying amounts of known or estimated path information to solve the problem. Performance assessment is in terms of maximum velocity, acceleration, and smoothness of the velocity proflles generated by the flve methods in simulation. The results demonstrate that a tracking method that incorporates both local geometric information and estimates of total path length performs the best. Analysis results concerning velocity and acceleration bounds are also presented.