New Simulated Annealing algorithm for continuous target monitoring using multiple UAVs
Teams of cooperating UAVs have the potential to: accomplish the missions in a shorter period [1], simultaneously accomplish many goals, cost less, be less detectable and more survivable than a single large vehicle; damage to a single UAV does not necessarily cause the entire mission to fault. The higher degree of coordination helps us to achieve better overall efficiency even though the Unmanned Aerial Vehicles usually have limited sensor capabilities relative to their large counterparts. The accomplishment of goal therefore relies heavily on cooperation between appropriate neighbours. Cooperative target observation [2] is a very good example for study of multi agent cooperation. We use a new simulated annealing [3] for coordination and cooperation of unmanned aerial vehicles to continuously monitor a group of moving targets. We compare with Hill Climbing algorithm and the new simulated annealing algorithm and find that the new simulated annealing algorithm is superior for almost all target speeds, UAV sensor ranges and various group sizes. The fact that the Hill Climbing Algorithm is stuck up in local minima is indeed overcome by new simulated annealing algorithm. We found that the UAV is placed as close to as many targets as possible so that the UAV can move the target whenever the need arises. The future work would include obstacles; and to avoid those obstacles we intend to incorporate an obstacle avoidance algorithm. This model also has worked with the assumption of limitation of communication. However, our next step would be to include communication between the UAVs, which will enable us to achieve a more realistic model to achieve our goal.
References
[1] Moses, B., Jain, L.C., Finn, A. and Drake, S. Multiple UAVs Path Planning Algorithms: A Comparative Study, Journal of Fuzzy Optimization and Decision Making, Kluwer, Volume 7, 2008, pp. 257-267.
[2] Moses, B. and Jain, L.C., Cooperative Target Observation of UAVs using Simulated Annealing, International Journal of Intelligent Defence Support Systems, Volume 1, No 2, 2008, pp. 116-129.
[3] Leng, J., Sathyaraj Moses B. and Jain, L.C., Temporal difference Learning and Simulated Annealing for Optimum control: A Case study, Lecture Notes in Artificial Intelligence, LNAI 4953, KES-AMSTA 2008, Springer-Verlag, Germany, 2008, pp. 495-504.