Online adaptation of path formation in UAV search-and-identify missions
conference paper
In this paper, we propose a technique for optimisation and online adaptation of search paths of unmanned aerial vehicles (UAVs) in search-and-identify missions. In these missions, a UAV has the objective to search for targets and to identify those. We extend earlier work that was restricted to offline generation of search paths by enabling the UAVs to adapt the search path online (i.e., at runtime). We let the UAV start with a pre-planned search path, generated by a Particle Swarm Optimiser, and adapt it at runtime based on expected value of information that can be acquired in the remainder of the mission. We show experimental results from 3 different types of UAV agents: two benchmark agents (one without any online adaptation that we call 'naive' and one with predefined online behaviour that we call 'exhaustive') and one with adaptive online behaviour, that we call 'adaptive'. Our results show that the adaptive UAV agent outperforms both the benchmarks, in terms of jointly optimising the search and identify objectives. © 2011 Springer-Verlag.
Topics
design and engineering for self-adaptive systemssearch and identifyunmanned aerial vehiclesExpected value of informationOfflineOn-line adaptationParticle swarmPath formationRuntimessearch and identifySelf-adaptive systemAdaptive systemsOptimizationSpace shuttlesUnmanned aerial vehicles (UAV)Adaptive algorithms
TNO Identifier
429746
Source title
10th International Conference on Adaptive and Natural Computing Algorithms, ICANNGA 2011, 14 - 16 April 2011, Ljubljana.
Pages
186-195
Files
To receive the publication files, please send an e-mail request to TNO Repository.