Integrated taskassignmentandpathoptimizationforcooperating uninhabitedaerialvehiclesusinggeneticalgorithms

The problemofintegratingtaskassignmentandplanningpathsforagroupofcooperatinguninhabited aerial vehicles,servicingmultipletargets,isaddressed.Intheproblemofinteresttheuninhabitedaerial vehiclesneedtoperformmultipleconsecutivetaskscooperativelyoneachgroundtarget.ADubinscar model isusedformotionplanning,takingintoaccounteachvehicle’sspecificconstraintofminimum turn radius.Byusingafinitesettodefinethevisitationangleofavehicleoveratargetweposethe integratedproblemoftaskassignmentandpathoptimizationintheformofagraph.Thisnewapproach results insuboptimaltrajectoryassignments.Refiningthevisitationanglediscretizationallowsforan improvedsolution.Duetothecomputationalcomplexityoftheresultingcombinatorialoptimization problem,weproposegeneticalgorithmsforthestochasticsearchofthespaceofsolutions.We distinguishbetweentwocasesofvehiclegroupcomposition:homogeneous,whereallvehiclesare identical;andheterogeneous,wherethevehiclesmayhavedifferentoperationalcapabilitiesand kinematicconstraints.Theperformanceofthegeneticalgorithmsisdemonstratedthroughsampleruns and aMonteCarlosimulationstudy.Resultsshowthatthealgorithmsquicklyprovidegoodfeasible solutions,andfindtheoptimalsolutionforsmallsizedproblems.