Health & Environmental Research Online (HERO)


Print Feedback Export to File
7180196 
Journal Article 
Heuristic and Genetic Algorithm Approaches for UAV Path Planning under Critical Situation 
Arantes, J; Arantes, M; Motta Toledo, CF; Trindade Junior, O; Williams, BC; , 
2017 
Yes 
International Journal on Artificial Intelligence Tools
ISSN: 0218-2130 
WORLD SCIENTIFIC PUBL CO PTE LTD 
SINGAPORE 
26 
1 
WOS:000397081200009 
The present paper applies a heuristic and genetic algorithms approaches to the path planning problem for Unmanned Aerial Vehicles (UAVs), during an emergency landing, without putting at risk people and properties. The path re-planning can be caused by critical situations such as equipment failures or extreme environmental events, which lead the current UAV mission to be aborted by executing an emergency landing. This path planning problem is introduced through a mathematical formulation, where all problem constraints are properly described. Planner algorithms must define a new path to land the UAV following problem constraints. Three path planning approaches are introduced: greedy heuristic, genetic algorithm and multi-population genetic algorithm. The greedy heuristic aims at quickly find feasible paths, while the genetic algorithms are able to return better quality solutions within a reasonable computational time. These methods are evaluated over a large set of scenarios with different levels of difficulty. Simulations are also conducted by using FlightGear simulator, where the UAV's behaviour is evaluated for different wind velocities and wind directions. Statistical analysis reveal that combining the greedy heuristic with the genetic algorithms is a good strategy for this problem. 
Genetic algorithms; unmanned aerial vehicles; path planning; risk allocation; uncertainty