Assignment of Cooperating UAVs to Simultaneous Tasks using Genetic Algorithms
Abstract
A problem of assigning multiple unmanned aerial vehicles (UAVs) to simultaneously perform cooperative tasks on consecutive targets is posed as a new NP-hard combinatorial optimization problem. The investigated scenario consists of multiple ground moving targets prosecuted by a team of heterogeneous UAVs carrying designated sensors and/or weapons. To successfully prosecute each target it first needs to be simultaneously tracked by multiple UAVs, from significantly different line of sight angles to reduce the position estimate errors, and then attacked by a different UAV carrying a weapon. Even for small sized scenarios, the problem has prohibitive computational complexity for classical combinatorial optimization methods due to timing constraints on the simultaneous tasks and the coupling between task assignment and path planning for each UAV. A genetic algorithm (GA) is proposed for efficiently searching the space of feasible solutions. A matrix representation of the GA chromosomes simplifies the encoding process and the application of the genetic operators. To further simplify the encoding, the chromosome is composed of sets of multiple genes, each corresponding to the entire set of assignments on each target. Simulation results conform the viability of the proposed assignment algorithm for different sized scenarios. The sensitivity of the performance to variations in GA tuning parameters is also investigated.
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 18, 2005
- Accession Number
- ADA445125
Entities
People
- Corey Schumacher
- Tal Shima
Organizations
- Air Force Research Laboratory