UAV Swarm Mission Planning Development Using Evolutionary Algorithms and Parallel Simulation - Part II

摘要

The purpose of this paper is to discuss the design and implementation of comprehensive mission planning systems for swarms of autonomous aerial vehicles (UAV). Such a system could integrate several problem domains including path planning, vehicle routing, and swarm behavior as based upon a hierarchical architecture. The example developed system consists of a parallel multi-objective evolutionary algorithm-based terrain-following parallel path planner, a multi-objective evolutionary algorithm (MOEA) for the UAV swarm router, and a parallel simulation. Generic objectives include minimizing cost, time, and risk generally associated with a three dimensional vehicle routing problem (VRP). The concept of the Swarm Routing Problem (SRP) as a new combinatorics problem for use in modeling UAV swarm routing is presented as a variant of the Vehicle Routing Problem with Time Windows (VRPTW). Various multi-objective VRPTW routing benchmarks result in very good Pareto-based performance with the MOEA which is also reflected in the results of the new SRP benchmarks. The culmination of this effort is the development of an extensible developmental path planning model integrated with swarm routing behavior and tested with a parallel UAV simulation. Discussions of this system's capabilities are presented along with recommendations for generic development of UAV swarm mission planning.