A Heterogeneous Aerial Platform Mission Planner using a Genetic Algorithm
Jonathan Rojas · 2015
Systems exist today that can plan a mission with more than one aircraft efficiently for surveillance. However, objectives in these missions do not change and are typically performed using a homogeneous set of aerial vehicles. An adaptive mission planner was sought to task a heterogeneous set of Unmanned Aerial Vehicles (UAVs) when an unknown Target of Interest (TOI) is located amongst a set of Points of Interest (POIs). First, two dimensional flight path models of fixed wing and quadcopter platforms were created. Next, the design of a genetic algorithm and its fitness functions were studied. Fixed wing fitness functions were developed to balance POI task loads amongst a set of fixed wing aircraft. A quadcopter fitness function was then designed to task a quadcopter to visit a newly located TOI. The quadcopter fitness function was also designed to maximize battery usage as it was desired that the quadcopter visit as many additional POIs on route to and from the TOI. Case studies were then simulated using varying heterogeneous UAV sets and TOI locations. Results of these simulations were then analyzed using mission times as a performance metric. Simulation results indicated that the deployment of the quadcopter to the TOI and additional POIs reduced overall mission times. Mission time reductions were also found to be depended on the number of fixed wing aircraft used in heterogeneous UAV sets.