Minimum Time Search in Unmanned Aerial Vehicles using Ant Colony Optimisation based Realistic Scenarios

Abstract

Unmanned aerial vehicles (UAV), or drones, are aircrafts without a human pilot on board. UAVs find a target in minimum time using Minimum Time Search (MTS) methods. Different optimisation paradigms, such as cross-entropy optimisation (CEO) and ant-colony optimisation (ACO) can be used for MTS. In this work, a set of simulation scenarios has been designed to test the ACO solution to the MTS problem. Simulations performed for each scenario take into account a heuristic function and its effect on the probability of detection of target and estimated time for detection. The results obtained for various scenarios based on external and internal factors in UAV trajectory planning (size of search grid, target distribution, etc.) are compared to categorise the best set of such factors across four input domains. Results show a huge variance in the role played by the heuristic function and choice of feature thresholds for each scenario.