GPU-based Concurrent Multi-UAVs Path Flow Analysis using K-MST Algorithm

摘要

Unmanned Aerial Vehicles have widespread applications in the fields of remote sensing, reconnaissance, product delivery, and search & rescue. With the advent of new technologies the manufacturing cost of typical UAVs have gone down whereas on the contrary the robustness and safety factors have improved drastically. Hence multiple UAV systems are being planned for coordinating specific large tasks such as search & rescue and delivering commercial products door to door. Many factors affect the optimized operation of these multiple flying machines in a given designated geographical area. The most crucial being the challenge of chartering an optimal flight path for each UAV based on distance, terrain and fuel costs. Subsequently, cumulative flying time costs are calculated and a final scenario is presented highlighting the optimal usage of multiple UAVs working in unison. The prime focus is on finding localized shortest path using Prim's algorithm and then clustering these localized paths into smaller sub-graphs optimally. Multicore processors are leveraged to provide accelerated computations as the paths change dynamically with respect to time.