Multi-hop Communications in a Swarm of UAVs

Authors: 
Rachael Purta, Saurabh Nagrecha, Gregory Madey
Citation: 
Proceedings of the 2013 Symposium on Agent Directed Simulation (ADS'13/SpringSim 2013). Society for Computer Simulation International, San Diego, CA. 2013.
Publication Date: 
April, 2013

Unmanned Aerial Vehicles (UAVs) are of increasing interest to researchers because of their diverse applications, such as military operations and search and rescue. The problem we have chosen to focus on is using a swarm of small, in-expensive UAVs to discover static targets in a search space.Though many different swarm models have been used for similar problems, our proposed model, the Icosystem SwarmGame, to our knowledge has not been evaluated for this particular problem of target search.Further, we propose to simulate the performance of this model in a semi-realistic communications environment. The challenge here is to find the optimal multi-hop configuration for the UAVs, so that they can find the most targets, avoid collision with each other as much as possible, and still communicate efficiently. We implement this through a weighted shortest-path problem using Dijkstra’s algorithm, with the weights being the transmission cost over distance. Testing has shown that our multi-hop communications perform, in terms of target-finding and collision avoidance with other UAVs, as well as an idealized communications environment.