Graph based algorithms for swarm intelligence with applications to MAVs

Alan Won-Ku Kim, Purdue University

Abstract

An individual micro aerial vehicle within a swarm of identical micro aerial vehicles can be programmed with a proper behavior logic which allows the swarm to carry out complicated surveillance and reconnaissance missions efficiently while maintaining communications with a command center. A hybrid-system approach to designing an obstacle avoidance behavior logic for such a swarm is presented in this study. An efficient algorithm for reduced visibility graph construction based on a general graph containing both convex and concave polygons is proposed. Also a method for the use of synchronized reduced visibility graph information for swarm flight path optimization is presented. Finally, a communication retention logic for the swarm is proposed. Both the obstacle avoidance algorithm and the flight path optimization method along with the communication constraint have been simulated and the results show that all of the proposed algorithms are computationally efficient and result in a desirable aggregate behavior of the MAV swarm.

Degree

M.S.A.A.

Advisors

Hwang, Purdue University.

Subject Area

Aerospace engineering|Robotics

COinS