Cooperative phototaxis using networked mobile sensors and centroidal voronoi tessellations

摘要

The main objective for this research is to design an economical and robust swarm system to achieve phototaxis. The system combines swarm intelligence with centroidal Voronoi tessellations (CVT) to localize and track a dynamic moving light source. A small group of static mesh sensors is used to observe the environment while mobile actuators move according to their CVT. This paper also proposes a solution to decrease the number of sensors needed for CVT convergence with a light estimation algorithm. This new algorithm can find the approximate location of a light source by a general light distribution equation according to light sensor readings. Experiments are conducted both in simulation and on an actual mobile robot platform which show the flexible and robust nature of CVTs over other behavioral swarm algorithms. The complete phototaxis algorithm is presented in a specific swarm design procedure.