Formation forming based low-complexity swarms with distributed processing for decision making and resource allocation
Jian Yang, Xin Wang, Peter H. Bauer · 2016
This paper introduces a new paradigm that combines in-swarm sampling/processing together with formation forming methods. The introduced concept is capable of generating a regular sampling lattice through emergent behavior of swarm agents and, in a second step, exploits this formation using 2-D filtering and processing algorithms based on the so-called multidimensional Roesser model. The swarm system scales well and the low complexity approach requires minimal hardware. Formation control is achieved through artificial potential fields generated globally by a few higher tier agents and locally by the massive amount of lower tier agents. The paper illustrates the proposed methods by applying line forming to a swarm example that scans an unknown area while performing in-swarm distributed filtering with subsequent marking of potential target areas. For the first time in the literature, a modified distributed 2-D Roesser model is combined with formation forming in order to achieve the desired results.