MilliSwarm: Leveraging Emergence for Energy Efficient Robotic Swarm Movement

Abstract

INCOSE Systems Vision 2035 outlines the need for greater understanding of both complex system dynamics and efficient systems. Energy efficiency and coordination in multi-agent systems is a key focal point of current research. Finding solutions to group movement that balance energy savings and group cohesion provide unique benefits to the system. While current approaches focus on improving the efficiency of individual swarm members (a reductionist approach), we approach this challenge through using biologically inspired design to harness emergence. Thus, the hypothesis examine is: “If rolling swarm inspired movement is applied to ground swarms encountering obstacles, then the system's overall energy efficiency will improve because cooperation in both climbing and moving reduces the energy required to operate in these environments.” In this article a method for moving groups of agents together as a team is presented, inspired by the behaviors of millipede swarms. The individual agents represent robots that are abstracted into uniform sliding blocks. Using simple behaviors based on individual knowledge, rather than global information sharing, we present a new efficient way to travel in groups. Data shows up to a 96% decrease in energy spent in certain scenarios for systems using the proposed MilliSwarm movement design. Additionally, this algorithm proves to be increasingly efficient at larger swarm sizes, while not requiring additional computing resources for coordination. Future work in this project will work on applying these functions to 3D environments, real-world case studies, and additional non-physical system implementation.