Effectiveness of Synchronization and Cooperative Behavior of Multiple Robots based on Swarm AI

Abstract

We investigated the effectiveness of Swarm AI, which allows multiple artificial intelligence (AI) robots to autonomously collaborate and achieve more advanced functions as a group. We built two models of robot groups, each equipped with its own AI, and ran simulations. The first one showed that multiple robots can sense the situation of another party and acquire synchronous behavior by reinforcement learning. The second one clarified that the efficiency of cleaning improved by 30-40 % as a result of introducing cooperative behavior based on Swarm AI into the group of cleaning robots. These findings demonstrate that the cooperation of multiple robots using Swarm AI is effective.