A mutual positioning relay method of multiple robots for monitoring indoor environments

Abstract

This article proposes a mutual positioning relay method that enables multiple robots to monitor indoor environments. Here, robots refer to a small number of parent robots with high positioning performance and a large number of child robots with minimal positioning performance. The parent robot can estimate its state accurately by itself. In comparison, the child robots have position errors that accumulate with time as they use only odometry information for positioning. Each robot can recognize other robots in its field of view by using a depth camera. It then performs relative positioning and estimates its position of itself or that of other robots on the map. The uncertainty of the position estimation is compared with each other, and the robot with a certain position becomes a positioning reference for the robot with an uncertain position. In this way, both the parent and child robots relay relative positioning with each other, following which all the robots accurately estimate their positions. In this study, six robots based on Roomba were used. The performance of the proposed method was verified based on experimental data of autonomous navigation tests. Our results showed that the proposed method could realize an accuracy that was four to five times better than that performed independently by a child robot. Our findings also revealed that the proposed method can recover the failure of the state estimation of the parent robot.