A Study on Accuracy of Self-Position Estimation of Mobile Robots by Sensor Fusion

Abstract

Currently, the aging of many infrastructures and the decrease in the number of workers due to the declining birthrate and aging population are becoming more serious, so it is necessary to reduce work force and improve the efficiency of measurement work using ground-penetrating radar. High accuracy of self-position estimation is required for the expected automation of inspection work by autonomous mobile robots. In a previous study, we attempted to improve the accuracy of self-position estimation by using sensor fusion of an encoder and an IMU sensor on a robot equipped with a mecanum wheel as a high-precision position estimation. As a result, although the accuracy was improved, it did not reach a satisfactory level. One of the factors was vibration and wheel slippage. In this study, we analyze the correlation between the vibration data and the position estimation accuracy by sensor fusion, and investigate a correction method.