Verification of Self-Position Estimation with m5go for Small Autonomous Mobile Robot

In this study, we investigated the accuracy of self-position estimation for a small autonomous mobile robot using the M5GO accelerometer, which estimates position by dead reckoning after removing noise from the M5GO acceleration data. Noise was removed by offset correction and setting a threshold value, and the accuracy of self-position estimation was confirmed when M5GO was moved 200 mm. As a result, although the noise could be removed when the M5GO stopped, it continued to move because the sum of the acceleration at the moment it moved and the acceleration at the moment it stopped did not equal zero. This may be due to the effect of gravitational acceleration and the size of the threshold value. Therefore, M5GO's self-position estimation needs to be improved.