Constructing IMU based On ADC and sensors calibration for Ballbot

Abstract

In this article the Ballbot described as being unlike a statically stable wheeled robot, is shown to be a dynamic stable robot. The stability of such inherently unstable robot is based on accurate angle and position feedback that is obtained from his IMU. We did not use commercial IMU in this Ballbot construction unlike the previous Ballbots, instead an IMU that is suitable for Ballbot mechanism was designed and implemented. Angle feedback produced by IMU must be as accurate as possible, similar to other sections of robot. Because of this reason we must calibrate sensors for best performance. But before sensors calibration, ADC calibration must be done in order to correct errors occurring during conversion of the sensor analog signals to equivalent digital numbers. By a calibrated ADC, sensor calibration is performed. A precise calibration method which is simple to apply is presented. The method does not need accurate equipment and is performed by a plate attached to an ordinary motor. After acquiring calibrated data, the Kalman filter is implemented for data fusion. The Kalman filter output is very accurate and IMU can track abrupt changes in orientation. It is implemented in Ballbot and could provide proper feedback to balance the Ballbot.