Regression-Based Position Detection for Navigation using IMU

Abstract

In this study, IMU (Inertial Measurement Unit) data was used and GPS (Global Positioning System) position detection was performed. Fine Tree Regression has been used for position detection. A new dataset has been collected to test the proposed method. The development board has been designed with Arduino mega, Altimu-10 IMU sensor, GPS module, and SD card module. This development board is fixed on an automobile. During the movement of the car, altitude, accelerometer, gyro, magnetometer, latitude, and longitude data have been collected. Obtained data have been used together with Fine Tree Regression. Both latitude and longitude regression results are calculated. Because of regression, latitude, and longitude results have been calculated according to the car's altitude, accelerometer, gyro, and magnetometer data. 0.97 R-Squared for latitude and 0.98 R-Squared for longitude have been computed with Fine Tree Regression. Obtained results have been compared with other regression models.