Noise Attenuation on IMU Measurement For Drone Balance by Sensor Fusion

Abstract

Stability is the key to maintain and control the drone, which is challenged by significant noise from drone motors during operation. The paper presents the Kalman filter and Complementary filter based on the quaternion to optimize drone stability. An exponential moving average (EMA) filter is used to minimize the significant vibration noise inside angular rates. The designed models optimize the misleading data from the Inertial Measurement Unit (IMU) sensor on the drone caused by noise. A real test bench was constructed to verify the proposed methods. An MPU 6050 (triaxial accelerometer and triaxial gyroscope) is equipped with a Racing Drone; then, the sensor data is logged in a MicroSD Card for signal analysis. The results demonstrate that the Complementary filter attenuates variation due to the noise, but it has an issue with drift. On the other hand, the Kalman filter accomplishes more stable output surrounding the drone's balanced point.