Research Of Attitude Estimation Algorithm For Quad-Rotor Aircraft

Quad-rotor aircraft is a sort of unmanned aerial vehicles which can take off and landing vertically and hover independently,it has been widely used in military and civil field.The speed and accuracy of attitude estimation can affect the stability of quad-rotor aircraft directly.Aiming at the problem of attitude estimation of a quad-rotor,a hybrid filter algorithm was proposed to estimate attitude.First,this article introduces basic principle of attitude estimation,and take quaternion algorithm as the way to update attitude angle,then deduce the whole calculation process of quaternion algorithm.Having analysis on the working principle of gyroscope,accelerometer and magnetometer,one-order low pass filter was designed to process the original data collected from gyroscope,Kalman filter was designed to process the original data collected from accelerometer,a simple calibration method was taken to process the original data collected from magnetmetor.Based on quaternion algorithm,a hybrid filter algorithm that combine the cross product of vectors and complementary filter was proposed to fuse the data collected from accelerometer,gyroscope and magnetometer,taking advantage of vector deviation of accelerometer and magnetmetor to modify the integration accumulated error of gyroscope during the process of attitude estimation.At last,establishing experimental platform of attitude estimation to test and verify hybrid filter algorithm.The experimental platform mainly includes main controller module,posture sensor module,wireless communication module,battery module,motor driver module and brushless motors.Experimental results show that,attitude angle estimated through hybrid filter algorithm can stabilize near the initial attitude angle effectively,the maximum error is not more than 0.2°,hybrid filter algorithm can modify the integration accumulated error of gyroscope effectively.In addition,hybrid filter algorithm’s average time of one iteration filtering is much shorter,it can reduce computational complexity,improve the operation efficiency and improve accuracy of attitude estimation at the same time.