Research And Application Of Attitude Angle Estimation Algorithm Based Onquaternion And Kalman Filter

The first application about pose estimation is in aerospace and military field, and is the most mature area in these fields. The volume of tradition inertial sensors(gyroscope and accelerometer) are usually very big, high cost, difficult to carry and has been effected the application fields. With the developing of MEMS technology, especially MEMS inertial sensor appear, the application range is becoming more and more larger, for example in biomedical field: endoscopic posture positioning and virtual-reality field: scene interaction. The relationship between one coordinate and another can be described as pose, that is yaw, pitch and roll. The key problem is attitude computation and the precise of attitude angle, our work carrying out also based on the two problem.There are many ways to describe attitude, for example Euler angle, direction cosine and quaternion. Euler angle describe attitude there will be a singularity problem, direction cosine method calculation complex, quaternion method has been widely adopted, our paper is based on quaternion method, runge kutta method is used to solve differential equation of quaternion to get attitude angle, realize the attitude algorithm. Due to the nature of MEMS gyroscope, the output data contain constant error and the error of the random drift, which adopt the MEMS gyroscope inertial sensor cannot achieve precise angle measurement for long time; Similarly, the accelerometer based on gravity field to calculate the horizontal angle(pitch angle and roll angle), and carrier is movement in practical application, will introduce the linear acceleration, cause horizontal angle calculation is not accurate, magnetometer based on the earth’s magnetic field to calculate the yaw angle, but as a result of geomagnetic field are susceptible to interference, then calculate the yaw angle is not too accurate. Therefore, at present widely used multi-sensor information fusion to estimate attitude angle, on the basis of predecessors fusion algorithm, we put forward our schemes: using kalman filtering algorithm for three-axis gyroscope, three-axis accelerometer and three-axis magnetometer for information fusion, to get the optimal estimated angle. First of all, compensation and filtering processing the sensor collected data, then calculation the angle of gyroscope as estimate value, the angle of the accelerometer and magnetometer as measured value, by using kalman algorithm achieve the optimal attitude angle estimation. From actual test and verify, prove its validity.Finally, according to the design of the algorithm, using nine-axis sensor MPU-9150 and Arduino Pro mini designed a attitude detection device that can be worn on one’s waist, realizing detect the attitude of fall down, through actual test proves its reliability. As an engineering practical writing, attitude estimation is applied to detect the attitude of fall down, promote attitude estimation in other areas of application, has the certain application value and practical significance.