The Method Of Attitude Acquisition For Walking Robot Under Vibration Environment

Abstract:Walking robot needs High accuracy attitude for its motion control and planning. This paper designs a method to acquire high accuracy attitude data by inertial navigation equipment for walking robot, this method could compensate the random error of micromechanical gyroscope under vibration environment, at the same time, it also could provide reliable attitude data for the quadruped bionic robot and help balance the robot.First of all, this paper analysis the quadruped robot’s vibration sources which is a typical representative of walking robot. And next we analysis the micromechanical gyroscope including its basic structure and its process of production, by that way, we know its measurement principle and acquire the basic kinetic equation. This paper also analysis the vibration environment which the quadruped bionic robot faces when it is working, for that factor, we could build an experimental platform which could guarantee the basic experiment which we could obtain the high-precision micromechanical gyroscope attitude data.Secondly, this paper designs a data acquisition system which collects the MEMS gyroscope data. At the same time does a detailed analysis of the impact of vibration environment to the MEMS gyroscope output data, analysis its random error model and its manifestations, in order to acquire further information, we use the Allan variance and discrete Fourier transform to analysis the output data of MEMS gyroscope. Above all, it could do a great help for subsequent modeling of random errors.We use the original sample data which collects under vibration environment to model random error using time series method, we also test the sample data for its trend term, its stability and normal detection to improve the accuracy of the model. Bases on the AIC and FPE principles, we model the ARMA prediction model, at last, verify the validity of the random prediction model by comparing with the actual output data. Finally, we use the Kalman filter and ARMA prediction model to compensate the output data of MEMS gyroscope under vibration environment, according to the comparison results with the XW-GI5700, the method of Kalman filter bases on the ARMA model could effectively compensate the MEMS gyroscope’s angle information under vibration environment.