A Compensation Method for Random Error of Gyroscopes Based on Support Vector Machine and Beetle Antennae Search Algorithm

Gyroscope is a common inertial sensor for inertial navigation system (INS). Its accuracy is affected by various error sources and noises. Among them, the deterministic error can be eliminated by calibration. However, random error is difficult to be evaluated due to its uncertainty, so random error has a great negative impact on the accuracy of gyroscope. Therefore, in order to improve the precision of gyroscope, it is necessary to compensate the random error. This paper proposes a compensation method for random error of gyroscopes based on support vector machine (SVM) and beetle antennae search (BAS) algorithm. Firstly, the original time series data of random error is processed in groups by a slide window, and then the SVM model is trained by using the original data of random error. The two key parameters in SVM are optimized by BAS algorithm. Finally, The trained SVM model is used to predict the gyroscopes random error data, and the prediction results are used to complete the compensation of gyroscopes data. Experimental results show the effectiveness of the proposed scheme, and the compensation accuracy is better than the traditional autoregressive integrated moving average (ARIMA) method.