A Fault Prediction Method of Quartz Flexible Accelerometers Based on AGO-RVM

A novel fault prediction method for quartz flexible accelerometers of inertial navigation systems is presented. Firstly, accumulated generating operation (AGO) is conducted on the original data sequence of the scale factor stability to enhance its regularity. Then, relevance vector machine (RVM), which has good quality in terms of prediction precision and generalization, is applied to the data sequence achieved by AGO. Moreover, the RVM model is updated continuously by a metabolism mechanism to improve the adaptivity of the prediction method. The gray relational analysis (GRA) is adopted to decide whether to update the RVM model. The performance of the proposed method is verified by the accelerated life test, and the experimental results show it can achieve high accuracy on the fault prediction of quartz flexible accelerometers.