Research On The Switching Strategy Of Data Fusion Method For Multi-mems Gyro

Micro-electromechanical system(MEMS)gyroscope is an important inertial device.It has developed rapidly in recent years and has also been used in practice.However.MEMS gyroscopes still have defects such as large noise and low measurement accuracy.Using wavelet domain multi-scale data fusion algorithm,the accuracy and reliability of the groscope can be effectively improved through signal fusion of multiple MEMS gyroscopes.Though the fusion value of the method is time-delayed compared with the original data when the method deals with issues such as the gyro dynamic data and jump data.Therefore,how to make full use of the advantages of wavelet domain multi-scale data fusion algorithm in processing gyro static data,and ensure the real-time performance of the system in the dynamic,so as to improve the overall output accuracy of the system is particularly important.For this reason,this paper adopts a data fusion method switching strategy.The main work of the full text includes:1.The basic principle of wavelet domain multi-scale data fusion algorithm and the advantages of multi-scale fusion are elaborated in detail.The problems of wavelet domain multi-scale data fusion algorithm in processing dynamic gyro data are emphatically analyzed.2.Based on the analysis of the problems in processing dynamic gyro data in the wavelet domain multi-scale data fusion algorithm and the common time series mutation detection method,the method of gyro signal mutation detection which combines the cumulative sum control chart algorithm with the preset threshold is determined.The feasibility of the method was verified with the measured gyro data.3.Considering that the gyro signal processing in dynamic state needs strong adaptability and strong ability to track the dynamic change signal,it is determined that the optimal weighted recursive least squares fusion method is used to process the gyro jump data.The weight factors are introduced to measure the reliability of gyro sensors in this method.The effectiveness of the data fusion method switching strategy is verified by experiments on MEMS gyros under static and dynamic conditions respectively.The experimental results suggest that the fusion method switching strategy adopted in this paper can well solve the time-delay problem caused by jump data when using wavelet domain numlti-scale data fusion algorithm.The real-time performance of the gyro signal processing and the frequency response of the system have been significantly improved.The results of static and dynamic experimental show that the noise of the fusion value is significantil smaller than the gyro original data noise,and the output accuracy of the system has also been improved.In summary,from the perspective of static and dynamic experiments,the fusion method switching strategy is effective and feasible.