Electromechanical inertial sensors test and analysis method based on identification using adaptive noise cancellation

Abstract

Inertial systems design and performance analysis critically depend on the development of the inertial sensors test and analysis. The requirements for the wide use of electromechanical inertial sensors necessitate the exploitation of the test and analysis method that allows optimal elimination of the disturbance in the test over the whole frequency band of interest. A method based on identification using adaptive noise cancellation is proposed to process the sensors’ signal. In theory, there is no correlation between the signal components reflecting the sensors noise and disturbance from the test environment. The semi-physical simulation based on the identification of sensor ontology is established to pick up the noise in the test as the noise background. The normalized least-mean-squares (NLMS) algorithm is applied to abstract the sensors output as adaptive noise canceller. The experiments of dynamically tuned gyroscopes and quartz accelerometers demonstrate the validity of the method, besides the results of tests in different bandwidths after denoising are consistent in accuracy level. Additionally, its implementation in different grades of inertial sensors is presented to support the effectiveness of the algorithm and the importance of bandwidth in the test.