Enhancing MEMS sensors accuracy via random noise characterization and calibration

This paper presents a design concept that can be used to monitor Micro-Electro-Mechanical Systems (MEMS) inertial sensors' random noise characteristics and dynamically track them for cancellation. The concept consists of a two-prong compensation approach offering both filtering and cancellation capability to effectively null out the MEMS sensor noise sources. The first path compensation will be fundamentally designed using high order filtering and calibration concept. This path is intended to effectively calibrate and remove high noise drift errors inherently existing in the MEMS sensors by using external aiding sensors data available on-board the spacecraft such as star tracker or GPS sensors. MEMS sensors' bias, scale factor, and misalignment stability errors will all be taken care of using this first prong design approach. The second compensation system will be designed using signal isolation and stochastic model propagation concept allowing on-line MEMS sensor's noise estimation and characterization. This second path is intended to dynamically monitor changes and identify MEMS inertial sensors' random noise parameters such as scale factor error, angular random walk, angular white noise, and rate random walk in a real-time fashion so that proper noise spectrum signatures can be obtained to update the process noise matrix of the calibration filter. This latter design approach can also be applied and implemented as a signal-conditioning device for MEMS sensors' internal self-calibration. The proposed algorithm is provided along with its preliminary results evaluated using simulation.