Evaluation of Methods for Estimating Noise Floor of MHD Angular Rate Sensors in Operation

Abstract

Noise floor plays a key role in evaluating the quality of MHD Angular Rate Sensors (MHD-ARSs). It determines the smallest angular rate signals that can be detected. Consequently, how to measure the noise floor accurately has become a critical issue. To the best of our knowledge, little research has been conducted into the noise floor measurement of MHD-ARSs. Moreover, noise floor estimators which have been used are not clearly defined and assessed. In this research, we aim at the evaluation of two methods for estimating noise floor of MHDARSs in operation, namely two-channel method and three-channel method. We present mathematical models and derive bias errors of both two methods. We set up a noise floor measurement system for two identical MHD-ARSs and test the two methods under the condition in which the two sensors operate in a high output SNR regime. We have obtained significant results demonstrating that: 1) Normalized bias error of two-channel method is approximately 80% to 100% in the case of SNR value greater than 10 dB. 2) Three-channel method exhibits no bias errors in theory, but it suffers from random errors originated from normalized random errors of PSD estimates.