Method and associated algorithm and firmware for extracting magnetometer signals from noise utilizing digital signal processing (DSP)

摘要

A compact, computationally-efficient, firmware-resident version of a DSP autocorrelation function, extracts periodic information from noise-corrupted signals. The function adapts autocorrelation parameters in real-time either autonomously or via up-link commands from a ground-based system. The function calculates specific determination values that optimize the performance of the autocorrelation process for a specific application. The per-bit multiplication used by the function can be performed by logically ANDing each of the respective samples in the two arrays. The summation of the per-bit multiplication is an integer summation, rather than the more time-consuming floating-point summation. A further aspect of the present function is the utilization of a threshold detection process to discretize the correlation output estimate to a one-bit binary sample.