Adaptive Measurement and Decoding of Frequency-Hopping Spread Spectrum Signals Based on Knowledge Enhanced Compressed Sensing

The frequency-hopping spread spectrum (FHSS) signals can be sparsely represented in frequency domain at any given time. In this letter, we render a dictionary and propose non-cooperative adaptive compressed measurement and decoding methods of FHSS signals, where the measurement kernels (i.e. non-zeros coefficients in the sensing matrix) are adaptively designed based on the gradually obtained measurement results and the decoding can be done without the signal reconstruction step. Besides the ideal adaptive compressed method that achieves the best decoding accuracy with short measurement kernel design time periods, two alternative strategies enabling longer periods in the adaptive measurement design stages are also proposed for economic computational cost consideration. Simulations show the proposed methods to get improved decoding accuracy than the compared non-adaptive and state-of-art compressed methods.