Wideband Compressed Sensing based Frequency Hopping Radio Network Identification

Multiple frequency hopping (FH) spread spectrum signals are challenging to detect, track and differentiate into distinct FH radio networks using conventional Nyquist rate samplers due to their wideband frequencies usage. In this work, a methodology to identify different FH radio networks operating in an overlapping frequency range is proposed. Wideband spectrum composed of multiple FH radio signals is sensed using sub-Nyquist rate sampling based on multi-coset sampler. The sampled signals are reconstructed in a segmented manner using accelerated iterative hard thresholding algorithm. Identification of the radio networks is performed based on the temporal, power, frequency statistical distribution, FH pattern and spectrum coverage features extracted from sufficient successive reconstructed sample segments. Multiple features extraction allows accurate identification of the FH radio networks and formation of Markov model for tracking of frequency usage in a particular radio network. Numerous simulations are performed to identify FH radio networks operating within the same spectrum.