A Mobile Wireless Channel State Recognition Algorihm: Introduction, Definition, and Verification - Sensing for Cognitive Environmental Awareness

Abstract

This research includes mobile wireless systems limited by time and frequency dispersive channels. A blind mobile wireless channel (MWC) state recognition (CSR) algorithm that detects hidden coherent nonselective and noncoherent selective processes is verified. Because the algorithm is blind, it releases capacity based on current channel state that traditionally is fixed and reserved for channel gain estimation and distortion mitigation. The CSR algorithm enables cognitive communication system control including signal processing, resource allocation/deallocation, or distortion mitigation selections based on channel coherence states. MWC coherent and noncoherent states, ergodicity, stationarity, uncorrelated scattering, and Markov processes are assumed for each time block. Furthermore, a hidden Markov model (HMM) is utilized to represent the statistical relationships between hidden dispersive processes and observed receive waveform processes. First-order and second-order statistical extracted features support state hard decisions which are combined in order to increase the accuracy of channel state estimates. This research effort has architected, designed, and verified a blind statistical feature recognition algorithm capable of detecting coherent nonselective, single time selective, single frequency selective, or dual selective noncoherent states. A MWC coherence state model (CSM) was designed to represent these hidden dispersive processes. Extracted statistical features are input into a parallel set of trained HMMs that compute state sequence conditional likelihoods. Hard state decisions are combined to produce a single most likely channel state estimate for each time block. To verify the CSR algorithm performance, combinations of hidden state sequences are applied to the CSR algorithm and verified against input hidden state sequences. State sequence recognition accuracy sensitivity was found to be above 99% while specificity was determined to be above 98% averaged across all features, states, and sequences. While these results establish the feasibility of a MWC blind CSR algorithm, optimal configuration requires future research to further improve performance including: 1) characterizing the range of input signal configurations, 2) waveform feature block size reduction, 3) HMM parameter tracking, 4) HMM computational complexity and latency reduction, 5) feature soft decision combining, 6) recursive implementation, 7) interfacing with state based mobile wireless communication control processes, and 8) extension to wired or wireless waveform recognition.