Standard recognising self organizing map based cognitive radio transceiver

Abstract

Cognitive Radio (CR) Systems have transceivers with the ability to adjust their operating parameters after observing the results in order to decide to operate in a specific radio configuration, expecting to move the radio towards some optimized operational state. Framed within this statement, this paper introduces and evaluates the use of Self-Organizing Maps (SOM) and Linear Vector Quantization (LVQ), which are both Unsupervised Neural Network (NN) Schemes as an effective technique for reconfiguring transceivers after recognizing the specific standard based on input parameters extracted from the signal. We exploit the inherent property of SOM, tonotopy, to recognize the standard in our proposed multi-standard cognitive transceiver. We target at cognition incorporation in our Software Defined Radio (SDR) waveform, designed on a Network-on-chip (NoC) based multi-core, single chip Xilinx Virtex-4 FPGA.