Design and implementation of a new detection system based on statistical features for different noisy channels

Day by day the frequency spectrum became unable to provide the service that user needs. The frequency spectrum became more crowded due to increasing the number of users. One of the solutions to minimize this problem is the cognitive radio (CR). This paper includes design and FPGA implementation of a new CR based on feature selection to increase the probability of detection (Pd) for different modulation systems. This design is based on statistical features that used to distinguish between signal and the noise. The system shows an excellent performance of a detection probability as compared with the traditional detection methods. Both AWGN and fading channels are tested in order to proof the performance of the proposed system. The obtained simulation results provide Pd equals to 100% at SNR is -18dB for 6000 sample. Also there are high similarity between the simulation and practical implementation. Xilinx Spartan-3A DSP 3400A is used to implement the proposed detection systems.