An FPGA testbed for spectrum sensing in cognitive radio

Abstract

Listen

Translate article

In this paper we present a research testbed for spectrum sensing in cognitive radio systems. We consider spectrum sensing based on Eigen values of sample covariance matrix of the input signal. When considering the hardware implementation of the algorithm, it will pose a question on how to calculate the Eigen values with low complexity. In practice, there is a tradeoff between complexity and detection performance. In theory, the Eigen values of the sample covariance matrix can be found by using a Discrete Fourier Transform (DFT), which can be easily implemented in hardware using Fast Fourier Transform (FFT). Simulations are performed in MATLAB and real-time measurements carried out in Xilinx Kintex-7 Field Programmable Gate Array (FPGA). The result shows that the probability of detection increases as the Signal to Noise Ratio (SNR) increases. The result from hardware shows a less performance compared to the simulation, that is because the threshold is being calculated from equation(5) and not obtained from the distribution of the received signal under hypothesis 0. In terms of the computational complexity, the Eigen values method which proposed in [6], is about M times that of the Number of samples Ns, where M is the autocorrelation length of the received signal. However our proposed method only needs 2∗M LUT(Look Up Tables) from the hardware resources as shown in table 3. The simulations and implementation show that the proposed method works well with low complexity, without using prior information of the primary user's signal, the channel and noise power compared to the current spectrum sensing state of the art.