Signal detection with quantization noise based on compressive sensing acquisition system

To promote the efficient use of spectrum resources, cognitive radio (CR) has been proposed. Spectrum sensing is the core technology of CR. In spectrum sensing, idle frequency bands should be detected for secondary users. Signal detection in spectrum sensing is a recent active research topic. Conventional spectrum sensing methods are based on Nyquist sampling and have high hardware requirements, thus limiting the use of these methods. The proposed compression sampling theory can overcome the limitation of Nyquist sampling and can solve this problem. In high-speed sampling, the effect of quantization noise is becoming more significant. In this paper, we study the signal detection in compressed sampling under the influence of quantization noise. Since the quantization noise is related to the quantization bits, we derive the relationship between the quantization bits and the detection rate (PD). We also consider the input noise and investigate its impact on signal detection. Due the amplification of input noise in compressive sampling process, in comparison to the quantization noise, the input noise exhibits more serious impact on detection performance.