Research And Implementation Of Adaptive Dual-threshold Wavelet Packet Entropy Spectrum Detection

In recent years,wireless intelligent terminal devices are more and more widely used in society,and the spectrum resources required for wireless communication are not inexhaustible,so society is faced with the urgent need to solve the problem of spectrum resource scarcity.Nowadays,countries generally adopt fixed static spectrum allocation policy,which may lead to the phenomenon that the authorized frequency band is idle and the non-authorized users who need it are not available,and the spectrum resources are wasted greatly.Cognitive Radio(CR)is an important method to overcome this difficulty,which has been studied extensively by many scholars.Spectrum detection technology is the first step of CR,and plays a vital role in CR research,so it makes sense to improve the detection performance of spectrum detection technology.In this paper,a wavelet packet entropy spectrum detection algorithm based on adaptive double-threshold is proposed,which reduces the whole computation and has good detection ability under the condition of low signal-to-noise ratio and uncertain noise,and it is verified on the hardware platform.The research work of this paper is summarized as follows:Firstly,on the basis of introducing the spectrum detection model,the classical single-point spectrum detection and cooperative spectrum detection are introduced in detail,and the merits and drawbacks of different spectrum detection techniques are contrasted and analyzed.Meanwhile,the difficult problems of spectrum detection technology are summarized.Secondly,the advantages and disadvantages of Double Threshold Energy Detection(DTED)and Wavelet Packet Entropy Detection(WPED)are emphatically introduced.The former takes into account the uncertainty of noise,however,it is not effective and robust in the case of low SNR,which has the characteristics of multi-resolution and metric signal,and can adapt to the low SNR environment,but has high computational complexity.Combining these two algorithms,an improved spectrum detection algorithm is proposed in this paper,which improves the performance and reduces the computational load.Through MATLAB simulation,the three algorithms are compared under different conditions,which proves that the new algorithm proposed in this paper has more advantages.Finally,on the hardware platform based on FPGA and DSP,it is designed to achieve adaptive double-door wavelet entropy spectrum detection function.First of all,on the basis of the requirements analysis,the spectrum sensing system developed in this paper is presented.The FPGA mainly implements DTED,and the characteristic of parallel operation makes FPGA more suitable for Fast Fourier transform(FFT).Because the detection of wavelet packet entropy spectrum needs a lot of multiplication,division,addition and subtraction and logarithm operation,so DSP is used to complete it.In order to solve the problem of half invalid operation caused by convolution and then extraction,the method of alternate point convolution is proposed.The high-speed data transmission interface SRIO is selected to complete the tansmission between FPGA and DSP,and the SRIO module is introduced in detail.The system function is verified on the hardware platform,and the comparison between the engineering results and MATLAB simulation results proves that the system has good detection performance.