Research On Spectrum Sensing Of Intelligent Network Based On Compressed Sensing

With the rapid development of wireless communication technology,the number of wireless users is increasing day by day.Users’ requirements for high-speed,high-bandwidth and other service quality of wireless communication are also increasing,and limited spectrum resources have become extremely valuable.Cognitive radio has become an effective solution to deal with low spectrum utilization and insufficient spectrum resources.Spectrum sensing is the most critical and basic process in cognitive systems,and the main purpose is to discover idle spectrum in fast,effective and reliable ways,making it possible for cognitive radio to perform dynamic spectrum access.The traditional spectrum sensing method has a huge sampling volume.In order to reduce the volume of data sensing and transmission in the cognitive wireless network,it is an effective solution by introducing compressed sensing technology into broadband spectrum sensing.Compressive sensing is a new method for acquiring and representing compressible signals with sub-Nyquist rate.It solves a sparsity-constrained process by exploiting the signal’s sparsity.Sparse or compressible signals can be reconstructed with fewer samples by introducing compressed sensing in cognitive radio systems.Using a set of wideband filters to measure the channel energy,only the energy of the channel representing the occupied state in successive time segments is needed to be recovered,instead of all the original signals in the channel.Although the method above can sample a signal at sub-Nyquist rate,it requires a large number of broadband filters for perceptual calculation and storage.Besides,it has high configuration complexity and large delay,which cannot meet the needs of fast and accurate detection of broadband signals.To satisfy the requirement of intelligent networks for efficient collection and transmission of spectrum data,this paper uses compressed sensing,semi-tensor product,deep learning and other theories to solve the problems of high transmission energy consumption,limited storage space of sensor nodes,and insufficient flexibility.This work is carried out focusing on the measurement matrix construction and the reconstruction algorithm optimization in compressed spectrum sensing.The main contributions and innovations of this paper are as follows:(1)A method of spectrum signal energy compression and reconstruction based on semi-tensor product is proposed to reduce energy consumption in transmission and storage.,This method proposes to replace traditional compression process with semi-tensor product compressed sensing in the process of collecting spectrum signals by secondary user.According to the characteristic that the dimension of the measurement matrix can be doubled by semi-tensor product,the proposed method could flexibly pick low-dimensional measurement matrix to compress the signal energy,so that reduces hardware resources required by the system.In addition,a parallel reconstruction mechanism is designed,and the semitensor product compressed spectrum sensing method can simultaneously perform the reconstruction process of channel energy on multiple compressed sensing decoders.Experimental results show that the proposed method improves the speed of the network node in sensing the spectrum usage and saves the required storage space;(2)A chaos-based spectrum detection and data transmission method is proposed to deal with the problems of information leakage and tampering caused by external intrusion in the process of information transmission of cognitive network nodes.By introducing chaos theory and making using of the initial value sensitivity of chaotic system,this paper proposes a method of generating compressed spectrum sensing measurement matrix based on chaotic parameters.In the absence of correct chaotic parameter package,attacker cannot recover correct channel occupancy results,which greatly improves the security of the cognitive intelligent network.Moreover,by only storing and transmitting the chaotic parameter package,the proposed method saves the storage and transmission energy consumption of the secondary user node in traditional compressed spectrum sensing algorithm;(3)A cognitive Internet of Things spectrum detection method based on residual network is proposed to improve the time efficiency of spectrum sensing algorithm in the reconstruction process.Considering the channel occupancy decision in compressed spectrum sensing as a binary classification problem,this paper uses residual network to recover signal energy.Based on residual network technique,an appropriate network model is constructed,and an adaptive threshold calculation method is designed to determine the channel occupancy.Simulation results show that the proposed method not only improves the accuracy of signal energy recovery,but also greatly reduces the computational delay of cognitive intelligent networks.