Study On Physical Layer Security Based On Compressed Sensing

Wireless sensor networks in the Internet of things have the characteristics of limited memory,computing and resources,which makes the traditional cryptography(such as RSA algorithm and AES algorithm)encryption with high computational complexity and more energy consumption unsuitable for this.At the same time,these sensor networks may perform critical tasks,playing an important role.Therefore,it is necessary to ensure the security of data transmission.Physical layer security uses orthogonal frequency division multiplexing(OFDM)technology,which has a bright future in protecting data transmission of current and future network.OFDM is the basic component of multi-carrier modulation in modern networks,such as the Internet of things,vehicular ad-hoc network and 4G/5G systems.Most existing security solutions of OFDM-based systems lack confidentiality and flexibility when combining secret keys with random information extracted from physical channels.In a variety of security solutions,the combination of compressed sensing and physical layer security is a very effective approach to achieve security.Compressed sensing,as a new physical layer security technology,breaks the traditional Nyquist sampling theorem which is regarded as the golden rule.According to CS theory,if the signal is sparse or shows sparsity in a certain domain,less data can be used to recover the original data at the receiver.The application of compressed sensing technology to channel estimation of OFDM-based systems can effectively save frequency band resources and then improves the performance of systems.Firstly,this paper studies the physical layer security schemes,analyzes the characteristics and problems of the existing schemes,and focuses on the basic principle of OFDM technology and the basic theory of compressed sensing.Through studying and learning these basic theoretical knowledge,laid a theoretical foundation for the follow-up work.Secondly,aiming at the PLS issues of key generation rate and security,this paper proposes a dynamic key generation scheme,use of CS technology.Through the security analysis and performance simulation analysis,it is concluded that the proposed dynamic key generation scheme can have a higher key generation rate on the premise of ensuring security.To be specific,on the one hand,the dynamic key generation scheme proposed in this paper can resist passive eavesdropping attack,computational attack against the measurement matrix and computational attack against the dynamic key.On the other hand,compared with several known key generation schemes,the proposed scheme has a higher key generation rate and better performance.Finally,on account of the main problems of the existing PLS schemes in OFDM-based systems,namely,most of them are lack of flexibility.We propose a flexible PLS scheme on the basis of the proposed dynamic key generation scheme in this paper.In the proposed scheme,the encryption module can be put before IFFT or after IFFT flexibly,which makes these two kinds of methods satisfy different requirements of BER and security.Through simulation analysis,it can be concluded that the proposed scheme has a better BER performance than the traditional typical channel estimation due to the use of CS technology.In addition,although the performance differences between the two types of methods are small,when the high-order modulation scheme(256-QAM modulation)is used,the difference becomes relevant and has an impact on system performance.Therefore,the proposed scheme is balanced between BER performance and security level,improving the flexibility.That is,compared with the time domain encryption scheme,the frequency domain encryption scheme reduces the effect of channel fading and further reduces the bit error rate;while time domain encryption is more secure.