Research On LoRa Radio Frequency Fingerprint Extraction And Recognition Method

With the rapid development of the Internet of Things,the Internet of Everything is gradually becoming a reality,and in the wireless communication process of the Internet of Everything,the access security of wireless devices has become a concern.LoRa(Long Range Radio)technology is a new technology emerging with the development of the Internet of Things.It has a wide range of application prospects in the field of Internet of Things.However,the access authentication mechanism in the current LoRa communication process is not perfect,and it is vulnerable to identity counterfeiting,and other malicious attacks.Aiming at the problem that the LoRa access authentication mechanism is vulnerable to be attacked,this paper proposes that LoRa radio frequency fingerprints can be extracted as the unique identification information of LoRa devices to distinguish and identify LoRa devices.The radio frequency fingerprint feature embodies the inherent hardware feature of the wireless transmitter,which is unique,stable and not easy to be tampered with.Because this feature comes from the hardware feature of the transmitter itself,it has good safety and reliability.At present,there is still a lack of research on LoRa radio frequency fingerprints.In response to this problem,this article has conducted research on LoRa radio frequency fingerprints,and in-depth study of LoRa's radio frequency fingerprint extraction and identification problems in some situations.It has important theoretical and practical significance for optimizing the access security mechanism of LoRaWAN and remedying the deficiencies in LoRa radio frequency fingerprint research.In this paper,both the radio frequency fingerprint of the LoRa steady-state signal and the radio frequency fingerprint of the transient signal are studied,and the method of extracting the radio frequency fingerprint in two states is proposed.The effectiveness of the proposed radio frequency fingerprint is verified through experiments;at the same time,it is aimed at low signal noise.The problem of radio frequency fingerprint extraction in the environment is studied,and two methods of noise suppression method based on symbol superposition and fingerprint extraction method based on adaptive filtering are proposed,and the impact on radio frequency fingerprints is verified;finally,based on LoRa radio frequency fingerprint technology,An optimized LoRaWAN access security mechanism is proposed.The main work of this paper is as follows:1.The problem of radio frequency fingerprint extraction of LoRa steady-state signals is studied,and a steady-state radio frequency extraction method based on the fusion of IQ offset characteristics and carrier frequency offset characteristics is proposed.Experimental results show that,compared with a single carrier frequency offset feature,the fusion steady-state radio frequency fingerprint feature improves the recognition accuracy from less than 40% to65.8% under the condition of 10 d B.2.Aiming at the problem that the carrier frequency offset characteristics will drift with the increase of working time,a carrier frequency offset tracking method is proposed.In the process of carrier frequency offset drift,according to the small carrier frequency offset deviation between adjacent frames,by calculating the frequency offset value between adjacent frames and performing feedback adjustment,the carrier frequency offset value is dynamically updated to control the frequency offset drift.Correction.3.The amplitude transient characteristics and frequency transient characteristics of the LoRa signal are studied,and the transient amplitude and transient frequency characteristics are reduced to 160-dimensional characteristics.The experimental results show that,through multiple classification algorithms such as linear discrimination,under the reference signalto-noise ratio of 40.6d B,the use of these transient features can achieve 100% recognition accuracy for all six test equipment.4.Aiming at the problem of radio frequency fingerprint extraction in a low signal-to-noise ratio environment,a noise suppression method based on symbol superposition is studied.The noise suppression is achieved by superimposing the same symbols,and the differential constellation trace figure(DCTF)method is used to verify the effect of this method on improving the performance of radio frequency fingerprints.Aiming at the problem of noise interference to radio frequency fingerprint extraction,a fingerprint extraction method based on adaptive filtering is proposed.By taking the signal in the ideal SNR environment as the standard signal,an adaptive filter using the LMS algorithm is designed and realized,and the filter coefficients are extracted as the radio frequency fingerprint feature.5.Aiming at the problem of weak authentication of current LoRaWAN access security,an optimized LoRaWAN access security mechanism combined with LoRa radio frequency fingerprint technology is proposed.By combining the LoRa steady-state radio frequency fingerprint and LoRa transient radio frequency fingerprint studied in this article with the LoRaWAN access security mechanism,the gap in the access security of LoRaWAN physical layer is made up.