Research And Implementation Of Key Technologies Of Key Generation Based On Channel Characteristics

In recent years,wireless communication systems and the services they provide have been a basic component of the industrial world,but wireless networks are vulnerable to attacks from illegal parties due to their inherent broadcasting characteristics.The traditional encryption technology used in wireless networks is mainly concentrated on the upper layers of the protocol stacks,but with the development of computer computing power,this security mechanism is beginning to face challenges.Physical layer security(PLS)is not affected by the computing power of illegal parties.Based on this,this thesis studies the physical layer security mechanism for the characteristics of wireless channels.The main tasks are as follows:The thesis analyzes the general process of the key generation mechanism based on the characteristics of the wireless channel,which mainly includes four steps of channel detection,quantization,key agreement,and privacy amplification,and three evaluation indicators: key disagreement rate,key randomness and key generation rate is also explored in depth.The thesis first investigates the derivation of secret capacity under the traditional time division duplexing(TDD)mode on the basis of the available literature,next widens it to the derivation of secret capacity in the frequency division duplexing(FDD)using two-way detection,and obtains its theoretical expressions.Analysis and simulation prove that the key generation based on FDD system experiences two channels in the detection process compared with TDD system,so the secret capacity obtained is larger.Based on the theoretical analysis,the randomness of key comes from the channel,and the key generation rate is low in the slow-varying channel.Consequently,the thesis investigates the secret bit rate enhancement technique using random perturbation,and proposes a key generation rate enhancement algorithm based on chaotic pseudo-random sequences for the problem that random perturbation can lead to an increase in reconciliation.The algorithm takes advantage of the characteristics of chaotic mapping which is sensitive to the initial value,takes the reconciliation key as the initial value of chaotic mapping,and generates enough pseudo-random sequences as the final secret key by iteration.For the quantization aspect of key generation,the thesis deeply analyzes and discusses a variety of representative algorithms,and evaluates the algorithms from the perspectives of effectiveness,complexity,and interactive information.Digital analysis and simulation results demonstrate that compared with other quantization algorithms,the bit string obtained by the equal-probability quantization has strong randomness,and does not require sacrificing the generation rate and additional signaling transmission,which is simple and practical.Referring to the derivation of the theoretical key disagreement rate in TDD standard,the thesis extends it to FDD standard and derives a closed-form expression for the theoretical key disagreement rate after multi-bit quantization.From the formula,it can be seen that increasing the guard interval or decreasing the quantization order can effectively reduce the key disagreement rate.For the reconciliation session of key generation,the thesis studies the Cascade algorithm that uses dichotomy to find error correction,and simulates the key disagreement rate after reconciliation and the amount of information needed to be exchanged during the reconciliation process.The digital simulation results show that the algorithm also has good error correction performance under low signal-to-noise ratio,but requires a lot of information exchange.In view of the large amount of information exchanged by the Cascade algorithm,the thesis deeply studied the reconciliation algorithm based on the error correction code.This type of algorithm uses the error correction ability of the code word to correct the error bit,and requires less information to be exchanged and easier to achieve.The thesis also further simulates and compares the error correction code reconciliation algorithm represented by Hamming code,repetitive code and LDPC code,and chooses repetitive code as the system scheme from the perspective of complexity and performance trade-off.After that,the thesis combines the previous research and gives the final scheme of the key generation system.the thesis selects two-way detection based on the FDD standard,denoising scheme based on improved wavelet transform,equal-probability quantization,reconciliation based on repeated code,and key generation rate improvement scheme based on pseudo-random sequence to jointly form the final overall scheme.Eventually,the thesis evaluates the property of the project through digital simulation and semi-physical platform.Finally,the full work of this thesis is summed up and the follow-up research is put forward.