| With the rapid development of electronic technology and the Internet, wireless communication gradually presents the trends of broadband and mobility, and its related applications have penetrated into people’s daily life. However, the openness, flexibility, instability and other characteristics of wireless communication make it extremely vulnerable. What’s more, along with the increase of the users and the development of computer technology, the traditional security mechanism shows its shortcomings increasingly, such as difficulties of key management and low capability of attack resistance, so that, it cannot meet the actual demand of security.As a new method of enhancing the safety of wireless communication and realizing theoretical security, wireless physical layer security has become a hot research topic in the field of information security. From the perspective of information theory, using the uniqueness, reciprocity, time variation and unpredictability of channel as the theory basis, it utilizes propagation characteristics of wireless channel to generate the secret key. This scheme makes the remote communication users generate the same key independently at the same time, namely the implementation of key sharing in the absence of key distribution, which avoids the transmission of the key, so it has a broad application prospect. The research on secret key generation technology based on the wireless channel characteristics in wideband OFDM system is studied in this paper, including extracting channel characteristics and secret key generation. The main work is as follows:1. A four channel ADC system with high sampling rate based on the MHFB is designed to realize the multi-channel joint reception of broadband signal in OFDM system, so that the sampling rate is decreased by 4 times. On this basis, the method of extracting channel characteristics is studied. According to the frequency characteristics of OFDM signal, communication terminals extract the amplitude of channel frequency response on each subcarrier as the channel characteristics. The channel characteristics obtained by this method have high similarity, and their correlation coefficient is about 0.99, which guarantees the success of key generation.2. In order to improve the reciprocity of channel characteristics, a new method of preprocessing based on the wavelet transform is proposed. This method decomposes the channel characteristic by the use of sym8 wavelet and reconstructs it after processing of the high frequency coefficient, which reduces the influence of noise effectively, and the differences between the channel characteristics is reduced by 20%, so that the BER during the key generation is reduced by 41% approximately.3. An improvement of calculation for quantization threshold by using median for reference is proposed to ensure the randomness of the secret key. Using the threshold calculated in this way to quantify the channel characteristics after pretreatment ensures that’0’and’1’have roughly equal proportion in the initial key sequence, which makes the sequence pass the frequency test.4. Segmented key confirmation is proposed to improve the key generation rate. This method divides the corrected sequence into 4 segments, and keeps the subsequence with the same SHA-1 value between the communication users. In ensuring the security of the key, the method aims to extract consistent bits as many as possible. In the UMa channel model, the key generation rate increases by 29% approximately compared to the rate in the case of confirmation with the whole sequence. Finally, the proposed algorithm interleaves the confirmed sequence to complete the key generation, which, improves the randomness of the key. According to the analysis, the 128-bit key generated from slow time-varying channel can pass 3 randomness tests at least and the key generated from fast time-varying channel can pass at least 5 tests, which proves that the key is random and can be used in secret communication. |
PDF Full Text Request