Wireless channel-based ciphering key generation: effect of aging and treatment

,


1.
INTRODUCTION Wireless communication has become a mandatory portion of our everyday life.However, this portion requires gigantic efforts to have the wireless transmission secured.This issue has become significant after the ever-increasing amount of data transmitted, wirelessly [1]- [3].Since electromagnetic waves can propagate anywhere, data transmission is fundamentally fraught with danger.The security and confidentiality of the transmitted information have therefore consumed a substantial amount of researchers' time and efforts.In order to ensure the confidentiality of transmissions, general transmission devices either use symmetric-key encryption or asymmetric-key encryption.However, neither of these methods is recommended due to the fact that asymmetric-key encryption necessitates a significant amount of complex processing power, and the second method necessitates a very secure link to transfer the unique encryption key, which is a crucial concern.Because sending sensitive information becomes impossible once the secret key is revealed [4], [5].An alternative assumption is a keyless or physical layer security scheme was proposed.However, the keyless scheme limits the resources of wireless transmission.Furthermore, users of public networks face severe restrictions and costly additional expenses [6]- [10].
The concept of key generation was initially thought of and investigated theoretically in [11], [12].Key generation is a crucial component of wireless transmission security.The process of generating a key entails the creation and distribution of an encryption code that can be used to secure the data at both ends of the transmission.Using a single code to encrypt the entire transmission is hazardous, however, because if the code ❒ ISSN: 2088-8708 is compromised, the entire transmission will be compromised [13], [14].
The improvement in the security of wireless networks is obtained by attaining information-theoretic secrecy, which provides network security with greater wireless adaptability.In contrast to conventional encryption key generation methods, a recent security technique exploits the randomness of the wireless channel.This method of key generation greatly complicates attempts to reveal the transmission by a third party (adversaries) other than the transmitter and receiver [15]- [17].
Hence, the randomness in wireless channels can be harvested to generate uncorrelated secret codes to guarantee transmission security [18].This assumption does not always hold true.Specifically in a static environment, the wireless channel does not vary enough, resulting in a low level of introduced randomness in the following time interval.A practical example of this environment is the indoor internet of things (IoT).Correlated successive channels result in the generation of correlated secret codes which endanger transmission confidentiality.Due to the high correlation between encryption keys, compromising one key could be used to easily decrypt the next [18], [19].
Nevertheless, recent work on encryption key generation in the static environment has attracted considerable interest.The capacity of the encryption key was determined and addressed the parameters that affect the capacity [20]- [22].However, efforts are directed toward reaping the advantages of randomness through the application of reconfigurable intelligent surfaces (RIS) [23]- [25].The current efforts concentrate on reducing transmission losses and restricting the information leakage to the eavesdroppers in order to maximize link efficiency.So far as we can tell, though, there appears to be little effort to draw attention to the consequences of utilizing encryption keys correlated to the generated encryption keys at consecutive coherence times.
The contribution of this work addresses the impact of channel aging on the network security level to propose an alternative solution to overcome the issue of correlated successive encryption keys.This can be clearly seen in the low randomness environment or in other words, in the static environment when the user is sitting or moving lowly.Furthermore, the results of this paper show the impact of the proposed scheme and the speed of the moving user in enhancing network security.This study focuses on the design of a robust encryption key generation that avoids the correlation of consecutive encryption keys.The remainder of this paper is organized as follows, section 2 describes the system model of this study.The problem is brought up and formulated in section 3. The results are shown and discussed in section 4. Finally, section 5 concludes this study.

SYSTEM MODEL
In this paper, the transmitter base station (BS) or access point (AP) ) is equipped with M transmit antennas and serves multiple users (total number of users is K) each user is equipped with a single antenna, as shown in Figure 1.Symmetric key encryption is considered that is used to encrypt the data transmission between the transmitter and each individual user, u k .The link between each single user and the transmitter is secured with a unique secret code.The encryption key for each link is derived from the randomness of its acquired channel state information (CSI).The CSI of all users is gathered by the transmitter.For each coherent time of transmission, the transmitter creates a new encryption key for each user and conveys it via a confidential channel to all of the active users.However, when u k is in a static or slowly moving environment, the channel aging of u k may be low, and the subsequent coherent time may be correlated.In this way, the transmission would operate with correlated secret codes, and the disclosure of a single transmission's encryption key would have adverse consequences [26].Thus, breaking a single transmission's encryption key would be used by adversaries employing brute-force attacks to break the next hop's secret code since they are highly correlated, breaking a single hop's secret code is a weakness that could cause a threat.As well, spatially correlated adversaries will have time to brute force attack.

PROBLEM FORMULATION
This paper focuses on capturing the impact of the user in the slow-moving or static environment at the generated encryption key from the randomness of the wireless channel.The process of encryption key generation involves three main steps: probing, quantization, information reconciliation, and privacy amplification.The channel is measured in the first step and converted into binary in the second.The error correction and secrecy evaluation are done in the last two steps.The comprehensive explanation of each of these stages is detailed in [15].However, the channel between the transmitter and u k is given in (1) according to the assumptions in [17].
where h[n − 1] represents the channel vector for the prior time slots while the effect of channel aging is represented by the error vector e[n].The temporal correlation parameter is given by α = J 0 (2πf D T s ).Where J 0 () is the zeroth-order Bessel function of the first kind.T s , and f D are the channel sampling duration and the maximum Doppler shift, respectively.The value of the Doppler shift is given by ( 2): where v is the moving velocity of the user u k , c is the speed of light, and f c refers to the carrier frequency.[17].
For the successive time slots, the successive channel is correlated which leads to generating a correlated encryption key.The key to mitigating and improving the quality of encryption keys is to introduce artificial randomness.The introduced randomness is added to the acquired channel during the process of generation of the encryption key.For simplicity, the permutations for the required channel vector can be used to generate uncorrelated successive encryption keys.After permutating the channel vector a decimal-to-binary conversion has to be done.
where the vector h 01 is binary vector of the length n bit.For a given vector h 01 of length n it is simply to generate a matrix with the dimensions P = n! × n, every single row in matrix P has the same length as v but with different permutations.The transmitter and receiver are able to use any permutated row vector according to a pre-decided algorithm.

RESULTS AND DISCUSSION
This section shows the simulation results that shows our proposal.The simulations of results were performed using MATLAB software.All generated simulations were made by averaging out 1,000 trails.
In Figure 2, the correlation of the successive coherent encryption keys for the two schemes was analyzed and compared.For the conventional scheme, highly correlated encryption keys result from the correlation between successive time slots.The correlation between the encryption keys decreases as a result of the effects of aging, However, the correlation between the first six encryption keys is considerably high.In contrast to the conventional approach, the proposed approach has shown considerable improvements.Hence, comparing ❒ ISSN: 2088-8708 the first and second-time slots generated keys, the correlation drops from 6.3799 in the conventional scheme to 0.0013 in the proposed scheme.The results reveal that the generated keys are uncorrelated with the encryption keys that come from the second time slot.Hence, the impact of the permutation of the received channel vector can be clearly captured.Figure 3, demonstrates the influence of user speed on the correlation between two successive encryption keys that come from successive time slots.It can be clearly seen that for the slow-moving user or the static user, the channel is highly correlated due to the low-level impact of channel aging.Hence, the generated encryption keys from the highly correlated channel vectors are correlated accordingly.Contrarily, the correlation level drops for the fast-moving user.Accordingly, the key will be more vulnerable due to this correlation and a third party can get an advantage.Eventually, the eavesdropper has to find the permutation index which changes in coordination between transmitter and receiver before breaking the encryption which requires a huge computation power and time.The permutation index sharing between transmitter and receiver gives the key high randomness in which high computational power is needed to find it by the eavesdropper.CONCLUSION Security of transmitted data is important in wireless communications.Consequently, data encryption with the proper key is necessary.The encryption key has to be random and challenging to revoke its regenerate by a third party.The randomness of the wireless channel has been utilized to generate the encryption key.The correlation between the generated keys at consecutive intervals of coherence times is high, particularly in slowmoving mobile phones.This correlation is a weakness in network security and causes a threat.A proposed key generation method to overcome this threat is adding new randomization of binary key bits based on the bits permutations.This method highly improves the randomness and makes key regeneration by eavesdropper harder.The proposed method has improved the performance of the key.