Quantization at the Physical layer using Affinity Propagation Clustering

In today’s world security of information is of utmost importance. Traditionally symmetric and asymmetric cryptographic methods are used to keep information secure. But they face problems of high computational complexity and key management infrastructure. These issues can be resolved by Physical layer key generation (PLKG) schemes which use channel properties to generate a symmetric key. A basic PLKG system consists of probing the wireless channel, quantizing the probed samples and converting them into bits, error detection and correction between the quantized output bits and privacy amplification using secure hash functions. Quantization plays a critical role in PLKG in order to obtain the symmetric secret keys. This paper proposes Vector Quantization with Affinity Propagation Clustering (VQAPC). When the two legitimate users (Alice and Bob) assign a set of samples to different quantization regions, there will be difference in the quantization results which is called as Assignment Disagreement Rate(ADR). Simulation is done using MATLAB software. The proposed quantization method-VQAPC gives very low ADR as compared to the existing methods.