Adaptive key generation based on quantization of channel characteristics

Length and mismatch rate of secret bit strings derived from channel characteristics by legitimate users for key generation are affected by quantization algorithm significantly. Existing schemes cannot achieve both a string and a low mismatch rate at the same time because of the static quantization method. In this paper we first propose an adaptive quantization algorithm formed of an quantizer offset with the quantization noise at one side and a quantization-level optimizing method, which maximize the length of secret bit string under the constraint of mismatch rate. Based on this algorithm, a key agreement scheme selection approach is presented to harvest longer bit strings. As a result, an adaptive key generation scheme based on quantization of channel characteristics is proposed using the above methods. With this scheme, bit strings with length close to theoretical limit can be generated, and the mismatch rate is lower than 5%.