ChirpKey: A Chirp-level Information-based Key Generation Scheme for LoRa Networks Via Perturbed Compressed Sensing.

Huanqi Yang,Zehua Sun,Hongbo Liu,Xianjin Xia,Yu Zhang,Tao Gu,Gerhard Hancke,Weitao Xu

DOI: https://doi.org/10.1109/infocom53939.2023.10228886

2023-01-01

Abstract:Physical-layer key generation is promising in establishing a pair of cryptographic keys for emerging LoRa networks. However, existing key generation systems may perform poorly since the channel reciprocity is critically impaired due to low data rate and long range in LoRa networks. To bridge this gap, this paper proposes a novel key generation system for LoRa networks, named ChirpKey. We reveal that the underlying limitations are coarse-grained channel measurement and inefficient quantization process. To enable fine-grained channel information, we propose a novel LoRa-specific channel measurement method that essentially analyzes the chirp-level changes in LoRa packets. Additionally, we propose a LoRa channel state estimation algorithm to eliminate the effect of asynchronous channel sampling. Instead of using quantization process, we propose a novel perturbed compressed sensing based key delivery method to achieve a high level of robustness and security. Evaluation in different real-world environments shows that ChirpKey improves the key matching rate by 11.03–26.58% and key generation rate by 27–49× compared with the state-of-the-arts. Security analysis demonstrates that ChirpKey is secure against several common attacks. Moreover, we implement a ChirpKey prototype and demonstrate that it can be executed in 0.2 s.