Generation secured cryptography keys in Wireless Body Sensor Networks using electrocardiogram (ECG) signals

Abstract

   Wireless Body Sensor Networks (WBSNs) consist of number of small size and limited sources sensor nodes. These sensors can be placed permanently or semi- permanently inside patient body or it can be located on the patient skin for medical issue. These sensors can sense certain parameters from the body that called biometric parameters. Considering the importance of transmitted data in these networks, so the study of security issues and protecting telecommunication are very important for these special-cases and limited-resources networks.  

   Using biometric parameters is very effective way to protect communication over WBSNs. We can use patient body to generate cryptography keys by biometric parameters, and we call them biometric keys. Even that using traditional biometric parameters increases security degree in protection systems, it has many disadvantages, which encouraged using more effective biometric parameters such as Electrocardiograms (ECG).

    In this research, we use ECG signal to generate binary strings depending on MIT_BIH Arrhythmia database. We could generate binary strings, each of them has size at 512 bits. We study the randomness of these strings by NIST. Also, we study distinctiveness by Hamming Distance. This research proved that these binary strings could be used as cryptography keys in WBSNs.