A medical sensor network can wirelessly monitor vital signs of humans, making it useful for long-term health care without sacrificing patient comfort and mobility. For such a network to be viable, its design must protect data privacy and authenticity given that medical data are highly sensitive. We identify the unique security challenges of such a sensor network and propose a set of resource-efficient mechanisms to address these challenges. Our solution includes (1) a novel two-tier scheme for verifying the authenticity of patient data, (2) a secure key agreement protocol to set up shared keys between sensor nodes and base stations, and (3) symmetric encryption/decryption for protecting data confidentiality and integrity. We have implemented the proposed mechanisms on a wireless mote platform, and our results confirm their feasibility.
Keywords: elliptic curve cryptography; medical sensor network; neural network; security.