ECG based unique identity for embedded devices

In the last decade, sensor nodes have been widely deployed in many application, owing to the their simple attributes and installation process. However, the sensors nodes are vulnerable to many potential threats such as node cloning and identity attack. The threats become more significant when the nodes are used in applications such as biomedical and e-Health applications where trust, privacy and security are of a major concern. In trusted systems, identity plays an important role in completing the chain of trust. One of the potential attacks on wireless sensor nodes is identity attack namely replication attack where one or more nodes illegitimately claim the identity (ID) of a legitimate node and replicate itself throughout wireless sensor network (WSN). Deployment of high security solutions leads to implementation challenges such overcoming the resource constraints that exist in sensor nodes and the need for arbitration by trusted third parties. This study explores the feasibility of using the human electrocardiogram (ECG) characteristics in generating the identity of the sensor node platform thus eliminating the need to store the identity on the node itself. An identity generation algorithm is derived from features extraction and reduction signal to cope with the constraint of the sensor node platform and provide the main solution to the above mentioned issues.