SecAP: Security Analysis of the Artificial Pancreas

The artificial pancreas system is a recent development to assist patients with type 1 diabetes controlling their blood glucose levels. It involves several medical devices that use wireless protocols for communication and thus are vulnerable to cyber attacks. This paper focuses on replay and denial-of-service attacks because of their commonality and potentially significant impact on the patient's health. These attacks were modeled and implemented on the GlucoSym Simulator, a program used to develop safer and more accurate AP control algorithms. The effects of each attack on the patient's glucose and insulin dose levels were observed and recorded. Observations of events of hyperglycemia and hypoglycemia over the course of the simulation were used to identify the ideal time for an attacker to cause the most harm to the patient. Finally, a detection method for these attacks is proposed to alert the AP user that corrective actions need to be taken in the event of an attack.