Data integrity attacks and defenses for Intel lab sensor network

Wireless sensor networks have been increasingly popular and they have been deployed in a wide range of areas including transportation system, healthcare, robotics, and smart home. Wireless sensor networks have facilitated our life using the remote sensing ability. However, these systems could be taken advantage by the malicious parties to mislead users since they are often not physically secured and are used in hostile environments. We have proposed theoretical and statistical framework for creating data integrity attacks and corresponding security defenses. The data integrity attack is defined as a type of attack that targets to mislead the system by altering some collected data of the system in such a way that to achieve the goal of attackers. We use the Intel lab sensor data and demonstrate that it is easy to attack only a few essential sensors values to change the behavior of system, such as the control of air conditioner. To be more specific, by changing the temperature and humidity data measured by a group of sensors, the attacker can control the air conditioner turning on and off without being detected by correlation check. On the top of data integrity attacks, we have further developed low overhead defense schemes.