Final Report-Rail Sensor Testbed Program: Active Agents in Containers for Transport Chain Security

摘要

Research was conducted on making transported objects (e.g., containers, pallets, and boxes) active participants in their own security. This effort focused on improving transportation security by enabling the objects being transported to become active agents in their own protection. Here, the objects are equipped with sensing and communication capabilities and are able to determine and communicate their sense of security throughout the dynamic transportation chain in a distributed manner. As part of the project, we have developed several data mining algorithms to enable intelligent agents to detect changes from their environment state. We have also designed algorithms to allow agents to communicate with each other for enhancing safety for the group of agents. Research issues of the designed algorithms have been applied to the Transportation Security SensorNet (TSSN) real transportation chain. We have tested all the new algorithms on real sensor data collected from transportation sensor network environments. The results have demonstrated the effectiveness of these algorithms for wireless sensor network security applications and provided useful insights regarding the challenges of the anomaly detection problem for distributed security in challenging environments. In addition the 'lessons learned' from our experiments are documented and a set of requirements for possible future systems were formulated.