Skip to content. Skip to main navigation.

News Archive 2001 - 2010

Trust-based Mechanisms Build Cooperative Control for Networks

July 16, 2009

Trust in the source of information is important in any situation, but especially so That’s why researchers at The University of Texas at Arlington are developing methods to establish protocols to evaluate and establish levels of trust that can be used in battlefield and emergency response situationsin situations where human lives are at stake.

Electrical engineering Professor Frank Lewis is head of the Advanced Controls and Sensors Group located in the College of Engineering’s Automation & Robotics Research Institute. He is leading a three-year, $250,000 project funded by the Air Force Office of Scientific Research, hoping to create a framework that combines interdisciplinary ideas from control theory and communications and results in a global consensus of trust among a network of individual systems (nodes).

Much as flocks, swarms and herds have built-in trust mechanisms to identify team members, team leaders and enemies, individual nodes in a network of unmanned air and ground vehicles could receive and evaluate information gathered by others in the group and conclude if action is needed and, if so, what kind of action.

For example, a semi-autonomous group of vehicles could be directed by one soldier to scout the interior of a building believed to contain enemy fighters. Acting on this command, the vehicles would move individually throughout the building, sensing and reporting to each other and the commanding soldier what they were encountering. If most of the first floor of the building had been searched, a number of the nodes could, as a group, decide to split from the main network and follow a trusted “leader” to begin a search on the second floor.

However, it is possible for one of the nodes to malfunction or be “captured” by the enemy and made to send spurious information to the other nodes. If this misinformation is discovered, the other nodes will become distrustful of the malicious node and further information sent from it will be disregarded. Professor Lewis is working with computer science and engineering Professor Sajal Das to develop efficient methods to detect malicious nodes in a network.

“Trust and leadership are based on past performance,” said Dr. Lewis. “Individual nodes can develop what could be called social standing through actions they have taken in other situations. Heirarchies can even develop within a network.”

Applications developed in the project will be tested using autonomous aerial vehicles and programs already developed by Dr. Lewis’ graduate students at theAutomation & Robotics Research Institute. These include a quadrotor helicopter created by Emanuel Stingu, ground vehicles developed by Chris McMurrough and Matt Middleton, and advanced control programs by Draguna Vrabie.

Though much of what Dr. Lewis and his team are developing will primarily be used in military situations, it will also be applicable to human-directed civilian operations such a search and rescue, firefighting, emergency evacuations and the like. The title of the project says it all: Trust-based Collaborative Control for Teams on Communication Networks.