ECE: Electrical & Computer Engineering
ECE News



Visit the Wireless @ Virginia Tech website.

VT cognitive radio technology to help improve railway communications, safety

Diagram of cognitive radio and trains

A Wireless@VT research team is adapting Virginia Tech cognitive radio technology to improve safety and operations of the nation’s railways.

The effort is funded through the Federal Railway Administration’s Office of Research and Development and dovetails with the Positive Train Control (PTC) technology being implemented under the U.S. Rail Safety Improvement Act of 2008. PTC is expected to be implemented by 2015 and to include wireless communication and GPS navigation throughout the rail system.

Under a PTC system, trains are expected to communicate with a base station and control center when they enter a rail yard. Using the highest bandwith available, a train would download the data regarding up-to-the-minute status in the yard. While traveling, trains will communicate with wayside wireless stations — delivering information about their location, speed, and direction, while acquiring information about the situation ahead and where they are allowed to travel.

Wireless communication with trains can be challenging, acknowledges Ashwin Amanna, a senior research associate and the project’s principal investigator. Not only is the train’s situation constantly changing, but so is the communications environment, with changing noise, multiple sources of interference, potential for multiple users contending for limited spectrum, and unpredictable usage by the users, he says.

“The railway industry is starting to use software defined radios (SDRs) that have a reconfigurable platform capable of condensing multiple radios into a single unit,” he notes. “A railroad-specific cognitive radio (Rail-CR), however, would incorporate additional adaptation to mitigate crowded spectrum, intentional jamming or learn from past experiences.”

The ECE team is developing a cognitive engine that uses AI tools such as case-based reasoning and optimization algorithms. The cognitive engine will also incorporate a policy engine that would prevent the radio from operating outside the policies of the area. “This is particularly important when a train runs along the border of another state or country,” Amanna explains.

The resulting cognitive engine will be provided open source for easy integration with any software defined radio. Amanna is working with Manik Gadhiok, Matthew Price, and co-principal investigator Jeffrey Reed on the Rail-CR effort. The research team is integrating the software-based cognitive engine with a private partner’s railway specific SDR.