News and Notes for Our Industrial Partners
July 2008

FACULTY

Allen B. MacKenzie Research Summary
Dr. Allen B. MacKenzie, 540-231-3565, is an Assistant Professor in the Department of Electrical and Computer Engineering. Over the past few years, a major research focus at Wireless @ Virginia Tech has been cognitive radio. Given the flexibility and reconfigurability that software defined radio has brought to radio platforms, research in cognitive radio attempts to determine how to intelligently control these powerful radio platforms for maximum benefit. In order to do so, cognitive radio employs algorithms and techniques from artificial intelligence to solve difficult computational problems, such as determining the right channel, modulation scheme, transmit power level, and coding rate to use in a given scenario. A key application of cognitive radio is dynamic spectrum access, which potentially relieves the “shortage” of radio spectrum by using spectrum more efficiently. In this case, a cognitive radio will assess the activity in a radio band and adjust its operation to provide maximum performance with minimal impact on other radio systems. Most cognitive radio work, though, has focused on improving the performance of individual radio links between two cognitive radios, without regard for the impact that this radio link may cause to other links in a given wireless network. As long as cognitive radios are relatively rare, this approach may be viable. But if cognitive radios are widely adopte, then this approach could lead to a “success disaster” in which cognitive radios destroy access to the radio spectrum through heavy-handed competition with each other. To address this problem, Dr. MacKenzie and Dr. Luiz DaSilva, another Wireless @ Virginia Tech faculty member, developed the concept of a cognitive network. A cognitive network is a set of wireless nodes that intelligently react to their environment in an effort to improve the end-to-end performance of a wireless network. One aspect of Dr. MacKenzie's current research on cognitive networks is topology control in dynamic spectrum access networks. MacKenzie is principally focused on how distributed nodes can organize their use of channels in a way that maximizes network performance while minimizing the amount of overhead required for network organization. A related area that Dr. MacKenzie is addressing is how to coordinate access to different radio channels, a problem known as medium access control, which is particularly difficult in dynamic spectrum access systems built with current software defined radio technology. In addition to developing theoretical solutions to channel assignment and medium access control in cognitive networks, MacKenzie and his team are demonstrating some of their ideas using a cognitive radio platform developed in conjunction with Dr. Charles Bostian. Prior to working on cognitive networks and cognitive radio, Dr. MacKenzie's research focused on applications of game theory in wireless communication systems and networks. Game theory is a branch of applied mathematics that studies what happens when decision makers with conflicting interests interact. Now, MacKenzie and his team are applying techniques from game theory to understanding what will happen in a cognitive network. Since cognitive networks are made up of autonomous nodes with conflicting interests, game theory is a good tool for analyzing the performance of these networks. MacKenzie's primary research sponsors include the National Science Foundation, the National Institute of Justice, and the Defense Advanced Research Projects Agency (DARPA) under a subcontract from BBN Technologies.