Using Interference Structure To Improve Wireless Communication
ECE/COS Professor Ken Duffy, in collaboration with Boston University and MIT, was awarded a $780,000 NSF grant for “Interferers in Our Midst.” The team will develop methods to improve communication performance in shared, congested, and contested spectrum bands.
Abstract Source: NSF
In the data-driven world that we live in, sharing digital information is a key component underpinning a vast body of technologies. Low-latency, high fidelity access to information is central to algorithms that impact how we work, are entertained, how we travel, and our healthcare. Systems that, in particular, rely on wireless communication to deliver their services have become ubiquitous. With the increase in data transmitted over the air, however, the central resource that they depend on, spectrum, can become congested with multiple communications overlapping and negatively impacting each other. This project brings together diverse researchers from Northeastern University (NU), the Massachusetts Institute of Technology (MIT) and Boston University (BU) to develop methods that improve communication performance in shared, congested, and contested spectrum bands.
The presence of interference in communications detrimentally impacts the throughput and reliability of systems. Interference and noise are often used interchangeably as they are commonly lumped together as general deleterious effects that corrupt communications. Interference, however, has a more structured form than noise. Central to this project is developing new means to leverage that structure to improve communication systems. By enabling more efficient use of scarce resources, more services can reliably co-exist, advancing national health, prosperity and welfare. By developing techniques that are receiver-only, it allows both backward compatibility and graceful adoption paths.
Interference management motivates substantial engineering effort at all levels, from hardware design, to signal processing, to error correction, to retransmission, and resource allocation protocols. A traditional approach to managing interference is to consider its impact as being part of noise. This project aims to do more, leveraging the structure of interference to improve performance through receiver-side approaches only, thus circumventing barriers to technological adoption. When a modulated communications signal experiences interference that arises from other modulated communications, those characteristics can be taken into account. Even when an interferers’ modulation may not be discerned, the interference can influence the noise experienced by a receiver in semi-predictable ways that can be exploited by a receiver. When interference is due to the presence of other communication systems where individual interferers’ modulation can be detected but the signal not decoded, unlike in a multiple user system, this project proposes an approach that takes into account both noise and the restricted forms the interference can take. When channel and modulation may not be available at the receiver, interference will still have characteristics that are different from, e.g., Gaussian noise. The statistical characteristics of such interference can be used to improve forward error correction decoding, enabling reliable communication with less overhead, which this project explores. When interference is due to signals that vary more slowly than the communication, such as from electronic devices, the receiver cannot rely on knowledge of the structure of the interference, other than the fact that it will exhibit a slowly varying profile. In that case, this project aims to discover post-decoding the interference experienced by some signals and use it as a starting point to remove pre-emptively at least partially that interference from other signals that are proximate in time, and thus subject to a similar interference.