CIF: Small: Fundamental Communication Latency Limits Beyond the Traditional Block-Coding Architecture

Advanced wireless communication networks have driven many technological advances ranging from the convenience of smartphones, personalized health monitoring devices, to smart soil / crop sensors for precision agriculture, to name but a few. While the amount of data that can be delivered through wireless networks has increased exponentially in the past decades through numerous innovations of signal processing, there has been significantly less progress on the reduction of latency (i.e., response time) of wireless communications.

This project aims to substantially shorten the wireless communication latency by first characterizing the drawbacks of the ubiquitously used design paradigm called the block coding architecture. The project will then design non-block-coding-based solutions that significantly lower the latency beyond what is considered possible under standard block-coding architectures.

The new ultra-fast response time of wireless communications will support complex, diverse, and highly dynamic applications in Internet-of-Things, edge computing, and cyber-physical systems, and may yet unlock countless new transformative applications beyond what is conceivable under today's technologies.

In the relentless pursuit of ultra-low latency, a system designer must utilize every source of delay reduction, and this project thus takes a multi-pronged approach that minimizes, simultaneously, the decode-&-forward delay in the network level, the queueing delay in the link level, and the synchronization delay in the signal level. For the network level, the project will reinvent amplify-&-forward schemes by designing a soliton-based concatenation mechanism that takes advantage of the ultra-low latency of amplify-&-forward without its main drawback of error accumulation.

For the link level, new random-walk-based analysis and design tools will be developed as the theoretical foundation for new blockless designs that eliminate queueing delay completely through continuous, on-the-fly packet encoding. For the signal level, this project proposes a rateless design under the framework of quickest change detection. The resulting scheme would significantly improve the link start-up / synchronization time, a major source of delay when sending short, sporadic messages to Internet-of-Things devices.

The theoretical results of this project should lead to new practical designs that harvest the ultra-low latency benefits beyond any block-based solutions and open up new ideas in other important disciplines, including reliable and fast information propagation over noisy social networks, and distributed computation over unreliable communication networks.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.