Theory and practice for optimum spectral efficiency for ad-hoc wireless networks with strict requirements on latency and reliability

Research Project, 2015 – 2019

There is an increased interest in short-packet wireless ad-hoc networks due to the emergence of applications with very short latency requirements and/or inherently short messages. However, there is a lack of knowledge on what the fundamental limits are for trading latency, spectral efficiency, and reliability agains each other and how to design practical systems to approach these limits. The challenges faced in obtaining theory and practical solutions for short-packet systems are quite different from traditional long-packet systems, and new thinking is therefore required. We propose to reduce this knowledge gap by

1) Quantifing with information-theoretic bounds the highest spectral efficiency achievable for a given reliability and latency constraint under varying availability of channel state information (CSI).

2) Exploiting the Doppler-delay structure of propagation channels to formulate new CSI estimators with low overhead and sufficient accuracy.

3) Designing feedback-exploiting protocols that operate close to the fundamental limits discovered in Step 1 and with the CSI estimators from Step 2.

We will use vehicular ad-hoc networks for traffic safety and traffic efficiency applications as the motivating example, since such networks are (a) useful for making the road transport system safer and more sustainable and (b) have challenging requirements on latency, spectral efficiency, and reliability in combination with highly time-varying propagation channels.