Optimizing Pilot Length for a Go/No-Go Decision in Two-State Block Fading Channels with Feedback

TL;DR

This paper develops an optimized approach for pilot length and channel identification in two-state fading channels with feedback, aiming to minimize total transmission symbols, improve throughput, and save energy.

Contribution

It introduces a method to optimize pilot length and detection thresholds in variable-length coding with feedback, enhancing channel efficiency and energy savings.

Findings

Optimal pilot length and thresholds are derived and numerically solved.

Significant reduction in channel occupancy demonstrated across various settings.

Method improves throughput and energy efficiency in fading channels.

Abstract

We propose an approach where each user independently seeks to minimize the amount of time that they occupy the channel. Essentially, we seek to minimize the number of transmitted symbols required to communicate a packet assuming variable-length coding with feedback. Users send a pilot sequence to estimate the channel quality and decide whether to proceed with a transmission or wait for the next opportunity. Thus a user may choose to leave the channel even though it has already gained access, in order to increase the network throughput and also save its own energy resources. This paper optimizes the number of pilots and the channel identification threshold to minimize the total number of transmitted symbols (including pilots) required to communicate the packet. We prove a sufficient condition for the optimal pilot length and the channel identification threshold. This optimal parameter pair is solved numerically and the reduction in channel occupancy is shown for various channel settings.