Millimeter Wave Receiver Comparison Under Energy vs Spectral Efficiency Trade-off

TL;DR

This paper compares analog, hybrid, and digital mmWave receiver architectures focusing on energy versus spectral efficiency trade-offs, using a flexible comparison method and multi-objective optimization to guide architecture choice based on component parameters.

Contribution

It introduces a reproducible comparison framework for mmWave receivers considering different ADC power models and interprets receiver selection as a multi-objective utility optimization problem.

Findings

DC does not always consume the most power as commonly believed.

Either DC or HC can offer the best SE vs EE trade-off depending on component parameters.

The comparison method provides a compact chart for receiver architecture selection.

Abstract

Receivers for mmWave systems suffer from high power consumption in Analog to Digital Converters (ADC), and there is a need to compare the three major receiver architectures: Analog, Hybrid and Digital Combining (AC, HC and DC). Moreover, the specific power consumption figure of merit of ADCs varies significantly between different component designs in the literature, so that comparisons performed for one ADC model - no matter how representative of the state of the art - do not necessarily carry over to other ADC designs with different figures of merit. In this work, we formulate a comparison method between AC, HC and DC that can be easily reproduced with different power consumption parameters and provides all information for receiver architecture selection in a compact chart figure. We also present an interpretation of the receiver selection decision problem as a multi-objective utility optimization to find the best Spectral Efficiency (SE) versus Energy Efficiency (EE) trade-off. We use existing results on the achievable rate of AC, HC and DC systems and an Additive Quantization Noise Model (AQNM) of the ADC capacity degradation. For some example commercial component parameters, we show that the usually held belief that DC requires the highest power is not valid in many cases. Rather, either DC or HC alternatively result in the better SE vs EE trade-off depending strongly on the considered component parameters and on the weight assigned to SE vs EE in the utility maximization.