Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Machine Learning Assisted Design of mmWave Wireless Transceiver Circuits

Xuzhe Zhao

TL;DR

This work addresses the challenge of designing mmWave ICs for 5G/6G by developing two 28-GHz transceiver blocks (transmitter and receiver) and embedding a machine learning–driven design pipeline that predicts circuit parameters from target specifications. It couples conventional ML models (Random Forest, SVR, MLP) and a Transformer within an end-to-end workflow, supported by Ocean scripts that automate Cadence simulations and data extraction. The contributions include a complete transmitter–receiver design demonstration, an automated data-generation and ML-evaluation framework, and a dataset plus benchmarking of models to accelerate analog/RF mmWave design. The results show improved design efficiency and accuracy, enabling rapid exploration of trade-offs in mmWave ICs with tangible implications for scalable 5G/6G prototyping and deployment.

Abstract

As fifth-generation (5G) and upcoming sixth-generation (6G) communications exhibit tremendous demands in providing high data throughput with a relatively low latency, millimeter-wave (mmWave) technologies manifest themselves as the key enabling components to achieve the envisioned performance and tasks. In this context, mmWave integrated circuits (IC) have attracted significant research interests over the past few decades, ranging from individual block design to complex system design. However, the highly nonlinear properties and intricate trade-offs involved render the design of analog or RF circuits a complicated process. The rapid evolution of fabrication technology also results in an increasingly long time allocated in the design process due to more stringent requirements. In this thesis, 28-GHz transceiver circuits are first investigated with detailed schematics and associated performance metrics. In this case, two target systems comprising heterogeneous individual blocks are selected and demonstrated on both the transmitter and receiver sides. Subsequently, some conventional and large-scale machine learning (ML) approaches are integrated into the design pipeline of the chosen systems to predict circuit parameters based on desired specifications, thereby circumventing the typical time-consuming iterations found in traditional methods. Finally, some potential research directions are discussed from the perspectives of circuit design and ML algorithms.

Machine Learning Assisted Design of mmWave Wireless Transceiver Circuits

TL;DR

This work addresses the challenge of designing mmWave ICs for 5G/6G by developing two 28-GHz transceiver blocks (transmitter and receiver) and embedding a machine learning–driven design pipeline that predicts circuit parameters from target specifications. It couples conventional ML models (Random Forest, SVR, MLP) and a Transformer within an end-to-end workflow, supported by Ocean scripts that automate Cadence simulations and data extraction. The contributions include a complete transmitter–receiver design demonstration, an automated data-generation and ML-evaluation framework, and a dataset plus benchmarking of models to accelerate analog/RF mmWave design. The results show improved design efficiency and accuracy, enabling rapid exploration of trade-offs in mmWave ICs with tangible implications for scalable 5G/6G prototyping and deployment.

Abstract

As fifth-generation (5G) and upcoming sixth-generation (6G) communications exhibit tremendous demands in providing high data throughput with a relatively low latency, millimeter-wave (mmWave) technologies manifest themselves as the key enabling components to achieve the envisioned performance and tasks. In this context, mmWave integrated circuits (IC) have attracted significant research interests over the past few decades, ranging from individual block design to complex system design. However, the highly nonlinear properties and intricate trade-offs involved render the design of analog or RF circuits a complicated process. The rapid evolution of fabrication technology also results in an increasingly long time allocated in the design process due to more stringent requirements. In this thesis, 28-GHz transceiver circuits are first investigated with detailed schematics and associated performance metrics. In this case, two target systems comprising heterogeneous individual blocks are selected and demonstrated on both the transmitter and receiver sides. Subsequently, some conventional and large-scale machine learning (ML) approaches are integrated into the design pipeline of the chosen systems to predict circuit parameters based on desired specifications, thereby circumventing the typical time-consuming iterations found in traditional methods. Finally, some potential research directions are discussed from the perspectives of circuit design and ML algorithms.
Paper Structure (27 sections, 12 equations, 21 figures, 4 tables)

This paper contains 27 sections, 12 equations, 21 figures, 4 tables.

Table of Contents

  1. Introduction
  2. Background
  3. Motivation
  4. Organization of This Thesis
  5. Literature Review
  6. Supervised Learning
  7. Reinforcement Learning
  8. Combination of NN and Optimization Algorithms
  9. A 28-GHz Transmitter Design
  10. Transmitter Architecture
  11. Performance Specifications Analysis
  12. Individual Performance
  13. System Performance
  14. Ocean Scripts Implementation
  15. A 28-GHz Receiver Design
  16. ...and 12 more sections

Figures (21)

  • Figure 1: Classification for prior automating techniques for circuit synthesis.
  • Figure 2: Machine learning techniques for analog circuit design reviewed in this thesis.
  • Figure 3: General learning loop for RL zhao_deep_2020.
  • Figure 4: Design flow for RL used in analog circuit design mina_review_2022.
  • Figure 5: An example of leveraging the knowledge from (a) simpler circuits to (b) more complex one fayazi_angel_2023.
  • ...and 16 more figures