AI-Driven Pixelated Matching Network Design with Automated Impedance Clustering Constraints for Power Amplifiers

Jingyun Bi; Xinyu Zhou; Jing Xia; Wenjie Feng; Yanze Wang; Wing Shing Chan

doi:10.1109/IWS65943.2025.11177942

AI-Driven Pixelated Matching Network Design with Automated Impedance Clustering Constraints for Power Amplifiers

Jingyun Bi
, Xinyu Zhou
, Jing Xia
, Wenjie Feng
, Yanze Wang
, Wing Shing Chan

Department of Electrical Engineering

Research output: Chapters, Conference Papers, Creative and Literary Works › RGC 32 - Refereed conference paper (with host publication) › peer-review

3 Citations (Scopus)

Abstract

This paper presented a fully automated inverse design method for power amplifiers (PAs) based on specific pixelated design spaces. Traditional PA design relied on manual Load-Pull analysis and late-stage manufacturability tuning, while defining impedance constraints remained a challenge for automation. A layout-level Electromagnetic (EM) model was incorporated early in the design process, with a fixed-size pixelated space allocated for the matching network. Various ideal microstrip topologies served as placeholders for pixel to explore impedance information with the goals of achieving the desired amplifier performance. The DBSCAN algorithm clustered impedance values of topologies at each frequency on the Smith chart, and elliptical envelopes were used to automatically define impedance constraints for pixel matching. This method automated the design process, shortened the design cycle, and delivered ultra-compact, high-efficiency PAs suitable for space-constrained applications. Results demonstrated a PA achieving 72% PAE and 40 dBm output at 3.35 GHz, with matching network size of 0.2 λ_g× 0.2 λ_g. © 2025 IEEE.

Original language	English
Title of host publication	2025 IEEE MTT-S International Wireless Symposium (IWS 2025) - Proceedings
Publisher	IEEE
Number of pages	3
ISBN (Electronic)	979-8-3315-3801-9
ISBN (Print)	979-8-3315-3802-6
DOIs	https://doi.org/10.1109/IWS65943.2025.11177942
Publication status	Published - 2025
Event	12th IEEE MTT-S International Wireless Symposium (IWS 2025) - Xi’an Qujiang International Convention Center, Shaanxi, China Duration: 19 May 2025 → 22 May 2025

Conference

Conference	12th IEEE MTT-S International Wireless Symposium (IWS 2025)
Place	China
City	Shaanxi
Period	19/05/25 → 22/05/25

Bibliographical note

Full text of this publication does not contain sufficient affiliation information. With consent from the author(s) concerned, the Research Unit(s) information for this record is based on the existing academic department affiliation of the author(s).

Research Keywords

AI driven design
impedance matching
inverse design
pixelated design space
Power amplifiers

Access Output

10.1109/IWS65943.2025.11177942

Other Access Options

Check CityUHK Library

Permanent Link

Right click to copy link

Cite this

@inproceedings{cb7616c6ffab45f4a061d80181051fde,

title = "AI-Driven Pixelated Matching Network Design with Automated Impedance Clustering Constraints for Power Amplifiers",

abstract = "This paper presented a fully automated inverse design method for power amplifiers (PAs) based on specific pixelated design spaces. Traditional PA design relied on manual Load-Pull analysis and late-stage manufacturability tuning, while defining impedance constraints remained a challenge for automation. A layout-level Electromagnetic (EM) model was incorporated early in the design process, with a fixed-size pixelated space allocated for the matching network. Various ideal microstrip topologies served as placeholders for pixel to explore impedance information with the goals of achieving the desired amplifier performance. The DBSCAN algorithm clustered impedance values of topologies at each frequency on the Smith chart, and elliptical envelopes were used to automatically define impedance constraints for pixel matching. This method automated the design process, shortened the design cycle, and delivered ultra-compact, high-efficiency PAs suitable for space-constrained applications. Results demonstrated a PA achieving 72\% PAE and 40 dBm output at 3.35 GHz, with matching network size of 0.2 λg × 0.2 λg. {\textcopyright} 2025 IEEE.",

keywords = "AI driven design, impedance matching, inverse design, pixelated design space, Power amplifiers",

author = "Jingyun Bi and Xinyu Zhou and Jing Xia and Wenjie Feng and Yanze Wang and Chan, \{Wing Shing\}",

note = "Full text of this publication does not contain sufficient affiliation information. With consent from the author(s) concerned, the Research Unit(s) information for this record is based on the existing academic department affiliation of the author(s).; 12th IEEE MTT-S International Wireless Symposium (IWS 2025) ; Conference date: 19-05-2025 Through 22-05-2025",

year = "2025",

doi = "10.1109/IWS65943.2025.11177942",

language = "English",

isbn = "979-8-3315-3802-6",

booktitle = "2025 IEEE MTT-S International Wireless Symposium (IWS 2025) - Proceedings",

publisher = "IEEE",

address = "United States",

}

Bi, J, Zhou, X, Xia, J, Feng, W, Wang, Y & Chan, WS 2025, AI-Driven Pixelated Matching Network Design with Automated Impedance Clustering Constraints for Power Amplifiers. in 2025 IEEE MTT-S International Wireless Symposium (IWS 2025) - Proceedings. IEEE, 12th IEEE MTT-S International Wireless Symposium (IWS 2025), Shaanxi, China, 19/05/25. https://doi.org/10.1109/IWS65943.2025.11177942

AI-Driven Pixelated Matching Network Design with Automated Impedance Clustering Constraints for Power Amplifiers. / Bi, Jingyun; Zhou, Xinyu; Xia, Jing et al.
2025 IEEE MTT-S International Wireless Symposium (IWS 2025) - Proceedings. IEEE, 2025.

Research output: Chapters, Conference Papers, Creative and Literary Works › RGC 32 - Refereed conference paper (with host publication) › peer-review

TY - GEN

T1 - AI-Driven Pixelated Matching Network Design with Automated Impedance Clustering Constraints for Power Amplifiers

AU - Bi, Jingyun

AU - Zhou, Xinyu

AU - Xia, Jing

AU - Feng, Wenjie

AU - Wang, Yanze

AU - Chan, Wing Shing

N1 - Full text of this publication does not contain sufficient affiliation information. With consent from the author(s) concerned, the Research Unit(s) information for this record is based on the existing academic department affiliation of the author(s).

PY - 2025

Y1 - 2025

N2 - This paper presented a fully automated inverse design method for power amplifiers (PAs) based on specific pixelated design spaces. Traditional PA design relied on manual Load-Pull analysis and late-stage manufacturability tuning, while defining impedance constraints remained a challenge for automation. A layout-level Electromagnetic (EM) model was incorporated early in the design process, with a fixed-size pixelated space allocated for the matching network. Various ideal microstrip topologies served as placeholders for pixel to explore impedance information with the goals of achieving the desired amplifier performance. The DBSCAN algorithm clustered impedance values of topologies at each frequency on the Smith chart, and elliptical envelopes were used to automatically define impedance constraints for pixel matching. This method automated the design process, shortened the design cycle, and delivered ultra-compact, high-efficiency PAs suitable for space-constrained applications. Results demonstrated a PA achieving 72% PAE and 40 dBm output at 3.35 GHz, with matching network size of 0.2 λg × 0.2 λg. © 2025 IEEE.

AB - This paper presented a fully automated inverse design method for power amplifiers (PAs) based on specific pixelated design spaces. Traditional PA design relied on manual Load-Pull analysis and late-stage manufacturability tuning, while defining impedance constraints remained a challenge for automation. A layout-level Electromagnetic (EM) model was incorporated early in the design process, with a fixed-size pixelated space allocated for the matching network. Various ideal microstrip topologies served as placeholders for pixel to explore impedance information with the goals of achieving the desired amplifier performance. The DBSCAN algorithm clustered impedance values of topologies at each frequency on the Smith chart, and elliptical envelopes were used to automatically define impedance constraints for pixel matching. This method automated the design process, shortened the design cycle, and delivered ultra-compact, high-efficiency PAs suitable for space-constrained applications. Results demonstrated a PA achieving 72% PAE and 40 dBm output at 3.35 GHz, with matching network size of 0.2 λg × 0.2 λg. © 2025 IEEE.

KW - AI driven design

KW - impedance matching

KW - inverse design

KW - pixelated design space

KW - Power amplifiers

UR - https://www.scopus.com/pages/publications/105019497022

UR - https://www.scopus.com/record/pubmetrics.uri?eid=2-s2.0-105019497022&origin=recordpage

U2 - 10.1109/IWS65943.2025.11177942

DO - 10.1109/IWS65943.2025.11177942

M3 - RGC 32 - Refereed conference paper (with host publication)

SN - 979-8-3315-3802-6

BT - 2025 IEEE MTT-S International Wireless Symposium (IWS 2025) - Proceedings

PB - IEEE

T2 - 12th IEEE MTT-S International Wireless Symposium (IWS 2025)

Y2 - 19 May 2025 through 22 May 2025

ER -

AI-Driven Pixelated Matching Network Design with Automated Impedance Clustering Constraints for Power Amplifiers

Abstract

Conference

Bibliographical note

Research Keywords

Access Output

Other Access Options

Permanent Link

Other Files and Links

Fingerprint

Cite this