Associate Professor Dr. Patrick Goh Kuan Lye
Last Updated on 05.02.2024
PhD MSc BSc Tel : 04 - 599 6033 Researcher ID : ABB-9389-2020 Last Updated on 19.06.2023 Latest publications can be viewed at the Google Scholar link below: Application of Artificial Intelligence for Signal Integrity Analysis MLP based ANN with an input layer, one hidden layer and an output layer. Modeling and Simulation of Power Distribution Networks on IC and PCB PDN modeling using triangular elements. A New and Fast Algorithm for Characterizations of High-Speed Interconnects Others: Prospective students who would like to pursue an MSc or PhD either full-time or part-time are invited to contact me at Last Updated on 03.02.2022Personal Info
University of Illinois at Urbana-Champaign (US)
University of Illinois at Urbana-Champaign (US)
University of Illinois at Urbana-Champaign (US)
Email :
ORCID : 0000-0002-4514-5065
Scopus ID : 36004664100
GoogleScholar ID : NEnFubYAAAAJ
USM Expertise Link : https://experts.usm.my/cvitae/eepatrickResearch Interest
Signal and power integrity
Modeling and simulation methods for interconnects and packages
High-speed and high-frequency digital signaling
Artificial intelligence and machine learning based methods
RF and microwave devices
Selected Publications
Google ScholarResearch Topics
Artificial intelligence (AI) techniques take advantage of advancements in machine learning to solve many challenging problems in engineering. Artificial neural networks (ANN) is an AI based algorithm which rely on a learning or training process where the algorithm is first fed with a well-established training set of data. The algorithm then, through a corrective process, adjusts its internal parameters such that it is able to solve subsequent inputs even though it has never encountered those before. An example structure of a multilayer perceptron (MLP) artificial neutral network is shown below. This research aims to apply ANN and other AI based methods to analyze signal quality in high-speed systems.
The analysis of power distribution networks (PDN) on IC and PCB is vital to ensure proper operations of digital circuits. In addition, poor power distribution can also impair the quality of the high-speed signals through effects such as simultaneous switching noise (SSN), power ripples and ground bounce. Research on PDN modeling involves geometry discretization (normally using triangular or tetrahedral elements), followed by a circuit element extraction using algorithms such as those based on the Delaunay-Voronoi triangulation. Then, DC and transient PDN analysis can be carried out on the developed model. This research aims to combine existing modeling methods with a novel circuit simulation algorithm that synergizes with the PDN modeling process.
Read the featured interview in IET Electronic Letters (vol. 52, issue 23) here:
http://ieeexplore.ieee.org/document/7728336/
Interconnect simulations for signal integrity analysis is a vital component to ensure proper operation of modern integrated circuit (IC) chips. However, interconnect simulations remain a challenge in the industry today due to the sheer large size of modern complex interconnect schemes. For example, it is common for a complete interconnect netlist to contain millions of nodes. When simulated in a conventional circuit simulation method such as SPICE, a single simulation could take hours or even days. This research attempts to find a new and novel idea for the characterization and simulation of high speed interconnects.
On-going projects include work on ensuring the stability of the algorithm, and improving the accuracy through the use of novel integration methods such as the Verlet integration and higher order Runge-Kutta methods.
In addition to the above projects, I also have topics in the following areas:
RF and microwave devices
Thermal issues in integrated circuitsPostgraduate Opportunities