Nick Cardamone

Published Work

IEEE AP-S Conference Paper (July 2025)

Title: An Additive Approach to the Shape Synthesis of Microstrip Circuits and Antennas
Authors: Nick Cardamone, Amal Mohammed, Igor Aćimović, and Derek A. McNamara
Conference: IEEE Antennas and Propagation Symposium (AP-S), 2025

Overview:

This paper introduces an additive shape synthesis approach for designing microstrip circuits and antennas. Instead of relying on predefined elements, this method builds RF/microwave layouts dynamically by iteratively adding conducting shapes within a defined space.

An evolutionary algorithm optimizes these layouts based on full-wave Computational Electromagnetics (CEM) simulations, refining the geometry to meet performance targets. The approach was successfully applied to a bandpass filter, demonstrating its effectiveness in generating high-performance layouts without traditional RF design heuristics.

Future work will extend this method to matching networks, power dividers, and arbitrary transmission line structures, making data-driven RF circuit synthesis more accessible.

---

Class Paper: Neural Network Buck Converter Control (Winter 2022)

Title: Controlling a Buck Converter with a Neural Network
Author: Nick Cardamone
Course: Modern Control Systems, University of Ottawa

Overview:

This paper explores the feasibility of controlling a buck converter using a neural network instead of traditional PID control. A NARMA L2 controller from MATLAB's Deep Learning Toolbox was used in simulation, showing that an intelligent control system is possible.

Key insights:

• Recurrent Neural Networks (RNNs) proved more effective than feedforward networks for controlling a dynamic system.
• Power consumption is a critical factor when implementing NN-based control on hardware like microcontrollers, FPGAs, or ASICs.
• The NARMA L2 controller was functional but required significant tuning, and further research is needed to improve practical implementation.

---

Class Paper: SEPIC Converter LED Driver (Fall 2022)

Title: SEPIC Converter as a Constant Current Controller for a LED Desk Lamp
Author: Nick Cardamone
Course: Electronics III, University of Ottawa

Overview:

This project involved designing and building a constant current LED driver using a Single-Ended Primary Inductor Converter (SEPIC). The converter was optimized through:

• MATLAB scripting for numerical component selection.
• LTSPICE simulations for validation.
• PCB prototyping and experimental testing.

Results showed the SEPIC converter successfully regulated LED current, but challenges included undesirable frequency deviation and suboptimal efficiency (max 72%). Future improvements could focus on reducing switching losses and enhancing thermal management.