Seminar

Seminar - Dr. Faycal Saffih

Phil Regier — Thu, 14 Jul 2016 14:41:32 +0000

Seminar - Dr. Faycal Saffih Phil Regier Thu, 07/14/2016 - 10:41

Speaker

Dr. Faycal Saffih

Topic

Intelligence Design (ID): From System to Circuit to Device Implementations

Abstract

In this talk we present the development of intelligence (vs intelligent) implementations from top-down and bottom-up approaches and from Electrical engineering design and Biological Biomimicry to Solid-state Physics prediction. Inspired by the Human Visual System, we postulated that intelligent vision requires embedding intelligence at the image signal acquisition and along the path to the visual cortex and beyond when involving image recall and pattern recognition. A feedback loop of this path might be deemed necessary if a full fetched image understanding and “intelligence extraction” is required. To start the above path, Smart CMOS imaging was chosen as the application of choice where multi-disciplinary research tracks need to interact suggesting a novel approach for research to design intelligent imaging system. We will provide our experience of this endeavor hoping to inspire other researchers to continue this promising track to build Artificial Intelligent Imaging (AI2) systems for a wide variety of applications where human vision needs assistance from an intelligent imaging system for biomedical, automotive and security applications to name a few.

Biography

Dr. Faycal Saffih (IEEE Member since 2000) received the B.Sc. (with Best Honors) degree in Solid-State Physics from the University of Sétif-1, Sétif, Algeria, in 1996, the M.Sc. degree in Digital Neural networks from Physics Department, University of Malaya, Kuala Lumpur, Malaysia, in 1998, and the Ph.D. degree in Smart CMOS Imaging from Electrical and Computer Engineering Department, University of ��ݮ��Ƶ, ��ݮ��Ƶ, ON, Canada. Taking a decade journey between academia and industry, Dr. Saffih enriched his experience multi- dimensionally spanning Microelectronics from devices up-to systems, and industry from R&D department to Entrepreneurship start-up, all of which from West USA (OR) to Singapore’s prestigious A*star Agency for Science, Technology and Research. Recently, Dr. Saffih endeavored into renewable energy research and business starting from Stanford certification in 2013 and currently undertaking an Online program from Renewables Academy (RENAC), Germany. Dr. Faycal Saffih is currently on the faculty of the Electrical Engineering Department of the UAE University and a regular visiting scholar at the University of ��ݮ��Ƶ, University of Alberta among others. His research is on intelligence extraction and implementation on devices and systems particularly smart CMOS image sensors.

Seminar - Professor Bram Nauta, University of Twente

Phil Regier — Thu, 05 May 2016 13:50:05 +0000

Seminar - Professor Bram Nauta, University of Twente Phil Regier Thu, 05/05/2016 - 09:50

Title

Circuit Techniques for Next-Generation Wireless Communication

Speaker

Professor Bram Nauta, University of Twente

Abstract

Due to the increase of wireless standards using different RF frequencies there is a need to have transceivers that can handle a wide range of RF frequencies. By abandoning the classical narrowband approach, new receiver architectures are explored in which noise and interferer robustness problems have to be solved. At the same time new features are wanted such as spectrum sensing for cognitive radio and self-interference cancelling for future full duplex communication. In this presentation several circuit and system techniques will be illustrated that may enable future radio systems.

Biography

Bram Nauta received the M.Sc and Ph.D. degrees in electrical engineering from the University of Twente, Enschede, the Netherlands. In 1991 he joined Philips Research, Eindhoven, and in 1998 he returned to the University of Twente. He is currently a distinguished professor, heading the IC Design group. Bram is an IEEE Fellow and has served as the Editor-in-Chief of the IEEE Journal of Solid-State Circuits and as program chair of the International Solid State Circuits Conference (ISSCC). Since 2016 he is Vice-President of the IEEE Solid-State Circuits Society.

Seminar - Dr. Earl McCune - The Physics of OFDM

Phil Regier — Thu, 24 Sep 2015 18:34:08 +0000

Seminar - Dr. Earl McCune - The Physics of OFDM Phil Regier Thu, 09/24/2015 - 14:34

Speaker

Dr Earl McCune, Visiting Faculty, TU Delft

Topic

The Physics of OFDM

Abstract

Orthogonal Frequency Division Modulation (OFDM) is an extremely revolutionary signal technology that actually is poorly understood by both academia and industry. Academia appears to not understand how revolutionary this signal actually is, and what the corresponding physical difficulties of its implementation are. On the industrial side, there is a general misunderstanding of the economic consequences of building hardware needed to generate and receive an OFDM signal, along with the problems that it solves and creates. In my entire career I have never encountered a signal type that has such a bimodal interest distribution – some adherents love it, and there are others who cannot loathe it more. The best way to examine this situation is to build up OFDM technology from first physical principles, in order to clearly understand what OFDM is, and isn’t.

Speaker's biography

Earl received his Bachelors, Masters, and Doctorate degrees at UC Berkeley, Stanford, and UC Davis respectively. His experience in RF circuits, signals, and systems goes back more than 40 years. Within this career he has founded two Silicon Valley startups; the first one doing modulated direct digital frequency synthesis in 1986 and merged with Proxim in 1991. The second start-up, Tropian, did switch-based efficient RF transmitters from 1996 and was acquired by Panasonic 10 years later. He retired from Panasonic in 2008 as a Corporate Technology Fellow. He now serves as visiting faculty at TU Delft.

Seminar - Dr. Earl McCune - Silicon Valley Style: Project Operations in the Presence of Extreme Risk

Phil Regier — Thu, 24 Sep 2015 18:28:54 +0000

Seminar - Dr. Earl McCune - Silicon Valley Style: Project Operations in the Presence of Extreme Risk Phil Regier Thu, 09/24/2015 - 14:28

Speaker

Dr Earl McCune, Visiting Faculty, TU Delft

Topic

Silicon Valley Style: Project Operations in the Presence of Extreme Risk

Abstract

The name ‘Silicon Valley’ is often synonymous with innovation and technology. This is not an accident, because the style developed in Silicon Valley is unusual in that it embraces project risks and uses them as tools. This impacts not only project organization, but also how teams and businesses are structured. A partnership between business operations and technologists is fundamental to the success of this model: neither is superior to the other.

Speaker's biography

MASc Seminar Notice: Approaching Memorization in Large Language Models

Andrea Pinos — Fri, 22 Aug 2025 11:51:17 +0000

MASc Seminar Notice: Approaching Memorization in Large Language Models Andrea Pinos Fri, 08/22/2025 - 07:51

Candidate: Xiao Cheng

Date: August 29, 2025

Time: 9:00am

Location: EIT 3145

Supervisor: Dr. Weiyi Shang

All are welcome!

Abstract:

Large Language Models (LLMs) risk memorizing and reproducing sensitive or proprietary information from their training data. In this thesis, we investigate the behavior and mitigation of memorization in LLMs by adopting a pipeline that combines membership inference and data extraction attacks, and we evaluate memorization across multiple models. Through systematic experiments, we analyze how memorization varies with model size, architecture, and content category. We observe memorization rates ranging from 42% to 64% across the investigated models, demonstrating that memorization remains a persistent issue, and that the existing memorization-revealing pipeline remains valid on these models. Certain content categories are more prone to memorization, and realistic usage scenarios can still trigger it. Finally, we explore knowledge distillation as a mitigation approach: distilling Llama3-8B reduces the extraction rate by approximately 20%, suggesting a viable mitigation option. This work contributes a novel dataset and a BLEU-based evaluation pipeline, providing practical insights for research on LLM memorization.

MASc Seminar Notice: Personalized Autonomous Driving and Motion Planning with Nonconvex Trajectory Optimization using Trajectory Sensitivities

Andrea Pinos — Fri, 22 Aug 2025 11:46:44 +0000

MASc Seminar Notice: Personalized Autonomous Driving and Motion Planning with Nonconvex Trajectory Optimization using Trajectory Sensitivities Andrea Pinos Fri, 08/22/2025 - 07:46

Candidate: Xiaofei Wu

Date: September 2, 2025

Time: 10:00am

Location: EIT 3145

Supervisors: Drs. Michael Fisher and Stephen L. Smith

All are welcome!

Abstract:

A key factor in increasing people’s level of acceptance of autonomous driving technology is trust. Personalized autonomous driving that mimics the driver’s own driving style is a viable approach. Many existing works explore learning-based approaches to achieve this goal but may be inadequate when dealing with unseen scenarios and enforcing safety guarantees. To mitigate these difficulties, we propose a hierarchical autonomous vehicle control framework, where the upper level mimics a target driver’s driving style and the lower level performs vehicle motion planning.

The lower-level motion planner solves a trajectory optimization problem, where nonlinear dynamics of the vehicle model makes the problem challenging. Many existing approaches formulate this as multistage programs and use derivatives of each stage to obtain a local approximation at each iteration. We develop a novel approach for obtaining improved local approximations using an input-to-state reformulation of system dynamics and trajectory sensitivities, which are derivatives of the entire system trajectory with respect to control inputs. The method is proved to converge with input-affine inequality constraints and is applied to generate trajectories for an autonomous vehicle in a variety of scenarios.

The upper-level driving style mimicking problem solves for weight factors that are used to parameterize the lower-level objective. We adopt a gradient-based approach to solve this problem. As differentiability is not guaranteed given the bilevel structure and the nonconvex lower level solution mapping, we use subgradients, which are generalizations of gradients for nondifferentiable functions, and a projected subgradient update algorithm to solve this problem. Simulations show that the proposed framework is capable of solving tracking and obstacle avoidance problems while mimicking driving style.

ECE Guest Seminar: Large Signal Analysis: Configuration, Calibration, Measurement, Data Analysis, and Design

Andrea Pinos — Wed, 20 Aug 2025 13:45:31 +0000

ECE Guest Seminar: Large Signal Analysis: Configuration, Calibration, Measurement, Data Analysis, and Design Andrea Pinos Wed, 08/20/2025 - 09:45

Speaker: Dr. Zacharia Ouardirhi, Senior Vice-President of Sales, Maury Microwave

Date: August 21, 2025

Time: 11:00am

Location EIT 3142

Host: Dr. Slim Boumaiza

All are welcome!

Abstract:

The presentation provides a comprehensive overview of large signal analysis techniques essential for the design and optimization of RF components such as transistors and power amplifiers. Emphasizing the non-linear behavior of active components under large signal excitation. Methodologies to enhance power efficiency, reduce signal distortion, and ensure compatibility with modern modulation schemes will be presented.

Biography:

Zacharia Ouardirhi received his PhD in Electrical Engineering from Ecole Polytechnique of Montreal in 2005 with an emphasis in RF & Microwave engineering.

He currently serves as Senior Vice-President of Sales at Maury Microwave, a leading company in RF and Microwave test and measurement solutions.

Over his career, Dr. Ouardirhi has been recognized for his expertise in device characterization. He is noted for participating in significant academic and industry events and for his collaborative spirit, regularly engaging with leading professors, researchers and Industry stakeholders in Canada and Internationally.

MASc Seminar Notice: Convex Reparameterizations for Efficient H2/Hinf Feedback Control

Andrea Pinos — Mon, 18 Aug 2025 16:33:19 +0000

MASc Seminar Notice: Convex Reparameterizations for Efficient H2/Hinf Feedback Control Andrea Pinos Mon, 08/18/2025 - 12:33

Candidate: Zhong Fang

Date: September 2, 2025

Time: 2:30pm

Location: EIT 3145

Supervisors: Dr. Michael Fisher

All are welcome!

Abstract:

The design of controllers with mixed H2/Hinf cost functions remains a challenging problem in control theory, with pervasive applications across diverse engineering fields. The main difficulties arise from nonconvexity and infinite dimensionality of the associated optimization problem for the design. Recently, several new approaches were developed to tackle nonconvexity by reparameterizing the variables to transform the optimization into a convex but infinite-dimensional formulation incorporating additional affine constraints in the design problem. For state feedback design, system level synthesis is focused since input output parameterization is primarily intended for output feedback. To make the problem tractable, and to address limitations of historical approximation methods, a new Galerkin-type method for finite-dimensional approximations of transfer functions in Hardy space with a selection of simple poles was recently developed.

However, prior applications of this simple pole approximation resulted in a design problem that required an additional approximation of a finite time horizon to compute H2 and Hinf norms for the closed-loop response. This finite horizon resulted in increased suboptimality, degraded performance, and increased problem size and memory requirements. To address these limitations, this thesis presents a novel control design framework that combines the frequency domain convex reparameterization affine constraints with a state space formulation of the H2 and Hinf norms using linear matrix inequality. This state space formulation eliminates the need for a finite time horizon approximation, and results in a convex and tractable semidefinite program for the control design. Suboptimality bounds are provided for the method which guarantee convergence to the global optimum of the infinite dimensional problem as the number of poles approaches infinity with a convergence rate that depends on the geometry of the pole selection.

The recently developed convex reparameterization methods have been challenging to adapt to continuous time control design in practice, because they typically rely on finite dimensional approximations for tractability that lead to numerical ill-conditioning or even closed-loop instability. In this work, the hybrid state space and frequency domain control design method is adapted to develop the first practical and tractable continuous time control design based on these convex reparameterizations that does not suffer from ill-conditioning and that guarantees closed-loop stability for stabilizable plants. Approximation error bounds are established for the first time for the simple pole approximation in continuous time. These bound the error based on the geometry of the pole selection, and show that this error goes to zero as the number of poles approaches infinity. These bounds are particularly challenging to obtain compared to the discrete time case due to the noncompactness of the domain of integration for computing the H2 and Hinf norms in continuous time. These approximation error bounds are then used to develop suboptimality guarantees of an analogous nature to those in discrete time. This is the first time that suboptimality bounds with zero asymptotic error have been developed for a control design method using these recent convex reparameterization approaches in continuous time. Again, the noncompactness represents a major challenge that must be overcome to establish these results.

There exist several recently developed convex reparameterizations for output feedback control design in discrete time (including system level synthesis and input output parameterization). However, all of these methods currently lack rigorous suboptimality guarantees that establish convergence to the solution of the infinite dimensional problem as the approximation dimension approaches infinity. This is largely due to the additional complexity introduced by the output feedback case compared to state feedback. This work develops novel output feedback control design methods using four different convex reparameterization approaches, each with different benefits and trade-offs, that result in convex and tractable control design formulations. Moreover, a single unified approximation theory is developed that simultaneously establishes suboptimality bounds for all four methods that recovers analogous results to the state feedback setting. In particular, they show a convergence rate to the global optimum that depends on the geometry of the pole selection in a similar fashion to the state feedback case.

The novel methods are applied to design controllers for power converter interfaced devices to provide frequency and voltage regulation to the power grid. Practical case studies demonstrate the ability of the methods to match desired dynamic behavior for these services. We consider multi-controller scenarios involving distributed energy resources where multiple power converters must coordinate to provide grid services while respecting physical and engineering constraints, including state, input, and output limits for each device, enabling coordinated control of dynamic virtual power plants. Case studies involving power converter voltage and frequency regulation, as well as multi-controller coordination in the IEEE 9-bus system, demonstrate superior performance.

MASc Seminar Notice: Polynomial Controllers for Optimal Trajectory Matching with Contractivity Guarantees

Andrea Pinos — Mon, 18 Aug 2025 14:04:26 +0000

MASc Seminar Notice: Polynomial Controllers for Optimal Trajectory Matching with Contractivity Guarantees Andrea Pinos Mon, 08/18/2025 - 10:04

Candidate: Alexander Kitaev

Date: September 2, 2025

Time: 11:00am

Location: EIT 3145

Supervisors: Dr. Michael Fisher and Dr. Christopher Nielsen

All are welcome!

Abstract:

The trajectory matching problem is a problem in control theory where a set of reference trajectories for a plant is given, and a control law that causes the plant's trajectories to be as close as possible to the reference trajectories is desired. This thesis presents an approach for solving the trajectory matching problem that generates explicit polynomial controllers. Additionally, the method presented in this thesis guarantees local contractivity of the generated controller.

This thesis presents several theoretical results that justify the method described here. Firstly, a proof that the local contractivity constraint can be expressed as a set of matrix inequalities is presented. Secondly, a theorem that describes how symmetries in the trajectory matching problem correspond to symmetries in its solution is presented and proven.

Finally, this thesis demonstrates the method it describes on two example problems motivated by real-world applications. The first of these is stabilization and disturbance recovery for a single-machine infinite-bus (SMIB) power system, and the second is a lane change manoeuvre for Dubin's vehicle, a simple vehicle model.

MASc Seminar Notice: Designing Legible and Predictable Robot Navigation in Crowded Spaces

Andrea Pinos — Thu, 14 Aug 2025 17:21:47 +0000

MASc Seminar Notice: Designing Legible and Predictable Robot Navigation in Crowded Spaces Andrea Pinos Thu, 08/14/2025 - 13:21

Candidate: Anna Moskalenko

Date: August 19, 2025

Time: 2:00 PM

Place: EIT 3141

Supervisor(s): Smith, Stephen L.

All are welcome!

Abstract:

The primary challenge in human-robot interaction, particularly in dynamic and crowded environments, is to design navigation that is both legible and predictable. For people to feel comfortable in the same environment as robots, their movements must be legible - ensuring quick understanding of the robot’s intentions - and predictable, meaning they align with human expectations. In this thesis, we introduce a learning-based navigation system that leverages a vector reward function to capture the dual objectives of legibility and predictability. Rather than relying on manually designed transitioning rules or fixed weighting parameters such as and, the reward function is learned from expert demonstrations generated by a planner (LPSNav) that blends between these objectives using a continuous parameter. Our approach does not attempt to replicate the planner’s handcrafted logic, but instead generalizes the emergent patterns in its trajectories through a supervised learning framework inspired by the structure of maximum entropy IRL. The proposed architecture includes:

an LSTM-based Robot Motion History Encoder,
a CNN-based Environment Encoder,
an LSTM-based Human History Encoder,
and a two-channel Reward Analysis Engine.

The system was evaluated both in simulation using the nuScenes dataset and in real-world trials using the Clearpath Jackal robot. In the user study, participants interacted with the robot in predefined navigation scenarios, and their feedback was used to assess the perceived clarity and predictability of the robot’s actions. Results show that our method generates more human-like trajectories and outperforms baseline models in scenarios requiring socially acceptable motion, such as intersections, sidestepping, and close-proximity passing. We provide both quantitative metrics (e.g., ADE and FDE) and qualitative visualizations demonstrating smooth and socially-aware navigation behavior. This work represents a step toward safer and more intuitive human-robot coexistence, offering a practical solution for real-world robotic deployment.