MENU 
2021 SURE/SROP Research Projects in ECE
Directions: Below are listed the most recent descriptions of 2021 Summer Undergraduate Research in Engineering (SURE) and Summer Research Opportunity Program (SROP) projects available in Electrical and Computer Engineering (ECE). Please consider this list carefully before applying to the SURE or SROP program. You are welcome to contact faculty if you have additional, specific questions regarding these projects.
*IMPORTANT*: In addition to their online application, SURE applicants for ECE projects must also submit a resume and statement explaining their interest in and qualifications for the project that most interests them, including why they want to work on the project, the relevant skills they bring, and what they expect from their experience. The statement should be no longer than one page (12 point font and 1” margins) and must be uploaded in “other” at the bottom of the online application. Applications without this information may not be considered. Please include your name and UMID on all documents submitted.
SROP applicants for ECE projects should follow the specific directions outlined in the online application.
| Research Area | Project Number |
|---|---|
| Applied Electromagnetics & RF Circuits | 1, 2 |
| Control Systems | 3, 4 |
| Integrated Circuits & VLSI | 5, 6 |
| Optics & Photonics | 7 |
| Power & Energy | 8, 9, 10 |
| Robotics | 11 |
| Signal & Image Processing and Machine Learning | 12, 13, 14 |
| Solid State & Nanotechnology | 15, 16, 17, 18, 19 |
Applied Electromagnetics & RF Circuits
ECE Project 1: Extreme Electromagnetics/Optics with Metasurfaces
Faculty Mentor:
Anthony Grbic
[email protected]
Course format:
Hybrid
Prerequisites:
EECS 230 required. EECS 330 preferred.
Description:
The research area of metamaterials has captured the imagination of scientists and engineers over the past two decades by allowing unprecedented control of electromagnetic waves. The extreme manipulation of electromagnetic waves has been made possible by the fine spatial control and wide range of material properties that can be attained through subwavelength structuring/patterning. Research in this area has resulted in devices which overcome the diffraction limit, render objects invisible, and even break time reversal symmetry. It has also led to flattened and conformal optical systems and ultra-thin antennas.
Electromagnetic metasurfaces are finely patterned surfaces whose intricate patterns/textures dictate their electromagnetic properties. Conventional field-shaping devices, such as lenses in prescription eye glasses or a magnifying glass, require thickness (propagation length) to manipulate electromagnetic waves through interference. In contrast, metasurfaces manipulate electromagnetic waves across ultra-thin thicknesses through surface interactions.
In this project, the student researcher will analyze, characterize and design metasurfaces, and test experimental prototypes at millimeter-wave frequencies. These prototypes will be fabricated using additive manufacturing techniques. The student will become well versed in industry-standard microwave/electromagnetic CAD packages. He/she will use these packages to simulate metasurface designs, and develop their own metasurfaces. The metasurfaces will be characterized in the laboratory using a Gaussian beam measurement system. The student will study wave propagation and develop analytical/theoretical models for complex 2D structures. He/she will employ modern techniques for the synthesis of millimeter-wave and photonic devices.
ECE Project 2: Space-Time Modulated RF/Microwave Circuits
Faculty Mentor:
Anthony Grbic
[email protected]
Course format:
Hybrid
Prerequisites:
EECS 215, EECS 216, & EECS 230.
Description:
Space-time modulation has attracted renewed interest within the fields of radio frequency (RF) circuits, applied electromagnetics and optics in recent years. Progress in the availability and performance of tunable semi-conductor devices as well as electro-/magneto-optic, phase change, and 2D materials have drawn researchers to examine the modulation of electronic circuits and electromagnetic/optical devices in both space and time. Spatio-temporal modulation enables filtering (n-path circuits), frequency conversion, parametric amplification, and more recently non-reciprocity (one-way transmission).
In this project, the student will develop an in-depth understanding of high frequency circuits, as well as methods to analyze time-modulated and space-time modulated RF circuits. He/she will then apply this knowledge to the realization of novel RF components such as sub-harmonic frequency translators, and magnet-free isolators and circulators. The student will gain a working knowledge of harmonic balance circuit solvers and RF CAD packages, as well as hands-on microwave measurement and high frequency circuit characterization experience. The student will work alongside graduate researchers to develop time and space-time modulated circuits and systems. They will develop space-time modulated RF circuits for various wireless applications relevant to 5G wireless systems, full duplex communications, and radar.
Control Systems
ECE Project 3: Explaining neural networks with optimization and control theory
Faculty Mentor:
Necmiye Ozay
[email protected]
Course format:
Student preference (in person/hybrid/remote)
Prerequisites:
Strong analytical skills, knowledge of linear algebra, basic knowledge of numerical optimization (e.g., linear programming or convex programming), programming experience in MATLAB or Python, familiarity with neural networks particularly for computer vision tasks is a plus.
Description:
Many modern autonomous systems involve machine learning components. Specifically, deep neural networks are commonly used for perception tasks when the autonomous system is equipped with a camera. For instance, a self-driving car can predict its distance to nearby cars, pedestrians, and obstacles using a neural network. However such machine learning models can fail in unpredictable ways risking the safetyof the overall autonomous system. This project aims to develop new ways to explain a trained neural network using ideas from optimization and control theory.
ECE Project 4: Privacy and Security of Cyber and Cyber-Physical Control Systems
Faculty Mentor:
Stéphane Lafortune
[email protected]
Course format:
Remote
Prerequisites:
Programming experience in Python, C, C++, or Java required
Description:
We are developing methodologies for (i) privacy enforcement in cyber systems and (ii) detection and mitigation of sensor and actuator attacks at the supervisory control layer of cyber-physical control systems. The student intern will work on implementing and testing the algorithmic procedures of these methodologies, either as stand-alone procedures or as part of our existing M-DES software tools (see: https://gitlab.eecs.umich.edu/M-DES-tools). The student will also work on the development of case studies to test these algorithms, as well as on visualization tools for illustrating these case studies. The student will work in close collaboration with graduate students.
Integrated Circuits & VLSI
ECE Project 5: Accelerating Whole Genome Sequencing
Faculty Mentor:
David Blaauw
[email protected]
Course format:
Hybrid
Prerequisites:
Interest in biomedical algorithm design, FPGA implementation, VLSI Verilog chips design. Responsibilities will depend on candidate background.
Description:
Whole genome sequencing (WGS) determines the complete DNA sequence of an organism’s genome. While it cost nearly $3 billion to sequence the first human genome in 2001, just over the last one decade, the production cost of sequencing has plummeted from ten million dollars to thousand dollars, making it a promising tool for individualized treatment plans and precision health. In the sequencing pipeline, hundreds to thousands of CPU hours of intensive computation needs to be performed on raw data to sequence one genome, which are opportunities for software algorithm optimization and hardware acceleration. In this project, we are developing novel algorithms tailored for hardware acceleration to speed up the entire secondary analysis for a number of emerging genomics applications, such as whole genome sequencing, RNA single cell sequencing, and microbiome analysis. We aim to build a heterogeneous hardware system to speed up the current software pipeline by 1000x or more using algorithm and FPGA/ASIC implementation co-design. The work will depend on the background of the candidate and can involve both software / hardware development as well as exploring new genomics applications.
ECE Project 6: Millimeter-scale Sensor System
Faculty Mentor:
David Blaauw
[email protected]
Course format:
Hybrid
Prerequisites:
Interest in embedded systems, circuit design. Responsibilities will depend on candidate background.
Description:
In this project we are looking for a student to help us with embedded hardware and software development of mm-scale sensor systems for Internet of Things (IoT) applications. In the last few years we have prototyped the world’s first complete and functional mm-sized embedded systems. The system incorporated a commercial ARM Cortex M0 processor with low leakage memory, ultra-low power flash, battery, harvesting and RF communication. Early versions sensed temperature and pressure and more recent versions can also record audio and images. This work is currently featured at the Computer History Museum in California as the world’s smallest computer and also in the atrium of the EECS building. Our team is currently working actively with other researchers to deploy the sensors for butterfly tracking, down-hole oil-reservoir exploration and medical implantable applications as well as commercializing the technology. We are currently working to expand our sensor capabilities to include new sensing modalities, better interfaces and increasing radio range. The student work will depend on the background of the candidate and can include embedded software development for low power operation, GUI design, development of new sensor applications, help with digital and mixed signal circuit design, and testing and diagnosis of fabricated chips.
Optics & Photonics
ECE Project 7: Using Smart Phones for Contactless Physiological Measurements
Faculty Mentor:
Mohammed Islam
[email protected]
Course format:
Hybrid
Prerequisites:
EECS 434 and some experience in app development (Android or Swift) preferred.
Description:
Undergraduate students will be working with our group (Professor and graduate students) to develop applications on smart phones for contactless physiological measurements for health screening. The parameters measured will include heart rate, respiratory rate, heart rate variability (proportional to stress), and potentially blood pressure. The measurements use the cameras included on some of the latest flagship smart phones, such as the iPhone 12 Pro or Samsung s20+. In particular, the wide-angle cameras (i.e. typical RGB cameras) output will be registered with the new time-of-flight, ToF, cameras (often called LiDAR), which use vertical cavity surface emitting lasers and direct or indirect ToF cameras to measure depth. The information from the ToF cameras is used to compensate for motion artifacts in the health tracking measurements, and machine learning or artificial intelligence processing leads to the health parameters. By using the smart phone app, participants will be able to build their own personal baseline health metrics, which can be used to then be used to predict the onset of health ailments, such as upper respiratory problems, including influenza and COVID-19. Other applications of the technology include health screening at building or facility entrances, as well as vehicle driver monitoring in advanced driver assistant systems.
Power & Energy
ECE Project 8: Advanced Cooling Methods for Power Electronics
Faculty Mentor:
Al-Thaddeus Avestruz
[email protected]
Course format:
Hybrid
Prerequisites:
Basic chemistry, thermodynamics, and heat transfer. Interest in learning how to build circuits and instrumentation. Labview and Python a plus.
Description:
We will be exploring new ways of cooling very high density power electronics with new types of fluids.
ECE Project 9: Energy Storage Systems
Faculty Mentor:
Al-Thaddeus Avestruz
[email protected]
Course format:
Hybrid
Prerequisites:
Python, C/C++, Some Embedded Programming
Description:
Future energy storage systems will be networked and heterogeneous. Digital control and optimization is needed for the best utilization of energy storage resources using power electronics. We will be designing and testing real-time control and optimization algorithms for interconnected energy storage.
ECE Project 10: Enhancing Power Electronics Education through Classroom Demonstrations
Faculty Mentor:
Al-Thaddeus Avestruz
[email protected]
Course format:
Hybrid
Prerequisites:
Must be comfortable building circuits and hardware; experience with 3d printing and laser cutting preferred
Description:
We will be designing and building a number of power electronics demonstrations for classroom use. These include induction heating, wireless power transfer, fluorescent ballast, etc.
Robotics
ECE Project 11: Machine Learning for Robot Motion Planning
Faculty Mentor:
Dmitry Berenson
[email protected]
Course format:
Hybrid
Prerequisites:
EECS 281 or significant programming experience.
Description:
This project focuses on exploring machine learning methods for use in robot motion planning. The project will begin by implementing and testing existing baseline algorithms for learning dynamics models and constraints for use by a motion planner. Then we will explore how to build and improve on the state-of-the-art methods to enable faster and more accurate planning of robot motion. Example applications will include manipulating deformable objects such as cloth and rope.
Signal & Image Processing and Machine Learning
ECE Project 12: Learning-based Methods for Supper Resolution in Microscopy Imaging
Faculty Mentor:
Qing Qu
[email protected]
Course format:
Hybrid
Prerequisites:
None.
Description:
Extracting fine-scale information from low-resolution data is a major challenge in many areas of the applied sciences. In microscopy, astronomy and any other application employing an optical device, spatial resolution is fundamentally limited by diffraction. A popular model for the data-acquisition process in such cases is the convolution of the signal of interest with a point-spread function (PSF) that blurs the fine-scale details, acting essentially as a low-pass filter. The problem of super-resolution is that of deconvolving the original image from the blurred measurements, which is a type of inverse problem.
On the other hand, deep learning has provided a new paradigm for solving such inverse problems in imaging sciences. Unlike traditional optimization-based approaches, which require precise knowledge of the measurement and noise models, deep-learning approaches make it possible to learn these models implicitly from training data, yielding a methodology that is robust to arbitrary noise statistics, aberrations, and non-linear forward models. Recent works report promising results for deconvolution of point sources in super-resolution microscopy imaging. Calibrating these models requires minimizing a highly nonconvex cost function. Unfortunately, there is essentially no theoretical understanding of this methodology. The goal of the proposed research is to a) investigate and understand such learning-based methods, and b) develop novel approach. To complement our investigations, we will apply the described techniques to real fluorescence-microscopy data.
ECE Project 13: New Machine Learning Algorithms
Faculty Mentor:
Clay Scott
[email protected]
Course format:
Remote
Prerequisites:
Desired qualifications: solid background in probability and linear algebra, proficiency in Matlab or Python, prior exposure to machine learning such as EECS 445 or Stats 415.
Description:
This project will involve developing and/or evaluating a new machine learning algorithm that addresses a fundamental shortcoming of some existing method. The focus is on developing a general purpose algorithm that would be useful in a number of different contexts, and demonstrating that algorithm on synthetic and real world data. The precise topic will depend on the current state of research in my group next summer and student interest, but current topics of interest include loss functions for multiclass classification, manifold learning, clustering with side information, domain adaptation, and weakly supervised learning. Multiple projects may be available.
ECE Project 14: Reinforcement Learning Algorithms for the Google Research Football Environment
Faculty Mentor:
Lei Ying
[email protected]
Course format:
Remote
Prerequisites:
Familiar with Python, PyTorch (or Tensorflow). Familiar with RL algorithms such as Deep Q-learning and Deep Policy-Gradient algorithms and their implementations.
Description:
The objective of this project is to develop single-agent and multi-agent RL algorithms for the Google Research Football environment. Students involved in the project will work with PhD students to develop new RL algorithms and evaluate them.
Solid State & Nanotechnology
ECE Project 15: Structural characterization of (In,Ga)N
Faculty Mentor:
Elaheh Ahmadi
[email protected]
Course format:
In person
Prerequisites:
Some basic understanding of semiconductors.
Description:
In this project, the SURE student will help with structural characterization of (In,Ga)N films. This includes performing atomic force microscopy (AFM), X-ray diffraction (XRD), and scanning electron microscope (SEM) on samples prepared by the graduate student. The student needs to have some basic understanding of semiconductors. Some experience with material characterization is preferred.
ECE Project 16: Improving the Quality of Inkjet-Printed Nanostructure with Interpretable Convolutional Neural Networks and Reinforcement Learning
Faculty Mentor:
L. Jay Guo
[email protected]
Course format:
Hybrid
Prerequisites:
Machine learning courses
Description:
Inkjet printing has been applied for fabricating functional nanostructures and devices, such as antennas and transistors. However, existing inkjet printing methods suffer from instability and low uniformity. Actively and optimally adapt multiple printing control parameters is the key challenge that hinders high-quality inkjet printing. In this project, we aim to gain an understanding of the relationship between printing parameters and printed structures with interpretable convolutional neural networks. We have collected a large image dataset that contains microscope images of the printed structures. We will train interpretable convolutional neural networks to predict the printing control parameters. With this approach, we hope to learn the correlation between the printed structures and printing parameters. The results will be interpretable by domain experts to facilitate scientific discoveries. With the gained knowledge, we will further develop reinforcement learning methods to actively control the printing process to achieve a better printing quality.
ECE project 17: Meta-Learning for Metasurface Inverse Designs
Faculty Mentor:
L. Jay Guo
[email protected]
Course format:
Hybrid
Prerequisites:
Machine learning courses, optics/photonics/EM courses
Description:
Metasurfaces are advanced optical devices that can be applied in many areas including but not limited to super-resolution, optical computing, VR and AR, and optical communications. However, the vast amount of different applications of metasurfaces all require human experts to come up with a specific design, which could be time consuming and suboptimal. The inverse design has recently been applied for automatic metasurface design. However, inverse design approaches usually require the users to train neural networks on a large dataset for each new task, which could be infeasible due to the time required for collecting the data. Rather than dealing with each task separately, we propose to develop meta-learning methods that can learn the shared structures of metasurface design tasks. We aim to train inverse design models with meta-learning approaches so that the trained model can inversely design near-optimal structures when only given a small dataset for a new design task. The proposed method could speed up the existing metasurface design pipeline.
ECE Project 18: Artificial Photosynthesis
Faculty Mentor:
Zetian Mi
[email protected]
Course format:
Hybrid
Prerequisites:
None
Description:
This project is related to the design, synthesis, and characterization of semiconductor nanostructures for artificial photosynthesis production of clean fuels directly from sunlight and carbon dioxide. The student will gain insight and contribute to the development of state of the art of solar energy devices and systems. This is a highly interdisciplinary project, and the student will have the opportunity to work with experts in electrical engineering, chemical engineering, and materials science and engineering.
ECE Project 19: Ultraviolet-C Light Emitting Diodes
Faculty Mentor:
Zetian Mi
[email protected]
Course format:
Hybrid
Prerequisites:
None
Description:
This project is related to the research and development of UV-C (<280 nm) LEDs, which are critically important for water purification, disinfection, and medical diagnostics. Students involved in this project will work together with PhD students on the design and characterization of semiconductor nanostructures, as well as the processing and packaging of UV-C LED devices and device testing.