Projects Spring 2023 | Ziad Youssfi

WHERE--Wrist-based HAR (Human Activity Recognition) with Embedded Real-time Equipment

Omkar Savkur, Nicholas Toldalagi, and Kevin Xie

Posted on May 10, 2023

With the rise of wearable personal electronics such as the smartwatch, more and more users have a capable edge computing device and sensing suite at their fingertips. Despite this, few projects have investigated how to perform continuous, power-efficient Human Activity Recognition (HAR) on the watch beyond traditional fitness tracking/pedometers. In... [Read More]

Sound-Based Trash Categorization (SBTrashCat)

Jacky Wang and Gordonson Yan

Posted on May 10, 2023

The key feature of SBTrashCat is to use the acoustic fingerprint of containers made with different materials to determine their categorization. For our baseline, SBTrashCat can identify containers including aluminum cans, plastic bottles, and tetra-pak. To ensure the robustness of the system, noise and human voice were also added in... [Read More]

Multinode Embedded Deep Learning

Jong-Ik Park, Chad Taylor, and Anudeep Bolimera

Posted on May 10, 2023

The rapid expansion of the Internet of Things (IoT) and the growing number of edge devices have significantly increased the amount of data generated and processed. This shift has led to a rising demand for efficient and cost-effective embedded systems to handle various tasks, including anomaly detection. High-specification devices, while... [Read More]

Motor Care - Condition Based Monitoring

Abhishek Basrithaya and Yuyang Xu

Posted on May 10, 2023

An electric motor is a rudimentary electro-mechanical device that is ubiquitous from wheels of an automobile, fans, automatic windows shades to magnetic disks for memory storage etc. Efficient operation of these fundamental devices is imperative for the smooth operation of the overarching device in question. Timely maintenance of these motors... [Read More]

Inventory Management on Edge (Empty Shelf Percentage Estimator)

Netra Trivedi, Rishi Pachipulusu, and Cathy Tungyun

Posted on May 10, 2023

In a warehouse or a retail setup: Given a frame of reference, our model indicates the percentage of empty spaces with reference to a shelf full of boxes (either rectangular or square) for new inventory. [Read More]

Feel the Vibe (Emotion Recognition)

Samuel Leong and Alex Strasser

Posted on May 10, 2023

Wearables are getting increasingly popular and feasible as computing power becomes smaller and more powerful. Indeed, the advent of advanced sensors and intelligence being packed into smaller packages opens up new opportunities for novel applications. In “Feel the Vibe”, we aim to create a wearable that is able to display... [Read More]

Dog Fitbit

Ajith Potluri, Eion Tyacke, and Parker Crain

Posted on May 10, 2023

This project investigates on-device deep learning models paired with an embedded system that can be used to track the movement of a dog for its owner. This valuable insight may help an owner to better understand the activities of the dog while the owner is away. The microcontroller with a... [Read More]

Crosswalk Aid

James Li, Pingyuan Liang, and Xuanye Li

Posted on May 10, 2023

The visually impaired encounter a number of difficulties when crossing streets. Without auditory information from parallel traffic, blind pedestrians can easily veer away from crosswalks. Complicated road junctions may also confuse a blind pedestrian’s ability to accurately determine the location of crosswalks. Last but not least, blind pedestrians are also... [Read More]

Bird Feeder

Eric Wu, Harry Rosmann, and Blaine Huey

Posted on May 10, 2023

The idea originated from an animal status monitor and evolved into a bird feeder monitoring system with squirrel detection. We are running two models on the Arduino Nano 33 BLE. The first model detects whether the approaching animal is a bird or squirrel. If a squirrel is detected, an alarm... [Read More]

Athletech

Adeesh Bhargava, Varun Jain, and Rohan Paranjape

Posted on May 10, 2023

The goal of our project is to use cutting-edge computer vision techniques in combination with embedded devices to help people improve their exercise form while also prioritizing their health and safety. While modern computer vision algorithms are able to detect a person’s pose with impressive accuracy, they often lack the... [Read More]