CE Capstone - Projects
2024-2025
LunarVitals
Members
Allan Garcia
Lucas Liang
Peak Manopchantaroj
Victor Garcia
Zeiler Randall-Reed
Description
The EVA Vital Monitoring and Predictive Analysis System is designed to support astronauts during extravehicular activities (EVAs) by continuously monitoring their physiological data and predicting task performance limits. As spacewalks require efficient management of life-support resources like oxygen, our tracker will have sensors to track key vitals such as heart rate, body temperature, oxygen levels, and motion. The data collected will enable real-time analysis of the astronaut’s physical exertion, aiding astronauts and mission control in managing task workloads. Through predictive analytics, the system will anticipate potential exhaustion or depletion of resources, allowing astronauts to adjust their physical activities accordingly, ensuring mission success.
Resources
Sponsors
SproutSense
Members
Max Cohn
Bharat Kathi
Julia Chan
Kriteen Shrestha
Pablo Sandoval Rivas
Description
SproutSense is an innovative agricultural monitoring system that employs a network of sensors to collect critical environmental data, including soil moisture, temperature, humidity that influence crop growth. This data is transmitted wirelessly, enabling them to make informed decisions on irrigation, fertilization, and other farm management practices. By providing insights on field conditions, SproutSense aims to enhance productivity, improve resource efficiency, and promote sustainable farming practices. The system is scalable and customizable, suitable for farms of various sizes and crop types, helping to address the growing need for agriculture in modern farming. We also plan to use machine learning algorithms to further analyze our data to find trends to further support farmers in their everyday work.
Resources
Sponsors
MANA-T
Members
Alex Xu
Jason Rodrigues
Kevin Fan
Neel Takale
Sreeganesh Siva
Description
Military reconnaissance demands an unmanned vessel capable of autonomously navigating waterways, gathering intelligence, and executing reconnaissance missions without human intervention. The USV is designed to be a lightweight, maneuverable vessel that enables the remote exploration and mapping of coastal lines without relying on classical wireless technologies like cellular. The vessel’s capabilities for coastal monitoring and terrain surveillance acts as a viable alternative to existing airborne solutions for military, environmental science, and civil use purposes for terrain reconnaissance. Our team is focused on the enhancement of the previous system: improving self-sufficiency in remote regions, implementing autonomous operating capabilities, and raising the surveillance fidelity. The major goal of this year is to fully implement SLAM (simultaneous location and mapping) functionality to an autonomously operational level.
Resources
Sponsors
Eco-Coustic
Members
Lizzie Crooks
Derek Dela Cruz
Max Stewart
Morgan Parslow
Tien Nguyen
Description
Our capstone team is working with a local company, Sightir, to protect endangered species from the environmental harm posed by wind turbines. Using an array of microphones in the audible and ultrasonic range, our system takes sound input and feeds it into a Raspberry Pi, which is running a deep neural network (BirdNET), to identify the species. The sound waves are then used to triangulate the position and heading of the bird, which are overlaid with thermal cameras to track targets in real time. The collected data is used to monitor the migration and flight patterns of local birds with respect to wind turbine fields. Our goal is to provide useful data which will be used to reduce collisions between active wind turbines and avians in order to create more eco-friendly green energy systems.
Resources
Sponsors
GauchoSat
Members
Nik Belle
Brady Gin
Dakota Barnes
Kpazawala Windross
Description
GauchoSat is a CubeSat project aimed at advancing space solar cell technology through in-orbit testing and data acquisition. The satellite, a 1U CubeSat (10x10x10 cm), will be equipped with state-of-the-art solar cells provided by industry partners and will be launched into Low Earth Orbit (LEO). The mission focuses on measuring key parameters such as the IV curves, cell temperatures, and sun angles of the experimental solar cells mounted on the CubeSat's surfaces. Using an Aerospace Measurement Unit (AMU) and a low-bandwidth communication system, GauchoSat will transmit this data back to Earth for analysis. With sun-pointing capabilities and sensor integration, the project contributes to the future development of more efficient solar technologies in space exploration.
Resources
Sponsors
Gibbous
Members
Tony Tran
Diego Mateos
Oszkar Augusztiny
Simon Mann
Description
The Moonlight Simulator is designed to study the behavioral patterns of ostracods in response to moonlight conditions. Ostracods, small bioluminescent crustaceans, display behaviors that are influenced by lunar phases, making precise emulation of the moon and control over its parameters that are essential to study their behaviors and responses. Our device provides dynamic control over key variables, including brightness levels, lunar phases, moon cycle durations, and the position of the moon on any specific day. Additionally, the system implements a movable track that adjusts to the LED’s position in real-time to simulate the moon’s natural movement across the sky. Allowing the user to manipulate these conditions will allow for experiments to be done with greater position and versatility, leading to a better and more comprehensive understanding of ostracods behavior.
Resources
Sponsors
InfraTrack
Members
Andy Gale
Dylan Jones
Ishaan Joshi
Kelly Flippo
Marin Lopez
Description
This project aims to develop advanced firmware and software upgrades for an existing Senseeker infrared camera system. Our main focus will be towards a high-speed multi-window mode, which allows us to read certain subsets of the image array much more efficiently than the rest of the image. Window choice will be communicated in real time to the sensor, so efficiency and time complexity of our programs will be highly emphasized. A combination of image processing algorithms, real-time data streaming, and display software will be required to analyze the input from the sensor and display it to a human observer in a timely manner. If time permits, advancements can also be made towards other modes of operation or enhancing the high-speed multi-window mode even further.
Resources
Sponsors
Intellitory
Members
Nico Shinozaki
Barry Wang
Josephine Nghiem
Justin Zhang
Tamer Tamer
Description
The Smart Shelving System is an innovative inventory management solution designed to streamline and enhance the accuracy of tracking and cataloging items in real-time. By combining UHF RFID technology and weight sensors, the system provides precise monitoring of inventory levels, reducing manual labor and minimizing errors in stock counting. This dual-sensor approach allows for automated item identification and weight-based verification, ensuring that inventory data is always up-to-date. The system is ideal for applications in retail, warehouses, and industrial settings, offering improved efficiency, cost savings, and better resource allocation. Through seamless integration, the Smart Shelving System aims to revolutionize traditional inventory management practices.
Resources
Sponsors
WaffleChat
Members
Logan Unger
Jason Nguyen
Josh Gianelli
Yang Liu
Description
We will present a multi-model, ChatGPT style AI Agent capable of comprehending and displaying data from the WM-811K dataset of wafermaps (labeled by lot and index for each lot). The user will type or speak its request (these are defined in the External Behavior Specification) to the agent. The frontend will strengthen the user input to make it work more favorably with GPT, which will then call the appropriate Python script in the backend to display the answer. The set of instructions to which the model can respond will be the “Spec,” which could be different for different applications. This is designed to help process engineers make better use of the large semiconductor datasets and eliminate the grunt work of searching for wafers matching a certain pattern.