Improving Bus Arrival Estimates
University of South Florida & Center for Urban Transportation Research (08/2017 – 10/2023)
- This project focuses on creating a machine learning experimentation platform used to compare state-of-the-art algorithms across transit agencies.
- Our approach has several strengths over the standard machine learning pipeline such as unattended real-time data collection over several cities, raw-data archiving and high-speed access to unstructured data structures. These improvements allow us to focus our attention on improving estimates by refining the feature extraction process in addition to parameter tuning for algorithms.
Reducing Patient Readmissions in Heart Failure Patients
University of South Florida (College of Engineering & College of Nursing) (05/2017 – current)
- This project aims to reduce patient readmissions and health risks after a cardiovascular event. This is done through a mobile app in which doctors, nurses and family members can monitor a patients health using wearables or assessments and providing interventions if needed. Patients can also utilize specialized health failure modules as self-care tools to improve their recovery and habits
- Our research focuses on identifying and applying novel intervention strategies in order to evaluate their efficacy.
Cluster Perception and Identification in Scatterplots
University of South Florida (Location-Aware Information Systems Lab & Graphics and Visualization Lab) (01/2020 – 12/2022)
- This work focused on understanding the visual perception of clusters by observers across several sampling methods, visual encoding methods and datasets. The results where then generalized and used to determine the optimal encoding parameters for a given dataset.
Pulmonary Disease Audio Classification
University of South Florida (Location-Aware Information Systems Lab) (01/2019 – 08/2019)
- Applied research directed towards capturing and classifying pulmonary sounds to diagnose chronic obstructive pulmonary disease (COPD) and Heart Failure (HF) using mobile phones.
Tennis Stroke Classification
University of South Florida (Location-Aware Information Systems Lab) (05/2016 – 05/2017)
- The project was focused on performing activity recognition in the sport of tennis using Inertial Measurement Units (IMU). Secondary objectives included determining the optimal position to wear IMUs while playing the sport and creating graphical labeling software for accelerometer and gyroscope data. We achieved a 95% classification accuracy between 7 types of segmented tennis strokes.
Autonomous Robot Navigation
University of Puerto Rico (Computer Science Department & Electrical Engineering Department) (08/2015 – 12/2015)
- Early stages of a project aiming to develop a self navigating robot that would perform collision avoidance using a stereoscopic camera. My relation to the project extended to generating depth maps and estimating the distance of regions in the robot’s field of view.
Floor Detection on Unstructured Environments
University of South Florida (Location-Aware Information Systems Lab) (05/2015 – 07/2015)
- This project was meant to serve as a module of a larger fall prevention system for visually impaired and elderly people. It focused on determining the ground in images retrieved from the camera of a mobile phone using computer vision techniques in real-time.
Lightning Detection and Monitoring System (PRLDN)
University of Puerto Rico & NASA Puerto Rico Space Grant Consortium (12/2014 – 12/2015)
- The purpose of this project was to develop a lightning detection system using an array of sensors placed around the island of Puerto Rico, which is able to detect low frequency radio signals generated by lightning. My contribution to the project extended to expanding functionality of lightning monitoring modules, improving cross-platform support and Introducing procedures for analysis of archived lighting data.