A new dawn brings with it a day promising a new technological breakthrough changing our lives. We live in an age where technology is everywhere around us and some of these technologies have become indespensible to our existence. I strive to be a part of this technological revolution by experimenting with some of these promising technologies so as to explore what is possible. In doing so, I hope acquire knowledge, research through the realm of possibilities, and one day contribute to this technological revolution so as to have a great positive impact on the world around us. I am fortunate to be able to share here a few of my recent projects and research ideas
I am Dhroov Bharatia, a high schooler from Plano Senior High School in Plano, Texas, USA. My current research interests include various topics in Artificial Intelligence and Machine Learning.
VisiAide is an intelligent assistant that helps the visually impaired and blind people to safely navigate along a path, recognize objects in front of them and perceive scenes details. It comprises a custom Raspberry-PI based wearable hardware with a camera to capture user's vicinity in real-time and a LIDAR sensor to aid the system to perceive depth of items in user's vicinity. This custom hardware is controlled by a custom VisiAide app on user's mobile device and enables interacting with the user using a rich speech based interface. VisiAide uses a novel context-based learning approach to guide users with a cloud-based self-learning AI machine learning models.
VisiAide learns to guide new environments and collaborates across users to continuously improve on its own.
This peer-reviewed research was published at the prestigious 2025 International Joint Conference on Neural Networks IJCNN 2025 in Rome, Italy on June 30 - July 5 2025 and can be accessed as Contextual Self-Learning Approach for Continuous Object Detections in Expansive Visual Domains at IEEE Xplore Digital Library.
Additionally, I have launched a public-domain project to use VisiAide for helping the visually-impaired and blind people world-wide. Please checkout this cool project in its entirety here.
Root2Word is my website to help students learn spellings and vocabulary using a novel root-based method that I mastered during my journey to the Scripps National Spelling Bee.
You don't have to memorize spellings - there is a better way. Many words evolved over time and they embed in them a succint history and hold great cultural significance. Word have components called roots that when understood provides you a way to remeber the spelling and its meaning. Explore with me this fabulous linguistic journey here. Many students are already using it and once you explore the magic of word roots, don't forget to checkout live free demonstration of how to learn using word roots.
QuakeWake is a mobile device based intelligent earthquake system that can warn people about an impending and ongoing earthquake and guide them regarding the safety of their residence. Users run a custom mobile app, that reports to cloud servers running AI-based earthquake detection and safety assistance machine-learning models. This enables detecting the earthquake tremors and using smart earthquake progression path it notifies other users in its path before quake arrives.
This peer-reviewed research was published at the prestigious 17th International Conference on Agents and Artificial Intelligence ICAART 2025 in Porto, Portugal on Feb. 23 2025. The research paper is, DOI: 10.5220/0013123200003890, as QuakeWake: A Novel AI-Based Early Earthquake Warning and Post-Quake Building Safety Guidance System and can be obtained through the SCITEPRESS Digital Library.
OmniGuide is an intelligent general-purpose assistant that automatically plans and learns to guide visually impaired and blind people to perform diverse tasks and to do, interact in numerous items in a wide range of environments. Using a novel Tree-of-Activities idea and prompt engineering, it plans using Large Language Model and learns to guide general purpose tasks for visually impaired people with innovative reinforcement learning approach.
Intelliguard is a smart device that detects conditions inside a vehicle with a baby using a thermal imaging sensor and an optical camera. It automatically analyzes and provides reliable, timely and actionable notifications when the baby is left alone or the temperature in the vehicle become hazardous for the baby.
