Ongoing Research

Blood Pressure Monitoring Model on Edge Device (TinyML)

Blood pressure monitoring is crucial for managing cardiovascular health and preventing related complications. Our research focuses on developing an innovative approach to blood pressure monitoring using edge devices and TinyML.

We are utilizing neural networks to build our model, employing Photoplethysmogram (PPG) signals as features and Arterial Blood Pressure (ABP) as labels. Our approach involves creating three models to predict Systolic Blood Pressure (SBP), Diastolic Blood Pressure (DBP), and Mean Blood Pressure (MBP) from preprocessed PPG signals.

To optimize for edge deployment, we have converted our models into TensorFlow Lite (TFLite) format, exploring various quantization techniques including Quantization Aware Training. The current phase of our research involves inferencing the model on an Arduino Nano 33 BLE Sense Lite, pushing the boundaries of what's possible with edge AI in healthcare monitoring.

Multimodal-based RAG Application

We are in the early stages of developing a cutting-edge multimodal-based Retrieval-Augmented Generation (RAG) application. This research aims to enhance the capabilities of AI systems in understanding and generating content across multiple modalities, such as text, images, and potentially audio or video.

Currently, our focus is on building a robust RAG pipeline that can effectively process and integrate information from various sources and formats. This research has the potential to significantly improve AI-assisted content creation, information retrieval, and decision-making processes across various domains.