The pediatric intensive care unit at centre hospitalier universitaire Sainte-Justine wants to be able to evaluate respiratory functions of a patient without using the current techniques which requires the patient to be mechanically ventilated. Currently, those respiratory functions can only be assessed by a specialist doctor with experience. The goal of this project is to create the first system which would help in the gathering and assessment of those parameters for a patient. To construct a real-time monitoring system through video data we needed to define what was the necessary data and the sensors available. The data in question was infrared data, point cloud data and RGB-D data. The infrared data was gathered with a Teledyne Flir Lepton 3.5; the point cloud data and the RGB-D data was gathered with an Azure Kinect DK. Both of these devices were connected to a Jetson Xavier NX board. Before this project, the data was gathered through three different applications. This resulted in the data being unsynchronized making it usable individually but unusable if you need the different data types to be synchronized for any kind of model training through machine learning or deep learning. The problems solved by this project were the implementation and development of the first tier of the data science pyramid (collecting layer), the synchronization of the data, the development of a web application and making it secure and remotely accessible. To do so we created a system of systems called Olympus with the following subsystems:

• Demeter (Data collecting software and API)
• Aglaea (Web application)
• Zeus (API Gateway, Authentication API and Authorization API)

We used various different technologies to create this system of systems such as: Nginx, FastAPI, Python3, SQLAlchemy, Alembic, PostgreSQL, ReactJS, NextJS, TypeScript and ThreeJS. The main focuses were efficiency, usability and security.