Project Aria

Project Aria is Meta's research initiative advancing augmented reality (AR) and artificial intelligence (AI) through egocentric data collection and analysis. It provides researchers with Aria Gen 2 glasses, a wearable device equipped with an enhanced sensor suite including multiple SLAM cameras, eye tracking, PPG sensors, contact microphones, and on-device compute capabilities. These glasses enable always-on sensing and human-in-the-loop features, expanding research in contextual AI, robotics, and machine perception.

The program supports a diverse community of over 200 academic and corporate partners, offering access to the Aria Research Kit which includes hardware, software tools, and services. Researchers can collect, stream, and analyze multimodal sensor data to create datasets and build models that improve scene understanding and AI contextual awareness.

Project Aria emphasizes open science by sharing extensive public datasets such as city-scale egocentric data, synthetic environments, and detailed object catalogs. It also provides open-source tools like EgoBlur for privacy-preserving data processing and SceneScript for scene geometry inference. These resources accelerate research and foster collaboration across computer vision and robotics fields.

The Aria Gen 2 glasses improve upon the first generation with longer battery life (6-8 hours), foldable design, programmable applications, and cloud streaming support. The device integrates advanced sensors like four SLAM cameras, RGB and eye-tracking cameras, inertial measurement units, microphones, GNSS, PPG sensors, and contact microphones, enabling rich data capture and on-device processing.

Meta offers a rolling application process for the Aria Research Kit, making this technology accessible to independent researchers and institutions. The program also maintains comprehensive documentation, research papers, and active community engagement through challenges and partnerships.

Overall, Project Aria bridges the gap between research and real-world AR applications by providing state-of-the-art hardware, open datasets, and a collaborative ecosystem. It enables researchers to explore human-centered machine perception and develop innovations that inform future AR and AI products.