Deep learning and smart sensors are both gaining wide interest in academia and commercially, but they both still lack research and a commercial breakthrough. DEBARE project researches novel gesture recognition using distributed multimodal deep learning and smart sensors.

DEBARE develops a groundbreaking intelligent sensing platform for intuitive interaction with physical objects. Key objectives are:

• Developing smart sensor-based data acquisition system for activity recognition
• Researching and developing multimodal deep learning for activity recognition
• Applying developed gesture recognition algorithm to an AR use case
• Assessing the feasibility and implementation of multimodal deep learning on smart sensors

Focus for the first 12 months will be in augmented reality applications for industrial maintenance and assembly. Our goal is to a build a scalable platform with initial industrial validation to enable new use cases and further commercialisation as next steps of the project.