Development and implementation of an Edge AI IoT device with domain specific architecture to autonomously monitor children around pools

Costs are continually declining for internet connected devices, known as the internet-of-things (IoT), offering solutions across various consumer, commercial and industrial applications. Concurrently, machine learning (ML) models are continually expanding their deployment platforms, from massive central servers down to microcomputers. Innovative IoT devices will run the ML models without support from central data centers or other computers. Running the models covering any aspect of artificial intelligence (AI) within only the device, which is called “Edge AI”, will play a role in disrupting each and every data acquisition process or product globally.
By combining novel hardware, ML model compression and federated learning as a customized solution, Spectergy’s new device will be able to dramatically improve the scope and speed of running compressed ML models. This specific research project will focus on the Edge AI IoT device application domain of renewable energy generation and will result in an increased ability to capture, process and leverage onsite sensor input data to increase the efficiency of energy generation.

Faculty Supervisor:

Steven Muegge

Student:

Jeff Cole

Partner:

Spectergy

Discipline:

Other

Sector:

Manufacturing

University:

Carleton University