Autobot: Data-driven metadata tagging of building automation systems

As Building Automation Systems (BAS) are becoming a standard in commercial buildings, and additional 3rd party applications can help buildings owners gain insights from their BAS, structured metadata management becomes the key to success. However, as converting traditional sensors naming convention to structured tagging systems is an expensive and time-consuming process, this project aims at automating the process. By leveraging modern Machine Learning techniques combined with Rule-based systems, Autobot is developed to automate the process. This project builds on previous research at Brainbox AI (Mishra, et al., 2020) to apply the software to additional building types and improve the process’s overall performance.

Faculty Supervisor:

Adam Rysanek

Student:

Claude Demers-Belanger

Partner:

BrainBox AI

Discipline:

Engineering - mechanical

Sector:

Professional, scientific and technical services

University:

University of British Columbia