Past talks‎ > ‎

Balancing energy and comfort through machine learning in commercial buildings

Original event:

Stephen Dawson-Haggerty, Co-Founder and CTO of Building Robotics

In looking at commercial building energy consumption, many of the known opportunities for positive behavior change have been found through turning off lights, turning off plug loads at night, and other similar easy actions for workers to take in their daily routine. But often the largest energy usage in buildings is HVAC- an energy use that is inextricably tied to people and their preferences and behavior. How can we create a system that encourages efficiency and good behavior in HVAC use? And what does that behavior look like?

At Building Robotics, we’ve developed software called Comfy that addresses these issues head on. Modern internet-based, mobile-ready technology allows us to provide much more granular and dynamic services to people, including comfort in offices. With Comfy, we do this in 2 ways. First, with machine learning, which allows us to provide dynamic environmental conditions, without taking the time of the facility manager. Equally critical is great user experience, which allows us to incorporate the intelligence and needs of real people, however unpredictable, while being simple, delightful, and sticky. In this talk, we'll discuss the results of our early pilots, showing how people are using Comfy, and how that in turn impacts the building around them.

At its core, Comfy presents a vision and framework for a system that needs both real people and machines to work together to produce great environments for people that reduce unnecessary energy use. We hope to inspire others to think of ways to use people-centric, machine-learning-based technology to solve environmental problems while helping people in their daily lives.

Stephen Dawson-Haggerty is co-founder and CTO for Building Robotics, a software startup focused on improving the way buildings operate by improving the interaction between occupants, facilities managers, owners, and data. He received a Ph.D. in Computer Science from Berkeley, where he developed a system for data acquisition and control in commercial buildings called sMAP. Steve has extensive experience with embedded sensing and networking and large-scale deployments of wireless sensors.