Machine learning at work: creating smarter skyscrapers

Several recent stories have served to underscore the advantages of putting the Internet of Things (IoT) to work inside large buildings. When heating and cooling systems, lighting, elevators and security are connected and controlled by a centralized management solution, the potential returns on investment — in time and money saved, in more efficient employees, and in environmental impacts — are great. Add to that the ability to predict usage trends and equipment failures using data and machine learning, and suddenly, wasted resources drop exponentially. It seems likely that before long, machine learning will be table stakes for successful skyscrapers.

That’s not lost on Takenaka Corporation, one of Japan’s largest architecture, engineering and construction firms. To help customers address aging infrastructure and strict governmental energy-use requirements, the company created an IoT “smart” energy management solution based on Microsoft Azure. Watch a video about the solution here:

The system collects and analyzes large volumes of data from equipment such as air conditioners and lighting, and also on the flow of building occupants. The data is sent via the Internet to the cloud, where it is analyzed using Azure Machine Learning to create predictions and trigger actions, turning the heat on, for example, when rooms are in use; choosing the right time to turn down the thermostat to prevent waste. Takenaka expects the services will reduce building management costs – both in utilities and in staff – for its customers.

The solution is being field tested at Takenaka’s main office in Tokyo; it’s expected to be introduced in about 20 buildings in 2015.