Can a Building Know You?

Sensor data combined with analytics trim energy costs and create new experiences.

Lights that flip on and off to timers and banks of lighting activated by area motion sensors are yesterday’s workplace technology. The lights of tomorrow will illuminate your path as you go about the office, adjusting to the precise brightness level you prefer and fine-tuning throughout the day to augment shifting natural light. Tomorrow has already arrived in millions of square feet of commercial office space worldwide.

Rich Green Enlighted

"Sensor information combined with analytics [will] give you denser information about what's going on in the physical world. This will open doors that people didn't even know were doors,'' said Rich Green of Enlighted.

Dense sensor deployments make this level of pinpoint control possible. Such systems may boost worker comfort and productivity, but it’s the potential to slash lighting bills by 50 percent or more that catches the attention of building owners. According to research firm Cleantech Approach, lighting typically accounts for almost 40 percent of electricity costs, and lighting control systems can reduce consumption 35 to 55 percent. Those potential savings will drive expansion of the global market for so-called intelligent lighting controls from $1.5 billion this year to beyond $4.3 billion in 2020, according to Pike Research projections.

As big as the potential light control market is, it’s just the beginning of what sensor technology can bring to the built environment, according to Rich Green, Enlighted senior vice president of products and technology. Green was previously CTO of Nokia and executive vice president of Sun Microsystem’s software division. Enlighted, a venture-funded company that includes Intel among its investors, is one of many startups and established companies vying for a piece of the burgeoning intelligent building systems market. Green was a speaker at the Verge @ Greenbuild conference Tuesday in San Francisco. During an interview at the event, Green discussed what combining sensor data with analytics can make possible and how sensors can make our world a calmer place.

Where are we now in the evolution of sensor technology?

We are now seeing that the cost of sensors is coming down to a price point that they can be densely deployed and that data can be scooped up and used to help you in your daily life. We’ve seen these virtual overlays from Nokia and others where you hold your phone up to a building and you see what’s in the building — if there’s a sale, what the exhibits are [at a museum] and so on. That’s all sensor information combined with analytics to give you denser information about what’s going on in the physical world. This will open doors that people didn’t even know were doors.

You have a platform background, in particular your work with the Java developer network. How does a platform for the built environment differ from those you worked on in the tech industry?

The technology is very similar. It is compressed from a cost and price perspective because this is a new field and people want to see a rapid ROI. The beachhead opportunity is to control and save lots of money on lighting. We save about 50 to 60 percent on your lighting bill. Then there are the opportunities for building applications on top of it. At Sun, we built Java that ran on handsets and then application servers and then computers with databases. [At Enlighted] we build sensors, we build networks, we build analytic systems, we build small servers that go in between. It’s quite similar, but because it’s focused on a set of applications it’s not starting out first as an open platform. It’s starting first as a solution and now there’s demand for the APIs to write applications so it’s blossoming into an open platform. It’s sort of a backwards play on what I’ve done before.

Enlighted Sensors at Flextronics

An Enlighted sensor is wired to each light fixture, sending data and receiving control commands over a wireless network. In this installation at a Flextronics facility in Milpitas, CA, a sensor is located on the ceiling to the right of each light. Courtesy of Enlighted, Inc.

What data do Enlighted sensors collect and who controls it?

We collect temperature, power consumption, occupancy and motion. We collect that and use it provide information back to the [building] owners. It’s stored on our systems, but it’s their data.

What are the potential uses?

Think about any built structure, whether it’s a parking garage, an outdoor area or a warehouse. We measure fine-grain temperature control so we can adjust the HVAC system to a greater level of comfort, to get rid of hot and cold spots. We can build scheduling applications so you can look at your mobile phone and see which conference rooms are occupied in real time. Rather than just scheduling, we can tell you what’s actually empty. We can help with security. We can say there’s been motion in an area at 3 o’clock in the morning. We don’t use video and we don’t take pictures of people, but we can indicate when there has been person-sized motion in a building that may link with a theft or a security violation.

What are the privacy concerns around sensor data?

An important point is that this [Enlighted's technology] is not a video, high-resolution application. We can see a circular object that seems to move like a person. We can’t identify an individual; there’s no facial recognition. There’s no capabilities that go from the general notion of “person-like movement” and map that to an individual. We stay out of that space. This is gross scale pattern observation, not fine-grained security identification.

Is that a technological limit or a policy decision?

It’s both right now. We provide a very dense array of sensors. The pressure on us to make that cost effective means the cost of each deployed sensor must be very low. There’s no practical means by which we can put a high-resolution camera in these. These are very basic PIR [passive infra-red] sensors, motion sensors, thermal sensors and power monitoring sensors. It’s not only a decision we’ve made that we don’t want to get into high resolution; the cost constraints of this dense network preclude us from doing so.

What is the cost of sensor today?

Sensors — deployed — cost roughly $50. The ROI is such that the system pays for itself in approximately 2 years at the average utility cost rate in the United States.

Sensor Panel at Verge Greenbuild

At the Verge @ Greenbuild conference in San Francisco, (from left) John Bradford of Interface, Michael Finegan of Sprint Nextel, Melissa O'Mara of Schneider Electric and Rich Green of Enlighted discussed sensors and the Internet of Things.

How aware should people be that sensors are around them gathering data?

It really has to be invisible. The last thing people need is for their world to be more complex. Having sensors operating passively where they are simplifying your interaction with the world around you makes for a calmer, easier world that is enriched without being made more complex. That has to happen without your participation and awareness, so that you’re being served by their presence.

What’s the relationship between the number of sensors deployed and data center capacity to support the resulting data?

It’s quite significant — certainly non-trivial from a computation perspective because of the I/O requirements. Think about 150MB every 5 minutes for 500,000 square feet, stored for 5 years or 10 years. That’s a lot of I/O capabilities, that’s a lot of storage capabilities, and running analytics across data sets of that size takes very significant computational horsepower. [Enlighted] will be pushing the envelope on both the computation and storage side, because we’re heavily write-centric, whereas a more classical system is often very read-centric. We’re pouring huge amounts of sensor data into a server farm and then we’re running analytics across that on a regular basis. When people want a monthly report of their savings, we’re swiping across these huge data sets to deliver the comparison of before and after. It’s very compute- and storage-intensive.

Looking ahead 5 to 10 years, what will sensors make possible in our daily lives?

Think about going to work and the building knows you, and instead of you adjusting to the physical world, you now control it. The lighting adjusts to the brightness you want; the temperature and humidity are tuned to your needs. As you walk down the hallway, the lighting goes from dim and power-saving to a row of light so that you’re safe and illuminated. When you walk to your car in the garage, there’s illumination where you have to go. Your car knows you’re arriving, it opens up and it knows where you want to go — having that link between you and the physical world — where you’re the dominant force and you’re not adapting yourself to the physical world, but it’s adapting to you.

Can a Building Know You?

Sensor data combined with analytics trim energy costs and create new experiences.

Lights that flip on and off to timers and banks of lighting activated by area motion sensors are yesterday’s workplace technology. The lights of tomorrow will illuminate your path as you go about the office, adjusting to the precise brightness level you prefer and fine-tuning throughout the day to augment shifting natural light. Tomorrow has already arrived in millions of square feet of commercial office space worldwide.

Rich Green Enlighted

"Sensor information combined with analytics [will] give you denser information about what's going on in the physical world. This will open doors that people didn't even know were doors,'' said Rich Green of Enlighted.

Dense sensor deployments make this level of pinpoint control possible. Such systems may boost worker comfort and productivity, but it’s the potential to slash lighting bills by 50 percent or more that catches the attention of building owners. According to research firm Cleantech Approach, lighting typically accounts for almost 40 percent of electricity costs, and lighting control systems can reduce consumption 35 to 55 percent. Those potential savings will drive expansion of the global market for so-called intelligent lighting controls from $1.5 billion this year to beyond $4.3 billion in 2020, according to Pike Research projections.

As big as the potential light control market is, it’s just the beginning of what sensor technology can bring to the built environment, according to Rich Green, Enlighted senior vice president of products and technology. Green was previously CTO of Nokia and executive vice president of Sun Microsystem’s software division. Enlighted, a venture-funded company that includes Intel among its investors, is one of many startups and established companies vying for a piece of the burgeoning intelligent building systems market. Green was a speaker at the Verge @ Greenbuild conference Tuesday in San Francisco. During an interview at the event, Green discussed what combining sensor data with analytics can make possible and how sensors can make our world a calmer place.

Where are we now in the evolution of sensor technology?

We are now seeing that the cost of sensors is coming down to a price point that they can be densely deployed and that data can be scooped up and used to help you in your daily life. We’ve seen these virtual overlays from Nokia and others where you hold your phone up to a building and you see what’s in the building — if there’s a sale, what the exhibits are [at a museum] and so on. That’s all sensor information combined with analytics to give you denser information about what’s going on in the physical world. This will open doors that people didn’t even know were doors.

You have a platform background, in particular your work with the Java developer network. How does a platform for the built environment differ from those you worked on in the tech industry?

The technology is very similar. It is compressed from a cost and price perspective because this is a new field and people want to see a rapid ROI. The beachhead opportunity is to control and save lots of money on lighting. We save about 50 to 60 percent on your lighting bill. Then there are the opportunities for building applications on top of it. At Sun, we built Java that ran on handsets and then application servers and then computers with databases. [At Enlighted] we build sensors, we build networks, we build analytic systems, we build small servers that go in between. It’s quite similar, but because it’s focused on a set of applications it’s not starting out first as an open platform. It’s starting first as a solution and now there’s demand for the APIs to write applications so it’s blossoming into an open platform. It’s sort of a backwards play on what I’ve done before.

Enlighted Sensors at Flextronics

An Enlighted sensor is wired to each light fixture, sending data and receiving control commands over a wireless network. In this installation at a Flextronics facility in Milpitas, CA, a sensor is located on the ceiling to the right of each light. Courtesy of Enlighted, Inc.

What data do Enlighted sensors collect and who controls it?

We collect temperature, power consumption, occupancy and motion. We collect that and use it provide information back to the [building] owners. It’s stored on our systems, but it’s their data.

What are the potential uses?

Think about any built structure, whether it’s a parking garage, an outdoor area or a warehouse. We measure fine-grain temperature control so we can adjust the HVAC system to a greater level of comfort, to get rid of hot and cold spots. We can build scheduling applications so you can look at your mobile phone and see which conference rooms are occupied in real time. Rather than just scheduling, we can tell you what’s actually empty. We can help with security. We can say there’s been motion in an area at 3 o’clock in the morning. We don’t use video and we don’t take pictures of people, but we can indicate when there has been person-sized motion in a building that may link with a theft or a security violation.

What are the privacy concerns around sensor data?

An important point is that this [Enlighted's technology] is not a video, high-resolution application. We can see a circular object that seems to move like a person. We can’t identify an individual; there’s no facial recognition. There’s no capabilities that go from the general notion of “person-like movement” and map that to an individual. We stay out of that space. This is gross scale pattern observation, not fine-grained security identification.

Is that a technological limit or a policy decision?

It’s both right now. We provide a very dense array of sensors. The pressure on us to make that cost effective means the cost of each deployed sensor must be very low. There’s no practical means by which we can put a high-resolution camera in these. These are very basic PIR [passive infra-red] sensors, motion sensors, thermal sensors and power monitoring sensors. It’s not only a decision we’ve made that we don’t want to get into high resolution; the cost constraints of this dense network preclude us from doing so.

What is the cost of sensor today?

Sensors — deployed — cost roughly $50. The ROI is such that the system pays for itself in approximately 2 years at the average utility cost rate in the United States.

Sensor Panel at Verge Greenbuild

At the Verge @ Greenbuild conference in San Francisco, (from left) John Bradford of Interface, Michael Finegan of Sprint Nextel, Melissa O'Mara of Schneider Electric and Rich Green of Enlighted discussed sensors and the Internet of Things.

How aware should people be that sensors are around them gathering data?

It really has to be invisible. The last thing people need is for their world to be more complex. Having sensors operating passively where they are simplifying your interaction with the world around you makes for a calmer, easier world that is enriched without being made more complex. That has to happen without your participation and awareness, so that you’re being served by their presence.

What’s the relationship between the number of sensors deployed and data center capacity to support the resulting data?

It’s quite significant — certainly non-trivial from a computation perspective because of the I/O requirements. Think about 150MB every 5 minutes for 500,000 square feet, stored for 5 years or 10 years. That’s a lot of I/O capabilities, that’s a lot of storage capabilities, and running analytics across data sets of that size takes very significant computational horsepower. [Enlighted] will be pushing the envelope on both the computation and storage side, because we’re heavily write-centric, whereas a more classical system is often very read-centric. We’re pouring huge amounts of sensor data into a server farm and then we’re running analytics across that on a regular basis. When people want a monthly report of their savings, we’re swiping across these huge data sets to deliver the comparison of before and after. It’s very compute- and storage-intensive.

Looking ahead 5 to 10 years, what will sensors make possible in our daily lives?

Think about going to work and the building knows you, and instead of you adjusting to the physical world, you now control it. The lighting adjusts to the brightness you want; the temperature and humidity are tuned to your needs. As you walk down the hallway, the lighting goes from dim and power-saving to a row of light so that you’re safe and illuminated. When you walk to your car in the garage, there’s illumination where you have to go. Your car knows you’re arriving, it opens up and it knows where you want to go — having that link between you and the physical world — where you’re the dominant force and you’re not adapting yourself to the physical world, but it’s adapting to you.