Computers Tell Smart City Lights What to Do
As thunderstorms approach a Minneapolis suburb, LED streetlights tune to a bluer shade of white in order to optimize visibility, a good example of how modern digital lighting can knowingly adjust to different conditions. Equally notable: in this case, the lights do not take their cue from weather sensors onboard the lighting hardware.
Rather, the streetlights in White Bear Lake, MN, know whether they are dealing with snow, rain, fog, or sleet, because IBM's Watson cloud computing system informs them. IBM sends information derived from its own vast global network of weather stations and sensors, the network the computer company picked up in its roughly $2-billion acquisition of much of The Weather Company in early 2016, and which now underpins its Watson Internet of Things push.
The data travels via Internet to a lighting gateway provided by controls company Echelon Corp. The intelligent control gear knows which hues and brightness levels best fit which weather conditions, and sends those instructions to each individual lamp via a powerline Internet connection, dialing up blue-tinged white light for thunderstorms, warmer tones for snowstorms, and other such made-to-order settings.
The small installation in a White Bear Lake park serves to illustrate that there is more than one way to implement the "smart" lighting schemes now emerging in the new era of LEDs, which are by their very nature digital (LEDs, light-emitting diodes, are actually semiconductors).
While venerable lighting companies such as Philips, Osram, and GE would like nothing more than to use the world's ubiquitous lighting infrastructure to house all of the chips and sensors that drive smart lighting and smart city schemes, the Echelon-IBM partnership is a reminder that sometimes it makes sense to tap outside sources of intelligence for smart city detection activities related to weather, traffic, parking, air quality, noise, and other factors.
"In my opinion, there's not one way that's better than the other necessarily," said Sohrab Modi, chief technology officer and senior vice president of engineering for Santa Clara, CA-based Echelon. "The primary essence here is, if you have a greater variance of the type of data, you could use that to make more accurate decisions. That is really the most important thing."
Cost vs. Latency
Modi noted that in instances where rapid delivery of information is important, it can make more sense to rely on locally mounted sensors, which could live in or on lighting hardware. For example, smart lights that switch on as needed in tunnels are effective only if they are aware a vehicle is approaching. "In those cases, you're talking about seconds, because the cars are traveling fast, so you can't wait for the data to come from somewhere else," Modi said. "It's got to be localized data."
Echelon has started trialing smart cameras at an intersection in Spokane, WA, loading the cameras with what it called "cognitive vision" algorithms that analyze traffic flow and adjust the roadway lighting accordingly.
Yet in instances where longer latency times are tolerable, such as with weather-tuned street lights, a remote source of intelligence can not only work effectively, but can save the city money by avoiding the extra capital expenditures required to purchase sensor-laden lights.
"In this case, you don't have to buy equipment and install it and maintain it to do the weather predictions," Modi said. "You don't have to install sensors all over the place in order to do it. So maybe there's a cost benefit here. For much more reduced cost, you're getting access to the data."
IBM's weather data is extensive, as it draws information from The Weather Company's conventional weather stations, as well as from many other sources. For example, it taps tens of millions of smartphones onto which users have downloaded the Weather Channel and Weather Underground apps, a process that gives IBM permission to use those phones to gather information on things like local air pressure. The Weather Company also uses sensors attached to commercial aircraft.
The White Bear Lake installation is the first undertaking of an alliance struck between Echelon and IBM early this year calling for Echelon to tap IBM Watson intelligence to support such smart lighting schemes. Echelon pays IBM a service fee to deliver pertinent intelligence to Echelon hardware; the Echelon system then transmits to lights either via powerline Internet as it does in Minnesota, or via 802.15.4 wireless communication.
Weather represents a likely use case scenario for the remote intelligence approach, but other applications can also benefit.
In Copenhagen, for example, outdoor wireless mesh communications company Silver Spring Networks has demonstrated a system in which it draws information from outside the lighting system to alert streetlights that cyclists are approaching at night.
Using a combination of sensors and the Silver Spring wireless network, traffic signal controllers instructed street lights to brighten up in time to give cyclists maximum visibility after stop lights turned green and allowed them carry on down the road.
An advanced version of the traffic system tie-in could eventually operate at hundreds of intersections around Copenhagen. It is an early example of what Silver Spring commercial director of smart city services Brian McGuigan expects to be more outside feeds into the smart lighting network.
Today, the network—run by France's Citelum with a number of subcontractors, including Silver Spring—controls 20,000 streetlights, instructing them to turn on, off, up or down at preprogrammed times, with its main focus on saving energy. Eventually, the lights will be able to change in response to real-time events and needs. Although onboard lighting sensors could provide the same functionality, the idea is to broaden the sources of information, or to make use of systems already in place, helping budget-challenged cities make better use of existing resources and assets.
"It helps to break down silos," noted McGuigan. "Departments are finding they have no choice but to collaborate, and solve problems in a more coordinated fashion. As technologies become more mature, it's becoming easier to integrate systems. There's a more universal embracing of this, of an ecosystem approach."
In another possibility, for instance, a city's traffic department might send information about road surface temperatures—which it uses to determine whether to dispatch costly salt trucks—onto lighting systems that would retune to cold-weather settings.
"No two cities are looking for the same thing," said McGuigan. "The old model that you might go to one of the major ICT vendors and give them a large check and they give you a smart city, I think it's kind of largely acknowledged that that's not going to be the model of how smart cities evolve, in part because there's not many large checks going around, but also because every city has different requirements. They might have the same list of requirements, but they're all sitting in a different order of prioritization.
"And that comes back to the point that those interfaces and APIs that let data from external systems come in and either inform the lighting system or any other system are becoming increasingly important in what cities are looking to achieve, because they might not have the budget and they want to deploy that which already exists. And in parallel, they want to improve the service they offer to citizens, whether it's the lighting level, public safety, road traffic safety, and other high-levels issues like crime, health, and education."
Lighting companies are turning to external computing power to help them provide smart illumination to cities.
Philips Lighting offers cloud computing to help change facade lighting when a situation calls for color schemes to honor one-off events that operators of Philips-supplied LED architectural lighting on a building or bridge have not pre-programmed.
For example, during the regional round of the NCAA basketball tournament in Memphis, TN, on the weekend of March 24, operators of the Philips-lit Big River Crossing pedestrian bridge that spans the Mississippi River between Memphis and Arkansas tapped a cloud computing system to change colors to the NCAA's blue and white scheme.
Likewise, last Christmas, operators of the Torre Cepsa, a skyscraper in Madrid, used the cloud to change the tower's facade lighting to a seasonal red and green.
While operators can use onsite systems to program regular preset schemes, the cloud gives them flexibility to broaden their possibilities. "For planned events, people have it in their calendars," noted Jim Anderson, business segment leader for Philips Lighting's outdoor architectural lighting group. "But if there's an unplanned event, you can get into the control system and change the content remotely." It's part of a remote management system that Philips calls ActiveSite.
Philips also relies on the cloud in many of its 800 or so smart street-lighting projects around the world, when instructions to change light settings can sometimes travel via Philips CityTouch system from an operators' computer to a cloud-based control system and back to individual luminaires, making the last hop through a wireless connection. In some of those cases, the intelligence remains within the cloud-based lighting system.
Many cities have yet to deploy smart lighting; by some estimates, only about 5% of the world's cities are now lit by LED, and according to Philips, only about 2% of the world's streetlights are intelligently connected. That leaves a lot of room not just to smarten up the lights, but also to broaden the sources of intelligence that feed them.
Two heads, or three, or more, can be better than one.
Mark Halper is a freelance journalist based near Bristol, England. He covers everything from media moguls to subatomic particles.