| The latest elevator technologies give tall buildings
an IQ boost
Elevator manufacturers are applying guidance systems run by Artificial intelligence and other technologies to get people around more efficiently in increasingly taller buildings.
By Alan Joch
If you yearn for elevators that can get you to your high-rise destination faster, the future looks bright. In the past, seasoned elevator riders employed two common tactics: guessing which car would open up next and positioning themselves accordingly, or the futile act of endlessly punching the button to induce the elevator to move more quickly. In reality, neither of these techniques alleviated the rider bottlenecks that build up in high-rise lobbies during peak traffic times.
Fortunately, the latest elevators have an answer to elevator gridlock. New technologies such as artificial intelligence (AI), genetic algorithms, and fuzzy logic (a technology employed by AI enabling computers to make calculations based on approximations or trends rather than absolute data), now control traffic management systems in leading-edge elevators and may finally accomplish what inveterate button pounding never could.
A number of major elevator vendors offer their own versions of high-tech guidance systems that use cyber-smarts to load elevator cars by grouping people going to similar areas in a tall building, rather than letting them randomly select cars themselves. Vendors concede that lobby wait times may be longer as these systems shuttle people to designated cars to optimize arrivals and departures. But the total time it takes to reach upper floors decreases. “In heavily populated buildings or buildings with multiple elevators, it groups people together so traffic is more organized and an elevator makes fewer stops,” says Sula Moudakis, director of high rise business in North America for Schindler Elevator.
Even more important, greater traffic efficiency may have farther reaching benefits, including decreasing the number of elevators needed to move a tall building’s traffic, thereby saving costs and opening up additional occupancy space.
With the new systems, riders log in their destination at central lobby kiosks and the guidance programs use this information to determine which cars and how many of them will travel to specific floors. An LCD display directs riders to the proper elevator. Once inside the cars, passengers don’t have to re-enter their floor destination.
Guidance systems for optimizing car loads aren’t new; rudimentary versions have been around for over a decade. What’s new today is the higher levels of intelligence, and, theoretically, efficiency with which the guidance systems operate. AI-based traffic control systems bring additional smarts to the job. Encoded in their circuitry is the ability to constantly evaluate traffic volumes and patterns as they change during different times of the day and to measure how long it’s taking the system to move people to each floor.
By constantly monitoring traffic flow, the systems become savvier about allocating cars and grouping passengers. “The system is able to consider historical information to learn the traffic variances in a building,” says Joe Rennekamp, vice president of engineering for Fujitec America. “The system knows at such and such a time there will be a peak in traffic, and at another time there will be a downturn.” Over weeks and months, an AI-based system can draw on these patterns to learn the characteristics of the particular building to become increasingly efficient at moving people around.
The lobby-based elevator touch screens can perform other duties as well. Passengers can view onscreen directories to select a tenant without knowing the floor number. Authorized building personnel can punch in special codes to designate cars for different types of service, including temporarily commandeering a car for freight service, for example. Riders with physical disabilities can tell the system to hold a car’s doors open longer than normal or load the car below its usual capacity to accommodate the extra space taken up by a wheelchair. Some systems also cater to special passengers. By punching in a special code, a senior corporate executive can take a solo express ride to the top.
In the future, special codes combined with identity management systems will be used to restrict access to upper floors to increase high-rise security, Moudakis adds.
Fujitec America, Kone, and Schindler say they’ve all developed their own guidance systems. The AI applications run on business-class server computers. The systems work best with taller buildings, those with at least 10 stories and three or more elevators. AI-based traffic control adds about 15 percent to the total elevator system cost, says Moudakis.
Less Is More
But others say the extra capacity is best for temporary capacity spikes rather than for day-to-day service. “We did some studies and decided that the real benefits are in managing high traffic,” says Frank Dugan, director of special projects for Fujitec America. Moudakis concurs. “If you take an elevator out of service for maintenance, you can still service the building with the remaining elevators,” she says.
The guidance systems, which can be retrofitted onto existing elevators, are also impacting modernizations, especially in buildings whose populations and traffic levels have reached higher volumes than what the original designs anticipated. “Rather than being stuck with elevator systems designed 50 years ago, you can take the same elevators and increase their capacity [with the guidance systems],” Moudakis believes.
Elevator companies say rider optimization systems will blossom in the U.S. as builders of new high rises, including the New York Times corporate headquarters in Manhattan and the 48-floor Hyatt Center in Chicago, adopt these technologies. But the U.S. market is playing catch-up with Europe and Asia, where AI-guided elevators are more common in tall buildings. “Originally, we felt it would work best for single tenant buildings and we focused on that market,” Moudakis says. “Then, as we used it in new installations and modernizations, we figured out [the tenant mix] doesn’t matter. Passengers get used to the system very quickly.”
Alan Joch is a freelance business and technology writer based in New England.