For Smart Building Infrastructure, is AI the Game-Changer?
By some definitions, a smart or intelligent building is one in which multiple systems gather data about their use, transmit that data to a centralized database, and that data is used as the basis for decisions that optimize those building systems’ use and efficiency. The means of physically connecting building systems, which are well-established, continue to evolve and improve to meet up-to-the-moment requirements of smart buildings. This article will examine some of the cabling technologies that support smart building networks. It also will discuss the potential for artificial intelligence (AI) to be the long-awaited final component that turns collected data into actionable intelligence.
In many smart buildings, sensors are the fundamental technologies that gather the data that ultimately becomes intelligence. While generations of sensors were wireless, the inefficiencies of battery-powered sensors that monitor building systems have been a dogged challenge. Occupants of residential dwellings everywhere are familiar with the hassle of changing the batteries on a smoke/fire alarm with appropriate frequency. In a smart building environment, maintaining a network of sensors embedded within building systems can become an exercise in battery management.
Ethernet connectivity advantages
The advent of Single Pair Ethernet (SPE) is a significant technological step forward for sensor networks. Bringing the stability of Ethernet communications, along with the ability to support lengths of up to 1 kilometer (in the case of 10Base-T1L) and the ability to deliver power to end devices (sensors in this case), SPE is an enabling technology that can dramatically change smart building network administration.
SPE is not new technology. Completed standards that have existed since 2015 are targeted toward short-reach engineered links in automobiles and industrial systems. SPE standards for these environments have evolved over the past decade. While they are not directly related to smart buildings, it’s easy to correlate the benefits of Ethernet in an automobile to benefits in a smart building. With modern vehicles equipped with numerous smart technologies including driving aids, systems that previously communicated internally using their own language now need to communicate with other systems inside the vehicle, to make decisions as quickly and efficiently as possible—like when the system detects your car is too close to something in front of you must communicate with the braking system.
The primary use for SPE in smart building automation is for connecting low-speed operational technology (OT) devices including sensors, in building control systems. Just like in a car, the use of Ethernet as a common language allows for communication and sharing of actionable data across these systems to more easily make decisions that improve efficiency and occupant experience. Rather than the line-of-sight sensor in a car talking to the brakes, it’s the occupancy sensor in the conference room talking to the HVAC system.
Emerging role of AI
But who, or what, within a smart building is making those decisions? In a recent interview with Cabling Installation & Maintenance, CommScope’s vice president of building and campus research-and-development Luc Adriaenssens pointed to artificial intelligence (AI).
“Sensors deliver a lot of data, but data can provide insight only if someone or something does the processing to determine what it all means,” Adriaenssens pointed out. Commonly in a smart building, people analyze data and develop programs that optimize systems—for example, the use of a building’s HVAC system according to current or anticipated weather conditions. Expending human capital on this responsibility is costly and subject to diversion of that asset to other organizational priorities.
“If you present the same data to AI, it can determine how to optimize building systems automatically—turning data into a tangible benefit. That time is coming, and it will be huge for smart buildings.”
The time may be coming soon, Adriaenssens says, because AI is already being put to practical use in some environments today. He points to automatic data mining as a concrete example. “Information in data streams may appear uncorrelated until you apply AI to it,” he explains. Acknowledging that correlation does not always equal causality, Adriaenssens says some AI-generated policies will need to be approved. He estimates approximately 5% oversight of AI-generated smart building policies is likely to be the norm.
“With limited control—management oversight of AI-generated conclusions—we will have a very powerful tool. The problem with many smart building concepts remains that they’re still complex. Turning raw data into actionable intelligence, if it can be automated, is a game-changer.”
Furthermore, he points out, a highly capable AI engine will be able to streamline the communication among data from different building systems—eliminating the requirement to have human capital manipulate application programming interfaces (APIs) to do so. “The amount of unstructured data generated within buildings each year is massive. Often that data is too expensive to store, and is discarded.”
If, on the other hand, a building owner gives an AI engine access to data generated from an access-control system and from a surveillance system, “it can automate the so-called ‘glue’ that makes the connections between these separate turnkey systems,” he emphasizes. Doing so “can ensure your infrastructure is not going to be limiting the possibilities” of capitalizing on the data gathered.
Edge devices and power needs
Another technological evolution with smart buildings is the practicality of data gathering and analysis happening at the edge devices, as opposed to those edge devices only gathering data and transmitting it to a location where it will be analyzed and acted upon. Reiterating that data storage can be an expensive proposition for a building owner, CommScope’s Adriaenssens observes, “When there is intelligence at the endpoint, you may not need to transmit all the data a device has gathered. The intelligent device can indicate there is nothing to report in the environment—no change. The intelligent device only reports changes to the norm. This capability can take a building owner down lots of avenues.”
Some of these intelligent edge devices, like surveillance cameras, are network-ready, manufactured with RJ45-based network interface cards. The more advanced and more sophisticated these devices’ electronics are, the more power they require. For many of these latest-generation (and next-generation) network devices, Category 6A is the preferred cabling media. Category 6A’s information-carrying capacity easily handles the data streams coming from the devices, and Cat 6A’s ability to deliver up to 90 watts of power via Ethernet's Class 8 PoE make it an enabling technology for such sophisticated network nodes.
Sometimes, as in the case of a smart campus with a sprawling network, Power over Ethernet may not be a perfect fit for the network’s power needs—or may require an underlying supporting technology of its own. More and more, networks with diverse powering and data-transmission needs are turning to fault managed power (FMP). For smart buildings and campuses, FMP achieves the efficiency that is the hallmark of these properties.
VoltServer, the inventor of Digital Electricity, describes how the technology can be used in smart buildings and campuses. “Next-generation education, healthcare, business, and industrial campuses are focused on flexibility and connectivity,” the company explains. “For convenience and continuous operation, these environments must be functional and purposeful. Instead of extending traditional grid infrastructure across campuses, remote power helps modern sites create safe and versatile spaces. VoltServer’s Fault Managed Power System supports fast deployment on campuses across North America. Our patented Digital Electricity solutions enable power and fiber deployment in a single pathway to streamline operations and improve connectivity, even in difficult installation environments.”
CommScope offers the Constellation connectivity platform, which includes FMP and data-communications technologies. When announcing a recent evolution of Constellation that accommodates geographies across the globe, CommScope noted, “Compared to traditional structured cabling networks, the Constellation platform has a radically simplified star architecture that is modular and technology-agnostic. It supports AC and DC power applications and a variety of connectivity standards for IT/OT networks. Eliminating the need for a telecom room or wiring closet on every floor, it enables centralization of uninterruptible power supply equipment, and Constellation Points can be deployed in ceilings, walls, or equipment racks—using fewer components while supporting both converged and segmented networks.
“The Constellation platform significantly extends service distances from equipment rooms and can deliver 10G and above speeds, and 1-kilowatt FMP to myriad connected devices in today’s increasingly dense urban environments.”
Adriaenssens added, “The Constellation platform enables our customers to support much denser urban centers and a proliferation of connected devices in their buildings. By offering a modular, technology-agnostic solution, we’re giving our global customers more choices for building connectivity and power to the network edge—and helping them to do it in a way that’s cost-effective, scalable, and sustainable.”
In terms of smart buildings and campuses, he refers to FMP as a relatively new tool for designers and planners to work with in these environments. The upper limits of direct current (DC) voltages that can be delivered safely through FMP have not yet been determined. The industry could find that it has the ability to deliver even more power via FMP.
This potential goes hand-in-hand with other objectives of smart buildings and campuses—namely, to proactively rather than reactively improve an environment, including working and living spaces.
About the Author
Patrick McLaughlin
Chief Editor
Patrick McLaughlin, chief editor of Cabling Installation & Maintenance, has covered the cabling industry for more than 20 years. He has authored hundreds of articles on technical and business topics related to the specification, design, installation, and management of information communications technology systems. McLaughlin has presented at live in-person and online events, and he has spearheaded cablinginstall.com's webcast seminar programs for 15 years.