The quantified building: IoT for FM

by Erik Jaspers & Eric Teicholz — Originally published in the March/April 2016 issue of FMJ—Imagine if every asset of your building — fans, doors, furniture, coffee makers, windows — contained a tiny embedded sensor that gathered simple data to determine the actual behavior of those assets. Imagine having systems in place that could capture this data and use it in real time to adjust behavior and signal human intervention as needed. Imagine being able to analyze the accumulation of this data over time to assess structural improvements and optimize operations.

This is the promise of the Internet of Things (IoT) for FM. So, what is actually happening in IoT-related technology and how can you plan for its adoption?

Trends in technology

IoT, smart buildings, analytics, machine learning … an array of terms is used to describe technology trends that impact, among others, the professions of real estate and facility management. These technologies, while initially disruptive, are positioned to revolutionize our business.

One current problem is that this terminology is often poorly defined, ambiguous or confusing. However, it is important to first understand what these terms mean in order to leverage these trends.

INTERNET OF THINGS

IoT is probably one of the most discussed phenomena these days. But what is it?

Gartner, Inc.’s1 comprehensive definition applies to both real estate and FM: “The Internet of Things is the network of dedicated physical objects (things) that contain embedded technology to sense or interact with their internal state or external environment. The IoT comprises an ecosystem that includes things, communications, applications and data analysis.”2

Within FM, this translates into the introduction of meters, sensors, building systems and devices that measure actual behavior and can communicate and interact with other systems, thus creating the ability to use data to adjust behaviors.

BIG DATA

As we increasingly connect more systems, devices and objects, the amount of data grows exponentially — hence the term big data. This term addresses not only the volume of information collected but also its variation (volatility) and speed. Technology allows us to capture, store and intelligently analyze these vast quantities of data in a variety of ways and at a low cost.

BUSINESS INTELLIGENCE AND ANALYTICS

The real value in acquiring data is in using it to your advantage. One area of use is reporting and analysis.

The term analytics has recently entered our vocabulary. We have traditionally thought in terms such as reporting (e.g., for operational data overviews) and business intelligence (e.g., for publishing KPIs and business dashboards). So how does the term analytics differ?

The most common characteristic within the variety of definitions for analytics lies in the principle of discovery. Analytic systems tend to provide support in discovering patterns that are not obvious or easy to see.

SMART

The adjective smart is frequently used (e.g., smart machines, smart buildings or smart meters) but can vary significantly in meaning depending on context. Smart meters, for instance, are merely connected devices that have the ability to transfer their readings. However, the word smart connotes a fundamentally different meaning when used in the context of machines. Gartner3 defines smart simply and sensibly by linking it to the property of learning, which in turn implies artificial intelligence (machine learning, deep learning).4

Machine learning technologies will provide significant value to FMs, allowing identification of correlations, rapid analysis and prescription of appropriate responses (also referred to as predictive or prescriptive analytics). This could apply in situations such as emergency response, energy management, asset/maintenance intervention, security system behavior, etc.

When we integrate these technology trends, the resultant paradigm allows us to create new applications for real estate and FM. IoT allows for connectivity and interactions, big data allows for data capture and storage, and smart technologies use this data, learn from it and act on the results of the learning. Putting this in the perspective of facilities, it is plausible that we are heading toward truly smart buildings.

The dominant underlying trend: Quantification of people, cities and buildings

We are installing sensors into our infrastructures at ever higher rates. Ubiquitous computing is occurring: these sensors and devices assess conditions and are consequently used as inputs for adapting system behavior.

All of these sensors are not yet integrated, but this will happen over time. Buildings, for example, currently use sensors to optimize energy utilization, such as adjusting lights based on room occupancy. These and hundreds of other functions communicate through the cloud with data centers that can be located anywhere in the world. Based on the information collected, the servers can return commands to control the sensors in order to operate a facility more effectively.

Cities are doing the same with their assets. Examples of urban applications include preventive maintenance data about urban transportation infrastructure, finding unoccupied parking locations, optimizing traffic flow on road networks based on current real-time density and accident data, locating the closest support needed for emergencies and other security applications, to name a few.

Ubiquitous computing is increasingly associated with people as well. Personal health is a major driver behind us adopting health apps. Today, wearable devices can track heart rate, pulse and a host of other fitness statistics. In January of 2016 at the Las Vegas Consumer Electronics show, there were dozens of new wearable computer apps related to telling users when and what to eat, when to take a break, timing and quantities of medication, and so forth.5

WHAT IS REALLY HAPPENING?

Applications associated with people, cities and buildings have a common denominator: they all start with gathering data. This is then used to describe actual behavior in simple terms. In fact, the process starts with quantification.

Quantification is necessary for understanding and describing the behaviors of the system being studied. Without quantification there cannot be science. Without data, machine learning is not possible. So, in order to acquire smart capabilities, we need to first capture data. Investing in sensor data capturing initially and then waiting for smart technologies is not effective.

There are currently apps that introduce sensor technologies and use data to immediately return value, even without the use of machine learning. For example:

• Data gathered from occupation sensors can help identify and inform users of vacant locations and allow FMs to analyze occupation patterns.
• Connected meters can register energy consumption and analyze patterns to avoid waste and inefficiency.

• Sensors installed on critical assets like escalator motors can monitor their behavior and signal maintenance needs in a timely manner, reducing safety risks and avoiding failure situations.

We can achieve real innovation in FM operations when we integrate various forms of quantifiable systems. Think of cleaning or energy management steered by actual occupancy, or building management systems operated based on climatic and air conditions (temperature, humidity, carbon dioxide).

This type of behavior is often denoted as smart; however it actually involves no element of machine learning. Every situation that is detected is processed in ways we have known about for years, leading to predictable responses. Google calls this type of data processing “if this then that.” In other words, the system always will respond in predictable ways: input defines output.6

So, although helpful and cost effective to FMs, these types of systems (and thus buildings) are not truly smart yet because no element of learning is in play. To distinguish this first and necessary step from the ultimate smart buildings, we should speak instead about quantified buildings. These are buildings that have the capability to describe their behavior and adapt their responses in effective ways.

When IoT is applied to quantified buildings — described as the Building Internet of Things — the low-hanging fruit currently relates mostly to building automation systems (BAS). Research and development is connecting and integrating data from all BAS systems, then analyzing and fine tuning systems in seconds without human intervention. This is possible because of open communications Internet protocol which enables such BAS connectivity. Using big data analytics, BAS systems will increasingly improve using smart learning practices. This type of integration is still in its infancy, as problems, such as security issues, need to be resolved.

It is just a matter of time before quantifiable cities, buildings and people start to share and integrate their data. FMs can now look beyond the walls of their own facilities to see vast amounts of available data and describe the behavior of the built environment.

In terms of IoT applications and solutions, technology vendors like Google, IBM, Microsoft and Amazon are developing platforms that allow for vast amounts of devices and systems to be connected, providing the scalability and allowing for the device diversity that are prerequisites for the data capturing needed. Cloud provisioning plays a pivotal role here, enabling systems to interact over the Internet. Once such functionality exists, FM vendors (e.g., CAFM, CMMS, BAS) will adopt these smart systems for FM applications.

Getting smart

Although IoT is beginning to take root in facilities, smart (learning) technology deployment in commercial buildings is still rare. And for a good reason: we need current and reliable data first!

The concept of turning the world into a smart object is fairly startling and represents a quantum shift in computing as currently practiced. The promise of smart systems for FM is the same as for business to consumer markets: process automation, replacing human activity to achieve efficiencies and providing great user experiences. The good thing is that we can start adopting this incrementally, making sure that each stage will return tangible value.

Implications

These technologies will deeply impact the professions of real estate and facility management over time. Here are some implications of these trends:

1. Data increasingly becomes an asset: Buildings can no longer be seen only in terms of their physical properties. Data management around buildings is something to plan for and manage. Although traditional vendors will provide for smart solutions, we need to understand principles and develop new and stronger partnerships with IT.
2. Planning and applying: There is so much opportunity and yet so little time (and money) to make this happen. RE and FM managers need to understand and plan for smart innovations. This requires understanding not only potential benefits but how to implement the new opportunities that smart technologies provide.
3. Engineering skills: To effectively implement smart systems, FMs will need staff with operational knowledge as well as an understanding of the principles behind the wider IoT technologies available. At present, there is a vast shortage of engineers with the appropriate skillsets.
4. Privacy: Issues related to what personal data is gathered and determining the appropriate level of privacy have yet to be resolved. Where data on the workforce and buildings is concerned, these determinations should involve stakeholders outside of IT.
5. Security: IoT’s ubiquitous nature is both beneficial and daunting. Options for hacking into systems multiply, producing new risks for facilities. Risk analyses, continuity of operations, mitigation and contingency plans will be critical tasks for organizations and should not be performed solely by IT staff.

IT increasing will play a more fundamental role in the lives and work of real estate and FM professionals as smart systems become available and proliferate. We need to prepare ourselves for a smart future.

REFERENCES

1. www.gartner.com/technology/home.jsp
2. Gartner publication, The Internet of Things and Related Definitions, 2014.
3. Gartner Digital Workplace Summits, 2015, by Tom Austin, one of Gartner’s Analysts for IoT.
4. In short, the term machine learning or deep learning points to the ability of computers to learn from large sets of data provided to them. A well-known example of this is the ability of systems to identify individuals by facial recognition. It takes millions of pictures of people in order for systems to learn this, but now they are more reliable at it than humans.
5. In the September/October 2015 issue of FMJ, the article Social Physics: A Science to Watch for FM describes principles of quantification of human behavior around the workplace. See also: Social Physics by Alex Pentland (MIT), 2014, and People Analytics by Ben Waber (MIT, Humanyze), 2013.
6. The essence of learning systems is that their behaviors