![page-6--_post-3](/_cms/media/page-6--_post-3.jpg)
Establishing a Smart Building Industry Standard
![page-6--_post-3](/_cms/media/page-6--_post-3.jpg)
Republished from AutomatedBuildings.com
We all spend the first hour of a meeting establishing what we mean by smart, how smart is smart, navigating disbelief, educating about new technology, and finally arriving at common ground. There is a better way.
The problem with today’s “smart buildings” is they’re not actually smart. At best, they are merely “connected.” Of course, we all know this to be true, but as an industry, we lack the vocabulary to distinguish how smart “smart” is, and thus, we lean on ambiguous terminology. This presents a serious problem as we chart the path from the First Industrial Revolution legacy to our inevitable future arrival at fully autonomous building systems.
Why Vocabulary Matters
The term “Smart Building” has been in use since at least the 1990s, coined to reference something “more” than just automation. It emerged from the “Smart” lexicon — smartphone, smart home, smart grid, smart cities — invoked when a product is more 4th Industrial Revolution than 1st.
Our industry — automation — was conceived of this 1st Industrial Revolution legacy. The basic control loop, typified by the mercury switch thermostat, still lives on as an emblem of the ingenuity of this early 19th-century paradigm shift. In fact, the senior statesmen of our industry — Honeywell, Siemens, Johnson, and Schneider — were all born out of this early era.
But as our industry transitions to a diverse world of autonomous systems, AI, and distributed IoT — new terminology is required. As anyone operating in the space of “smarter” systems can attest, the building automation industry’s lack of a clear lexicon is becoming a barrier to communication. We spend the first hour of a meeting establishing what we mean by smart, how smart is smart, navigating disbelief, educating about new technology, and finally arriving at common ground.
How can we as an industry provide clarity, chart a path to the future, and distinguish between technological architectures — while also communicating on a mutually understood foundation with our customers, investors, and the public?
I have a proposal.
Autonomous Vehicles Standard
Other industries have faced this before, and we can borrow from their success. Foreseeing a similar challenge, the Society of Automotive Engineers (SAE) came together in 2014 to establish a common taxonomy and terminology for autonomous driving systems. This standard charted the course from conventional 1st Industrial Revolution vehicle technology to the future of fully autonomous vehicles — with defined thresholds for each level of autonomy, from level 0 to level 5. With clear definitions charting a technological path forward, the autonomous vehicles market has thrived with technological development, architectural evolution, customer identity, market investment, and cultural speculation.
We in the building controls industry don’t have to reinvent the wheel. In fact, the future of autonomous buildings looks more similar to autonomous vehicles than different, and much of their existing taxonomy applies to the building industry.
Autonomous Buildings
What are autonomous buildings? Just like autonomous vehicles are the pinnacle in the automated vehicle standard taxonomy, fully autonomous buildings are the end-point of the “smart building” revolution. Like autonomous vehicles, these next-generation buildings “navigate” in real-time. Only instead of navigating a single car in a 2D spatial map, autonomous buildings navigate a whole “fleet” of sub-systems simultaneously in a multi-dimensional temporal map.
Fully autonomous buildings develop their own control sequences on the fly in response to changing conditions. They’re not based on static sequences, set-points, PID, or simple state machines. They understand their own underlying physics of operation and generate continuous control paths. They can introspect those same physics and provide deep insights, analytics, or more importantly — analysis. This analysis can, in turn, be used to automate the commissioning and optimization of systems. Fully autonomous building systems are aware of the future implications of their control decisions, enabling them to navigate around system “collisions” or energy “congestion” hours before it even occurs. This future-forward control is uniquely possible because buildings have accurate prediction horizons of many hours, or even days, in contrast to the mere seconds that vehicle systems have to work with.
Perhaps most importantly, autonomous buildings cooperate. Because buildings are the primary building-blocks of cities, you obviously can’t have smart cities without having truly smart buildings. Autonomous buildings will act as agents in energy networks, buying and selling energy futures using smart contracts, working with utilities and district systems, and ultimately building the backbone for future peer-to-peer grids.
Why do autonomous buildings matter? Buildings are the largest controlled infrastructure in the world economy. They are also the most complex. A typical commercial building has thousands of sensors and hundreds of control points. When we add in complex dynamics like occupancy, weather, and equipment variance, a simple state analysis shows that the typical building has trillions of possible states. This is well beyond the complexity of other industries, and beyond the limits of today’s static approach, based on forms of Finite Automata — which rely on sequence programming, PIDs, and setpoints.
Taxonomy for Smart Buildings
We propose starting with 6 levels of autonomy that are orthogonal to the SAE Autonomous Vehicles Standard. In addition to this base set, we propose adding 3 more levels to account for the greater levels of autonomy buildings can achieve due to the fact that their control and automation systems are site assembled, not limited by the factory constructed uniformity of cars. These additional 3 levels account for systems that guide their own installation (i.e. self-assembling), self-commissioning systems, and self-networking systems.
Our proposed taxonomy — including levels 6–8 that go beyond the Autonomous Vehicles Standard — is defined below.
Working Together
As we develop a language to discuss the future of our industry, the future of the control systems, and the buildings we interact with — we will start to chart a course to buildings that are truly “smart.”
Troy Harvey is the CEO of PassiveLogic, a company developing the future of automation, built on digital twins from the ground up.
![1svc1gz3vnnaiygefgjxmsw](/_cms/media/1svc1gz3vnnaiygefgjxmsw.webp)
Developing a language to discuss the future of our industry, the future of the control systems, and the buildings we interact with will prepare us to chart a course to buildings that are truly “smart.”
Level 0 — AutonomyConventional set-point or proportional control. It is directly analogous to cruise control. This is the historical industry approach.
Level 1 — Adaptive set-pointsAt this level, the system adaptively adjusts set points of otherwise static manual sequences, set-points, and/or PID control. This is directly analogous to an automobile’s adaptive cruise control, where the setpoints of a fixed control scheme are adjusted during operation. For buildings, this is often done off-site, in the cloud, in less than real-time.
Level 2 — Sequence generation assistAt this level, in addition to adaptive set points, the control system assists in generating the system sequences. The sequences are still static, though they may cover a larger state space than conventional control programming. This is analogous to driver-assist in the vehicle market.
Level 3 — Adaptive sequencesAt this level, in addition to automated sequence generation, the system can adapt its control sequences during operation. This is less than real-time sequencing, often updated from the cloud. This is analogous to the SAE conditional automation level.
Level 4 — Real-time sequencingAt this level, the system no longer uses set-points or proportional reactive control. Instead, building control uses continuous paths. Sequences are generated in real-time, at the edge. Cloud control is no longer suitable due to the resilient real-time needs of the system. This is analogous to the SAE high automation level.
Level 5 — Full autonomyBuildings that self-optimize using adaptive real-time sequencing compare future outcomes of different control schemes are introspective about their own operation, automatically analyze the building’s behavior, and continuously adapt to their environment. This level is analogous to fully autonomous vehicles.
Level 6 — Self-Guided installThis level utilizes the underlying intelligence of Level-5 autonomy, and the introspection of its own physics and interconnectivity, to guide the installer through the installation process, assessing errors and providing installation guarantees based on the system design goals and/or engineering plans.
Level 7 — Self-CommissioningThis level utilizes the automation system’s own understanding of the building and mechanical system’s physics to self-commission, validate the interconnectivity of equipment and zone-sensor interactions, and automate point mapping. Level 7 autonomy enables system guarantees for architectural, engineering, or energy design intent.
Level 8 — Peer-to-Peer InteractiveThese fully autonomous buildings can act as real-time agents in smart city networks. Given a common interface and a self-validating smart contract currency, buildings can operate as free-market actors on behalf of their owners. These fully autonomous buildings enable real-time demand-response, district energy systems, and peer-to-peer decentralized grids.
PassiveLogic is developing the the first fully autonomous platform for buildings, built on digital twins from the ground up.