In today's world of smart building, smart equipment, and IoT, the number of networked devices and the data they contain is immense. From building automation systems to central plant controllers to smart sensors. These devices are providing tons of real-time and historical data within the silo they're designed for.
There are platforms that bring these disparate systems together and more importantly, normalize their data to provide structure (relationships) and meaning through metadata tags. We are at the beginning of arguably an artificial intelligence evolution, which at its very foundation requires data that is clean, structured, and easy to consume. The intersection of these two points is where we live today. Here's a look at why building automation is perfectly suited for AI.
As an industry, we have done a great job getting devices talking to each other. In the early days of open standards, such as LonWorks and the proliferation of BACnet for HVAC, Modbus for power, and KNX for lighting, all contributed to open and interoperable devices. Even the physical layer has evolved over time from serial twisted pair wire to IP Ethernet-connected devices and even wireless using Wi-Fi and other technologies such as LoRaWAN.
These networked systems not only aggregate devices at the local site, but also can be connected to the internet for remote access and cloud-based services (AI). Of course, cybersecurity and secure connectivity are even more important for this edge-to-cloud architecture.
In the building automation space, these devices often contain dozens – even sometimes hundreds of points representing sensors, actuators, and set points. Modern Building Automation Systems (BAS) have the ability to share both real-time and historical data, typically used for dashboards and analysis. This high density of points becomes a wealth of information for other applications such as LLMs and agentic AI.
With this vast amount of data coming out of smart building systems comes the realization that it's both unstructured and difficult to consume. This is not a new challenge and has been solved by data standards such as Project Haystack, which models and defines metadata for subsystems in smart buildings. Now the data is self-describing, meaning that it's easy to determine what it is and what it's related to.
The BAS protocols are transported across various physical layers. So the very first step in normalizing data is to get each of the physical layers (serial, Ethernet, wireless) consistent. Modern day supervisory controllers take care of the translation of the electrical signals coming from the devices and connect it to the Ethernet backbone. Similarly, the protocol or language spoken by the devices is normalized in these controllers (and at the server if they are IP protocols).
J2 Innovations' FIN Framework is the platform that unifies devices and their data independent of physical layer or protocol. It's the same platform whether it's at the supervisory controller or the server providing the normalized data, complete with Haystack standard built in.
As AI use cases become more prevalent, the need for "clean data" has become clear. We all have heard of AI hallucinations and AI slop (to name a few of the negative outcomes). Conversely, given a clean dataset, good context, and a solid model for the data, AI-based applications have demonstrated incredible insights and utility.
In smart buildings, devices generate data across multiple subsystems, including HVAC, lighting, energy, and others. There has been major progress made in device interoperability through open standards. In addition, applications can now easily access the data thanks to the combination of unifying platforms, leveraging a data standard such as FIN Framework / Project Haystack.
Some of the first AI-based applications in building automation, such as chatbots and intelligent agents, are already in use, and the potential applications extend far beyond what we see today. This progress will be driven by the availability of clean, structured data, brought together on a unified platform and governed by an open standard.
Topics from this blog: Smart Buildings Technology Industry Building Automation System BAS AI GenAI
Back to all posts
