Transforming Building Automation with AI Technology

00:00

00:00

1x

• 0.25
• 0.5
• 0.75
• 1
• 1.25
• 1.5
• 1.75
• 2

Alex Shatalov Data Scientist & ML Engineer

AI in building automation is changing how commercial and residential properties are managed. If you already use a building automation system (BAS) to control heating, cooling, lighting and security, you’re already halfway there. Now imagine taking that system to the next level by adding artificial intelligence.

But here’s the thing: integrating AI into an existing system requires planning. It’s not just about adding technology for technology’s sake, but about creating a more sustainable and cost-effective system.

This article reveals all the secrets of building automation with AI based on our experience.

Building Automation System in Construction: CHI Software Case Study

Buildings become smart long before occupants start to live there. Take this story as an example. Our client aimed to simplify the construction process through AI automation. Now, let us provide some context before we dive into technical details.

A Few Words About Ensuring Building Stability in Japan 

As you know, Japan experiences a lot of seismic activity, so ensuring structures’ stability during construction saves lives. 

Typically, builders have to attach colored markers to each section of the reinforcing mesh to show they have the correct number of reinforcing bars. Moreover, these markers must be attached both horizontally and vertically, which, as you might guess, is a repetitive task.

How to Make an AI App: A Massive Integration Guide for 2024 Continue reading

But the main problem is that plastic markers often fall off, and the whole marking process has to be done over and over again. On top of that, builders attach a ruler to the photo to measure the distance between the reinforcing bars.

Imagine how physically demanding and time-consuming it is to carry out this work of marking the reinforcing bars and calculating the distances. Can AI-based building automation systems simplify the whole process?

At First, There Were Some Challenges

Yes, technology indeed can be helpful. But great achievements do not come easy. This is what we had to face at the beginning of the project:

• Figuring out how to pick the first layer from all those reinforcing bars and measure the gaps between them was pretty tricky. Plus, we had to do it using only an image, with no additional sensors and smartphone-based tools;
• After thinking it over, we realized we needed something to help measure stuff. So, we came up with a special tape, which worked like a filter and a ruler for our computer vision system;
• Making this tape was not a walk in the park. We had to get it just right: see-through enough to spot the bars in front clearly but not so much for the ones at the back;
• We went through a bunch of trials and errors to create an algorithm that turned our idea into something tangible;
• And finally, we had a tight deadline: our client wanted the project up and running in three months.

How Our Team Transformed the Marking Tasks

This solution differs from a typical BAS, as we had to meet very specific business requirements. So, how did we act? These are the main project milestones: 

• First, we used image pre-processing techniques inspired by classical computer vision (CV), as we didn’t have enough images to apply more advanced neural networks. We cleaned up the image, eliminated unnecessary objects, and highlighted the main rebars. We wanted to make sure we separate the right rebars from the wrong ones;
• Then, our engineers implemented Line Detection with Hough Transform. This nifty technique helped us detect rebars on images super quickly and efficiently;
• The CHI team also dabbled in different color spaces and morphological image processing. No deep learning models were involved here. We managed to detect tape without all the complexity.
• Of course, we had to do some post-processing magic. Our engineers wrote a bunch of logic to filter through the rebars and select only those that truly mattered. 
• Last but not least, we created a genius automatic distance calculation mechanism between the rebars. No need to break out the measuring tape anymore!

Read this case study in full to learn more about building automation with AI.

What Are the Best Practices for Combining AI with Building Automation Systems?

Regardless of the goals you have in mind, some time-tested practices will help you build an outstanding AI-powered BAS. 

These simple steps will allow you to take your BAS to the next, innovative level.

1. Embrace the Power of Data

AI thrives on data. The more quality info your BAS can provide, the smarter your algorithms become. It is like feeding a brain: the better the food (data), the stronger it grows. Make sure your system is well-prepared to collect, store, and analyze information efficiently.

2. Continuous Learning is Key

Algorithms are not set-it-and-forget-it tools. They should be a learning entity. Regularly update your system based on new data and feedback. Think of it as continually training an athlete for peak performance.

Looking for a team to implement these AI practices? Leave your message to our engineers

3. Focus on User Experience

Remember, technology is for people. Your AI-driven automation system should make life easier, not more complicated. Strive for intuitive interfaces and seamless integration. It should feel like a helpful assistant, not a complicated puzzle.

4. Collaborate with Experts

Do not do it alone. Integrating AI with your BAS is usually a complex task. Work with specialized experts to tailor solutions to your specific needs. Building a dream team is a proven way to achieve your goals faster and more precisely.

5. Prioritize Cybersecurity

With great power comes great responsibility. AI BAS can be a goldmine for cybercriminals, so prioritize fortifying your cyber defenses. Regular updates, strong encryption, and vigilant monitoring are your digital knights in shining armor.

The Role of Computer Vision in Business Automation: Insights Based on a Real Case Study Read more

6. Test, Test, and Test Again

Before going all in, test your AI integrations in a controlled environment to identify potential issues and fine-tune the system. Think of it as a dress rehearsal before the big show.

7. Balance Automation with Human Oversight

While AI can do wonders, human oversight remains crucial. Ensure there are protocols for human intervention in case of anomalies or emergencies. Algorithms can help, but they will not replace human intelligence.

By following these best practices, you can harness the full potential of AI in building automation to create smarter, more efficient, and more user-friendly environments. Welcome to the future of building management! 

What to Expect in the Future of BAS with AI Integration?

The future of BAS and AI is bright. As intelligent tech continues to advance, so do BAS capabilities. Here are a few ways AI-powered building automation may impact the industry in the coming years.

Predictive Maintenance

AI can analyze vast amounts of data from BAS sensors and equipment to detect patterns that indicate potential issues before they become problems. This allows facility managers to perform preventative maintenance and avoid costly downtime or repairs. Artificial intelligence in building automation may get so advanced that it can predict failures of individual equipment pieces months or even years in advance.

The Big Change: Top AI Trends Transforming the Business World in 2024 Read more

Optimized Energy Usage

By analyzing historical data on a building’s energy usage, occupancy, and other factors, AI can determine the optimal settings for HVAC, lighting, and other systems to minimize wasted energy. It may even learn the preferences and schedules of a building’s occupants to maximize comfort while reducing usage. Some companies have already implemented AI to cut building energy usage by up to 30%.

Enhanced Occupant Experiences

AI has the potential to create highly customized experiences for building occupants. It could analyze data from smartphones or access control systems to anticipate individuals’ needs and preferences. For example, as someone enters the building, algorithms may adjust the temperature in their office to their ideal setting while turning on their preferred lighting. AI could also provide personalized wayfinding, recommendations, and other services.

Autonomous Operation

Eventually, AI may reach and exceed human-level intelligence for building automation tasks. At this point, BAS could operate largely autonomously using AI to monitor systems, detect and address issues, optimize performance, ensure occupant comfort, and more with minimal input from facility staff. 

However, humans would still maintain oversight and the ability to override AI controls if needed. This level of autonomy could significantly reduce costs while improving efficiency, sustainability, and occupant experience.

Conclusion

Could we imagine in the 20th century how far the construction industry will go? But moving into 2024, you will hardly surprise anyone with the phenomenon of smart cities and all the opportunities it brings. 

We have just shown you how AI becomes a new era in people’s safety long before they live in a building. How about predictive maintenance or optimized energy usage? All these capabilities can be implemented in one building, eventually impacting the whole neighborhood or even the city’s prosperity. 

But all great beginnings start with a small step. We are sure you will not stand aside and wait. So, start now. Pick the best AI engineers and share your ideas with them. CHI Software is a place where dozens of amazing AI projects originate from. Will yours be next? It is for you to decide. Just let us know if you are ready for a conversation.

About the author

Alex Shatalov Data Scientist & ML Engineer

Alex is a Data Scientist & ML Engineer with an NLP specialization. He is passionate about AI-related technologies, fond of science, and participated in many international scientific conferences.