Improving Wireless Design With AI
Can AI actually help network engineers do their jobs better? That’s what Hamina CEO Jussi Kiviniemi and WLAN Pros founder Keith Parsons discussed on a recent episode of the Heavy Wireless podcast. The episode focused on how machine learning is starting to change how networks are designed.
Practical ML, Not Sci-Fi
When we’re talking about AI and machine learning, we're not talking about robots taking over. We're talking about practical machine learning that handles specific tasks, like interpreting floor plans, identifying walls, or improving heatmap accuracy.
As Jussi explained: "Unlike traditional computer programs that are deterministic and have to follow exact rules, machine learning just takes the best guess at what to do in this situation based on its training data... Machine learning can be like, oh, I've seen this kind of stuff before, let me take a stab at this."
Traditional tools rely on pre-written algorithms, such as "calculate signal loss based on free space path loss and certain object attenuation rules”. But interpreting walls from a messy JPEG image or a fire escape plan captured with your phone camera? That's where machine learning shines.
Hamina AI automatically builds a 3D model of a building, saving wireless engineers hours and days of manual work
Automating the Tedious
Keith noted the love-hate relationship with wall drawing: "I kind of enjoy drawing walls. It's a little relaxing, and you're taking away because your system actually draws walls for me."
When AI handles wall drawing, engineers focus on what matters: access point placement, coverage needs, and optimization. And the machines are getting good at it - sometimes better than people:
"It's not been once or twice when I've looked at a drawing and the results of machine learning, and I've been like, huh, it made a mistake there. And then I look closer and I'm like, no, actually I was wrong," Jussi shared.
AI handles the drawing, you choose the materials. Select pre-made materials for auto-detected walls and doors or create your own custom material.
What Makes Good ML?
As Bob Friday likes to say, “Good grapes make good wine.” Similarly, good training data makes good machine learning. But you also need the right team. As Jussi explains, "You need incredibly good, carefully annotated data, and then you need the smartest engineers you can get. And they need to know math."
Jussi also emphasized the importance of culture: "These incredibly talented people... they will only choose to stay there if your people and culture are a fit for that person."
AI Won't Replace Network Engineers
Keith put it perfectly: "Don't be afraid of losing your job because as humans, we're still better. But the things that might have been boring... If the tool can make that go faster, then I can focus on the things I really need to learn."
Jussi agreed: "I think if your job was to only click walls, that job might be obsolete. But the network engineer's job was never to build walls; it was a necessity we had to do."
Hamina Network Planner automatically detects the scope of your project.
What's Next?
According to Jussi, the future lies in complementing, not competing with, what infrastructure vendors do well:
"We definitely don't want to create overlapping features with the Wi-Fi networking vendors. They already do a fantastic job... But anything that's left outside of that realm, and often it comes down to real-time visual type of things... that's where our tools vendors have played."
Final Thought: Smarter Tools, Better Networks
AI isn't replacing engineers; it's eliminating tedious tasks so they can focus on building great wireless experiences. As Keith noted, the basics haven't changed, but now those basics are enhanced by tools that learn and adapt.
And when it's done right? We'll spend less time manually tracing walls and more time designing networks that just work.
🎧 Listen to the full conversation on the Heavy Wireless podcast
Heavy Wireless is a deep dive into Wi-Fi, IoT, wireless security, and more. Host Keith Parsons, CWNE#3, draws on his decades of experience to educate, entertain, and inspire those working in the WLAN industry.