I didn't start out planning to build a robotic mower portal.
I already owned an older Husqvarna Automower 450X that I call Old Faithful. Then I added a newer Automower 520, Mrs. Mow-It-All, for my backyard in Phoenix, Arizona.
That's when I started noticing something interesting.
Old Faithful (450X) could mow for hours on a charge. Mrs. Mow-It-All (520) would mow for roughly an hour and a half, return to charge for close to an hour, and then go back out again. The issue was that if the session was set at 3 hours, after charging for 51 minutes, there was hardly any session time left for the mower to mow. There wasn't necessarily anything wrong with either mower—they were simply behaving very differently.
That got me wondering: How long does a robotic mower really need to mow?
The Automower Connect app does a good job of telling me what my mower is doing. But I wanted to know more about what it had been doing over time. How long was it actually mowing before it needed to recharge? How much time was being spent charging instead of cutting grass? And most importantly, was all that mowing actually necessary?
I asked Chat GPT to help me collect and analyze the mower data.
I also began using ChatGPT extensively throughout the process. It helped me work through the mower data, design tests, analyze the results, troubleshoot problems and eventually build the Portal you're looking at now. I supplied the observations and real-world testing, and together we kept asking the same basic question: What does the evidence actually tell us?
That answered some questions, but it created another one.
Testing Mrs. Mow-It-All (520) showed that about 100 minutes looked like a sensible mowing session because it allowed her to do most of her mowing before needing a lengthy recharge. But there was an obvious problem with that conclusion:
How did I know 100 minutes was enough to mow the lawn?
I didn't.
The mower's own location data wasn't accurate enough to answer that question reliably, so I put an independent GPS logger on top of Mrs. Mow-It-All (520). I named it Miss Daisy. Yes, I actually drove my robotic mower around with a GPS logger riding on top of it. 😄
After testing the logger itself, I ran a controlled series of mowing sessions and measured where Mrs. Mow-It-All (520) actually traveled. That showed something I couldn't see from mowing time alone: areas missed during one session were progressively reached during later sessions.
Now I wasn't just guessing that 100 minutes seemed reasonable. I had evidence that it was actually working on my lawn.
The project became even more interesting when I moved Old Faithful (450X) from Phoenix to our family vacation home in Pinetop-Lakeside, Arizona.
Suddenly I had the same mower working in a completely different environment—a larger and more irregular lawn, pine trees, different growing conditions and a different climate. What works well for Mrs. Mow-It-All (520) in my Phoenix backyard may make no sense at all for Old Faithful (450X) in Pinetop-Lakeside.
That reinforced something ChatGPT and I had already begun to recognize: there probably isn't one “right” mowing strategy for every robotic mower or every lawn. The useful question is what works for this mower, on this lawn, under these conditions.
That's also why the Portal keeps the history for each mower and lawn separately. Eventually I'd like to see whether the same approach can help owners of other robotic mower brands as well.
And that's really how this Portal grew.
I didn't want another version of the Automower Connect app. I wanted something that could take the information my mowers were already producing, combine it with other measurements when necessary, and help answer the questions I actually had as an owner.
Even now, I'm still learning. A recent change in cutting height appears to have affected how long Mrs. Mow-It-All (520) can mow on a charge. I changed the height back, and now I'm waiting to see what happens. The Portal gives me a history and a baseline to compare against instead of relying on what I think I remember.
That's the idea behind My Robotic Mower Laboratory.
I'm not trying to collect more data just for the sake of having more data.
I want the data to answer useful questions.