I have been using AgOpenGPS for a bit now and love how it helps streamline my farming operations. Recently, I have been thinking about how AI could be integrated into AgOpenGPS to improve precision and decision-making even further.
For instance, could AI be used for things like optimizing field coverage patterns, predicting crop or soil conditions based on historical data, or even automating some of the decision processes during planting and harvesting? Iam not entirely sure how to go about adding AI to the system, but it seems like there could be a lot of potential here. I have gone through some resources/articles AgOpenGps for a beginnerGenerative AI Roadmap, however they are quite useful but I want to learn from members of community.
Has anyone in the community explored AI integration with AgOpenGPS? Or maybe you have some thoughts on how it could be done? I would love to hear your ideas or experiences.
First of all, a disclaimer I’ve done very little with AI, beyond using ChatGPT, and some AI auto completion. So some of my ideas could be way wrong!
I have thought about using AI for autotuning the steering parameters. But the more I think about it, the more downsides I see. First, either the tablet would need an internet connection at all times, with the model running on a server ($$$), or the AI model would have to be running locally on the machine. And for sure running locally (I think) would overpower the tablets most of us use. And even remotely might use too much CPU, or cause network lag. But a lot of this could only be verified by trying.
And all of us have had AI feed us answers completely, obviously wrong ( one of this community told me ChatGPT told him to include Windows.h in an Android C++ application ). When that happens, no big deal, the human fixes it. But on a tractor? A lot of damage (or worse) could be done quickly, especially if AI was telling AOG where to steer next.
And time. AI integration would be slow, and require a lot of learning. Whether it would be worth it all, I don’t know. But I for one don’t have the time.
But if Brian had walked into his local dealer and told the tech guys there that he was going to make his own autosteer, likely they would have had reasons why it wouldn’t ever work. And it obviously did!
Maybe all I’m doing here is giving the reasons why I haven’t tackled this.
AI isn’t required for autotuning, it’s the wrong solution… code to analyse the XTE results and change parameters one way or the another, measuring any improvement or worsening, and so on.
I’ve done very little with AI, beyond using ChatGPT, and some AI auto completion. So some of my ideas could be way wrong!
). When that happens, no big deal, the human fixes it. But on a tractor? A lot of damage (or worse) could be done quickly, especially if AI was telling AOG where to steer next.