<@U9TLLJDD2> thanks for the contribution. It is fi...

# mathematics

[altavir]

@ESchouten thanks for the contribution. It is finally merged. I want to do some post-merge changes. Mostly related to computation of the "costs" of path. We have cases, where it is not "path length", but some value that depends on the specific point. So it will be some kind of visitor. Also I will probably change structure somewhat to better conform to the other code. Meanwhile, you can check the complementary project: https://github.com/mipt-npm/maps-kt. It allows (or will allow) to visualize the path on maps and schemes.

[ESchouten]

Good to hear! Excited to see the coming changes!

[ESchouten]

@altavir Turns out the maps-kt project is useful for me as well! The reason I created the Dubins project was to steer a tractor driving rows doing field work. Might have to revisit and continue the project 😜

[altavir]

We also are thinking about tractors, but it is a second priority right now. First is ships. For tractors, you will probably need schemes, not maps: https://github.com/mipt-npm/maps-kt/tree/main/maps-kt-scheme. This moudle is currently independent from maps. I did not manage to make reasonable common API between them so I will be just porting features from maps there as soon as they are needed.

[ESchouten]

Exactly, IMO tractor "autosteer" is not very complex, but is very effective for a huge portion of all farmers. Instead of 50 cm overlap due to driver inaccuracy it can easily be reduced to 5-10 cm with cheap hardware or a dual band gps phone. Plus it makes the drivers work a lot more relaxed.

[altavir]

OK, we have some contacts with probable investors for such job. If ships work well, we will move to it as well. I am not sure about the financial model, but I want to do an open-core development.

[ESchouten]

@altavir There are several more expensive, proprietary solutions and one open-source, AgOpenGPS.
I believe AgOpenGPS is overly complex, if i remember correctly, client code not properly separated from business logic, and thus can only run on x86 windows machines sending signals to a complex PCB you have to construct yourself.

[altavir]

Interesting. Please write to me (privately) if you have any ideas, we well discuss it with our management. I plan to do a modular solutions. And we have a way (https://github.com/mipt-npm/controls.kt) to integrate it with the hardware.