Get ready with a training problem on the sandbox area

[gmacario]

From https://challenges.reply.com/tamtamy/challenge/5/detail

On 15th March, you will have 4 hours to figure out a typical Reply engineering problem. Would you like to help us? Team up to solve the problem that will be released on 15th March, at 4.30pm CET using your favourite programming language and win a Razer Blade Gaming Laptop. Your team can have two, three or four people.

Get ready with a training problem on the sandbox area.

[ludusrusso]

The training problem is a simple path planning algorithm to be developed. It can be easily solved with a well known class of algorithms called RRT*, see below:

• https://en.wikipedia.org/wiki/Rapidly-exploring_random_tree
• https://github.com/s880367/RRT

[gmacario]

It can be easily solved with a well known class of algorithms called RRT*

@ludusrusso I beg your pardon but as far as I understand your answer does not qualify as a complete fulfillment of what detailed in https://challenges.reply.com/tamtamy/challenge/3/detail

How about issuing a new https://github.com/ludusrusso/rokers-hunting-razerblade/pulls with the proper files in attachment?

[xrmx]

@gmacario I trust @ludusrusso as the algorithm guy :)

[gmacario]

@ludusrusso @xrmx and I am eager to extend my Jenkins pipeline to run the .py against the provided inputs in order to create the output files - See https://github.com/ludusrusso/rokers-hunting-razerblade/issues/11

[ludusrusso]

@gmacario It can be easily solved with RRT* algorithm. There is some boring stuff to do to:

1. Generate a proper input file for RRT* from the input file of the challenge
2. Pass those file to the RRT* or one of the derivated algorithms.
3. Generare an output file from the output of RRT*

@FiorellaSibona can write the real code! She's studying this stuff 🥇

[FiorellaSibona]

@ludusrusso @gmacario @xrmx I am actually still not aware of how RRT* algorithm works but I can try to take a look and also try to understand that library!

[FiorellaSibona]

@ludusrusso @gmacario @xrmx If I well understood, the tricky part lies in the definition of triangular obstacles. I mean, given the code Ludo linked, the only part which must be modified is the obstacles definition (triangles are not defined in the pygame library as far as I understood) and relative methods to check collision. Here a good alternative to proper triangle objects is provided, but there should be a way to identify all needed rectangles approximating the desired triangle obstacle. Any ideas? For what concerns the input and output "conversions" it should be a matter of reading a .txt file and extrapolate related info and once RRT is performed, the travelled nodes should be gathered and saved in a new output .txt* file. Do you agree?

[xrmx]

@FiorellaSibona Maybe something like this https://stackoverflow.com/questions/2049582/how-to-determine-if-a-point-is-in-a-2d-triangle ?

[FiorellaSibona]

@gmacario @ludusrusso @xrmx I finally identified a polygon definition library called pylygon and I succesfully modified the code to perform RRT* with triangular obstacles. Meanwhile I am trying to gather values from input file and save output and then we're done

[gmacario]

@FiorellaSibona

Meanwhile I am trying to gather values from input file and save output and then we're done

How about starting a PR "(WIP) Implement code to perform RRT with triangular obstacles" so @ludusrusso and @xrmx may help you - and I will learn Python while enjoying the conversation?

[gmacario]

@FiorellaSibona Any updates about the Pull Request?

[FiorellaSibona]

@gmacario Sorry I have just read your comment. I have created a training branch and pushed my code. Not able yet to solve some issues since it is not trivial as it seemed. Sorry but I am also a sort of newbie in Git so still learning. If I have done something not correctly, please feel free to clean it up and explain. Thank you

[gmacario]

Deferring resolution to a later milestone, now focusing on the real challenge