search

quimortiz / dynoplan

https://quimortiz.github.io/idbastar
MIT License
47 stars 10 forks source link

Volume/shape based conflict detection for focal search #21

Closed akmaralAW closed 7 months ago

akmaralAW commented 7 months ago

It would be helpful to know how would you solve it @quimortiz, this is what I am doing to check two motion primitives (one is a node being checked for expansion, second one is from the solution of other robots).

quimortiz commented 7 months ago

can you first explain to me what is focal heuristic? When you check two primitives are they "static", in the sense that we erase the notion of time and just check if their traces collides?

quimortiz commented 7 months ago

in the current status i cannot review the code: 1) I don't know what it is supposed to do. 2) Can you add some comments, e.g., says what is each argument to your new functions. 3) Is the code working and running correctly?

akmaralAW commented 7 months ago

@quimortiz, comments are added to the script. I will merge it to ecbs branch for now. Answering to your questions: 1) this heuristic function is used to count the number of conflicts for the discrete search. It checks each motion primitive, that is considered as applicable for expansion, for collision with the solution/result of neighbor robots. 2) Comments are added 3) Yes, it is working and runs. At the moment I call it inside tdbA*-epsilon for focal set maintenance.