Home » Community » Forums » » Vehicle Control with Neural Networks
  Intel sponsors gamedev.net search:   
[Control Panel] [Register] [Bookmarks] [Who's Online] [Active Topics] [Stats] [FAQ] [Search]

Add Forum to Favorites |  Send Topic To a Friend | View Forum FAQ | Track this topic


 Last Thread Next Thread 
 Vehicle Control with Neural Networks
Post Reply 
Excellent article.

Just a shame there's no demo program with it, or at least none that I could find .



GDSAUBY GameDev Society For UnBanning YodaTheCoda
If you want to see yoda unbanned then put this in your sig

 User Rating: 1001   |  Rate This User  Send Private MessageView Profile Report this Post to a Moderator | Link

Great article, but no demo. what a shame!

 User Rating: 1015   |  Rate This User  Send Private MessageView Profile Report this Post to a Moderator | Link

Very interesting! I was curious to see how it works, so I have created a demo with Delphi. I don't know how to attach it, so if you want to see it, tell me and I'll send it by mail (300kB). The code is included, but it's totally messy because I took my old character recognition neural net to start with.

 User Rating: 1015   |  Rate This User  Send Private MessageView Profile Report this Post to a Moderator | Link

gr8 article, but it would have been nice if there was a running demo with it :-)


 User Rating: 1034   |  Rate This User  Send Private MessageView Profile Report this Post to a Moderator | Link

Why is there not a source implementation with this?

 User Rating: 995   |  Rate This User  Send Private MessageView Profile Report this Post to a Moderator | Link

Seems a little silly to me to be using a supervised learning algorithm for a control task. Why not use policy-gradient reinforcement learning? It'd be about a dozen lines more code, you wouldn't need to specify input-output pairs, and the behaviour would be a lot more realistic.


 User Rating: 1015   |  Rate This User  Send Private MessageView Profile Report this Post to a Moderator | Link

Since everyone seems to want to see a demo, here it is...
http://www.kuenzle-family.ch/philippe/gl/SeniorProject.htm

Philippe Kunzle

 User Rating: 1027   |  Rate This User  Send Private MessageView Profile Report this Post to a Moderator | Link

quote:
Original post by nobbis
Seems a little silly to me to be using a supervised learning algorithm for a control task. Why not use policy-gradient reinforcement learning? It'd be about a dozen lines more code, you wouldn't need to specify input-output pairs, and the behaviour would be a lot more realistic.



It seemed a little silly to build an airplane to fly, and it seemed a little silly to build a rocket to go on the moon... So it seems a little silly to explore new horizons?


 User Rating: 1015    Report this Post to a Moderator | Link

Sry :) Just a little correction - in the 3rd part of this very interesting article I think it should be "Steering = Output[1] - ..." instead of "Steering = Output[0] - ...".
And how is this distance formula working :P ?

 User Rating: 1015    Report this Post to a Moderator | Link

Quote:
Original post by RamboBones
Excellent article.

Just a shame there's no demo program with it, or at least none that I could find :P.



GDSAUBY GameDev Society For UnBanning YodaTheCoda
If you want to see yoda unbanned then put this in your sig




 User Rating: 1015   |  Rate This User  Send Private MessageView Profile Report this Post to a Moderator | Link

All times are ET (US)

Post Reply
 Last Thread Next Thread 
Forum Rules:
You may not post new threads
You may post replies
You may not edit your posts
You may not use HTML in your posts
Jump To:
Administrative Options: