My point of it being on a separate board was that if something happened to the main system itself it could restart it. A watchdog program running on the same system it's watching could fail along with it's system. But one that isn't governed by it would still function.

Now, when invalid data is generated by the motherboard, wouldn't that cause(or have to be caused by) crashing or breaking in some way since computers don't just make mistakes? And wouldn't that cause an irregularity in the heartbeat or stop it entirely?
DShiznit wrote:
My point of it being on a separate board was that if something happened to the main system itself it could restart it. A watchdog program running on the same system it's watching could fail along with it's system. But one that isn't governed by it would still function.

Now, when invalid data is generated by the motherboard, wouldn't that cause(or have to be caused by) crashing or breaking in some way since computers don't just make mistakes? And wouldn't that cause an irregularity in the heartbeat or stop it entirely?


It would never lead to an "irregularity" in the heartbeat. The system would simply have a thread that pings the watchdog program (or vice versa - doesn't really matter). Either it replies or it doesn't. Invalid data in simply leads to invalid data out - not to crashing. At least, not to the crashing of a decently written program. It will, however, lead to the robot to fail (in that it crashes into something, misbehaves, does something stupid, etc...). A watchdog won't help with any of that. And if the system hardlocks and/or crashes to the point that it would take down a separate process on the same system - you aren't going to recover from that as that almost guarantees a hardware issue (unless its Windows, of course Wink ). A watchdog on a separate board will simply be able to tell you that the other system died - it can't fix it for you.
Point taken. I guess then it would be up to the person programming the robot to not be retarded and it should do fairly well.
What you can do is add an routine that calculates the risk of the decision based on the inputs, and do an certan move or another thing based on some of the inputs that may be correct, for this you will require a NN, that should be trained trought all sorts of directions or events!
Indeed, but it still needs at least some modicum of irregularity or randomness in the decisions it makes to give it some creativity or at least flexibility to respond to situations with subtly different requirements in different ways.
rayden wrote:
... for this you will require a NN, ...
Define "NN" I'm lost Confused
neural network! and almost Kerm we can include bayesian decision! based on the probability density function! it should be noticed that no pic can run this it is meant to be used with fast processors or with the use of a computer! Is the way we use on ours!
  
Register to Join the Conversation
Have your own thoughts to add to this or any other topic? Want to ask a question, offer a suggestion, share your own programs and projects, upload a file to the file archives, get help with calculator and computer programming, or simply chat with like-minded coders and tech and calculator enthusiasts via the site-wide AJAX SAX widget? Registration for a free Cemetech account only takes a minute.

» Go to Registration page
Page 2 of 2
» All times are UTC - 5 Hours
 
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum

 

Advertisement