Subject:
|
Re: What I would do (2)
|
Newsgroups:
|
lugnet.robotics
|
Date:
|
Fri, 27 Jan 2006 14:03:23 GMT
|
Viewed:
|
2474 times
|
| |
|
|
In lugnet.robotics, Ed Manlove wrote:
> In lugnet.robotics, Alexander Sack wrote:
> > Anyway, this is just a high-level sketch (I make no claims that this
> > architecture is best and I'm sure it would be heavily revised when the rubber
> > hits the road (crosses fingers) in February or August)...
>
> What you are describing seems more like an end application (multiple autonomous
> robots that do X and Y and can communicate using Bluetooth) then an RTOS design.
> An analogy of your description of using the NXT processor for just the
> communication link and letting the real computing power be done on my host PC is
> like having a G5 Mac count the drip rate of a leaky faucet, i.e. a big waste of
> processing power. An ARM processor (if this is what is in the NXT) is used in
> many "powerful" applications including possibly your automotive airbags (see
> http://www.arm.com/markets/showcase/).
I'm not so sure about this. Figure its an ARM7 clocked way below 50Mhz (for
power) and has 64k of memory. How much runtime state can you really process
with 64k after the kernel and sensor/motor processes are loaded?
Maybe more than I think...
Btw, thanks for the document! The current paradigm in RTOS design seems to
embed your application with the operating system. I think for the majority of
use cases, the NXT provides enough resources. However, I still think that a
messaging approach may prove to be more flexible for complex robots (how about
the host as another NXT talking to several slave NXT marshalling events and
processing logic and even sending down more tasks to schedule).
Ed, I wholeheartily agree with your comments regarding the NXT software. I
actually think the main focus SHOULD indeed be expanding the existing NXT
software stack. Come August (or hopefully sooner!), everybody who uses the NXT
should have a very positive experience installing and developing with the
existing NXT tools! I suppose if the MDP wanted folks to help with fine tuning
and testing the existing stack, I would be all for it.
-aps
|
|
Message has 1 Reply:
 | | Re: What I would do (2)
|
| (...) The processor speed (even underclocked) will not be an issue. There real issue which I see many people have talked about is the amount of memory will be in the NXT. On the website for uCOS-II there is a RAM calculation (for the Intel 80x86 (...) (19 years ago, 27-Jan-06, to lugnet.robotics)
|
Message is in Reply To:
| | Re: What I would do (2)
|
| (...) Alexander, What you are describing seems more like an end application (multiple autonomous robots that do X and Y and can communicate using Bluetooth) then an RTOS design. An analogy of your description of using the NXT processor for just the (...) (19 years ago, 26-Jan-06, to lugnet.robotics)
|
24 Messages in This Thread:
  
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
This Message and its Replies on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
