[Dev] Tizen use of FIFO & Nice

Stoppa, Igor igor.stoppa at intel.com
Wed Dec 18 09:49:05 GMT 2013


On 18 December 2013 11:35, Jussi Laako <jussi.laako at linux.intel.com> wrote:

> On 12.12.2013 22:18, Stoppa, Igor wrote:
>
>> And even the reporting of speed is such that it generally has quite
>> high tolerance to errors in the reporting (higher tolerance at higher
>> speeds, up to 10-15%)
>>
>
> That would be a bit too much from the measurement system, because wheels
> (wearing, temperature, etc) already generate large part of the error.
>

Actually some years ago I worked on a dashboard where the requirements for
the error on the end of the scale were even more relaxed (120-140 km/hour,
[0-20%] )
At lower speed also the error got smaller (0-50 km/hour, [0-5%])
And no, I won't say which brand/market it was. =)

Digital speedo and tacho on my motorcycle has about 10 Hz update rate. In
> my car digital speedo is slower, but tacho is still pretty fast.
>
> On engine with light weight flywheel it doesn't take long to hit red
> line...
>
> When I was doing ECU-style software, much higher update rates were needed,
> but I guess IVI system is not supposed to handle ECU, ESC or ABS? For
> example some modern engines don't have throttle flaps in the intake
> manifold anymore, but adjust throttle completely by adjusting camshaft
> timing...


With the sort of car models that are likely to mount an IVI system I would
expect that all the stuff would be chewed by some other uC that is located
closer to the sensor in the food chain, and has also the role of
(in)directly controlling the operational parameters, like temperature,
air/fuel ratio, etc.

IVI should just get the post processed values over CAN bus and make sure
that any virtual needle or similar widget is rendered smoothly and conveys
the feeling of lots of horsepower under the hood.
That's what most customers pay for anyway, also with traditional dashboards
=p

-- 
cheers, igor
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.tizen.org/pipermail/dev/attachments/20131218/cdf48040/attachment.html>


More information about the Dev mailing list