Advice using i2 pro

[GaryA]

Hi Guys, I have a few quires regarding the i2 pro software.

To give a bit of background i am using an ACL which logs GPS (position, speed, etc.) from a CAN input using a VBox i also have a VCS recording video (the enable video sync has been ticked) and there are also 7 other cameras placed around the vehicle. The idea is that the VCS will allow me to sync the data to the video and use this to sync the data to the other camera views. To help with the time slip a clock was placed in front of each camera before each test.

Now the issues,

1) the tests we were conducting were in very hill terrain with dense vegetation which caused the GPS to dropout, unfortunately when this happened the ACL recorded a speed of zero and a lat and long of zero which is more annoying, because now when i try and export a Google earth file it gets confused and tries to plot some points of the coast of Africa. This is annoying but the real issue comes when trying to generate a map in i2, I believe i had found a solution which would involve creating a maths channel that would eliminate these erroneous values however this does not work, i think i2 requires an uninterrupted data source to generate a map., can you suggest a way i could address this?

2) is it possible to setup camera vies for all cameras that will automatically load when the project loads?

3) i think i noticed someone else ask this (or similar) question, is it possible to generate the map over the Google earth image in i2. so that when using i2 you can see the google earth image.

Thanks
Gary

[Holmz]

1) the tests we were conducting were in very hill terrain with dense vegetation which caused the GPS to dropout, unfortunately when this happened the ACL recorded a speed of zero and a lat and long of zero which is more annoying, because now when i try and export a Google earth file it gets confused and tries to plot some points of the coast of Africa. This is annoying but the real issue comes when trying to generate a map in i2, I believe i had found a solution which would involve creating a maths channel that would eliminate these erroneous values however this does not work, i think i2 requires an uninterrupted data source to generate a map., can you suggest a way i could address this?
Gary

I think it was Darrin that told me.. Something like is num-Sats = 0 then "Invalid" or "NaN".
I have not tried it, but it is somehow making the values not a number value.
So yes it can be done.

3) i think i noticed someone else ask this (or similar) question, is it possible to generate the map over the Google earth image in i2. so that when using i2 you can see the google earth image.

Someone asked it this past week.
I usually export the data as cvs, and then do the work off line in MATLAB or whatever, maybe python, and then create the KML to get it into GoogleEarth.

[GaryA]

Thanks for the reply,

Problem 1 is solved, I Thought I tried your solution but when i tried it this morning it worked fine!

I have the KML files and they all look good on Google earth, however what i mean is i would like to use the Google earth background in the i2 software as i think it will look a little better than a plan white background. Although it is no great loss if this cannot be done.

[Holmz]

Thanks for the reply,

Problem 1 is solved, I Thought I tried your solution but when i tried it this morning it worked fine!
...

Can you please post exactly what you did symbol by symbol...
That will save me some time.

[GaryA]

I did something slightly different to excluding all values of 0. the reason being that occasionally when dropping out and reconnecting the value is not 0 so i limited the rate of change of the lat and long coordinates. Below is the code i used, it is the same code for longitude with the appropriate variables changed.

choose(derivative('GPS Latitude1' [deg])<0.001,choose(derivative('GPS Latitude1' [deg])>-0.001,'GPS Latitude1' [deg],invalid()),invalid())

Although i cant think what i did different to make it work.

my next challenges are:

to calculate distance traveled using either GPS coordinates or figureing a way to correct for speed when the GPS dropped out.

and to calculate time in HH:MM:SS from a value of seconds after midnight.

[Holmz]

I did something slightly different to excluding all values of 0. the reason being that occasionally when dropping out and reconnecting the value is not 0 so i limited the rate of change of the lat and long coordinates. Below is the code i used, it is the same code for longitude with the appropriate variables changed.

choose(derivative('GPS Latitude1' [deg])<0.001,choose(derivative('GPS Latitude1' [deg])>-0.001,'GPS Latitude1' [deg],invalid()),invalid())

Although i cant think what i did different to make it work.

Thanks.

my next challenges are:

to calculate distance traveled using either GPS coordinates or figureing a way to correct for speed when the GPS dropped out.

GPS speed is almost always low as it is "ground Speed". For instance going off a cliff GPS-speed is zero.
For GPS look at topographic to ENU transforms.
Look for this "Geodetic to/from ECEF coordinates" about 60% of the way down in here [url]http://en.wikipedia.org/wiki/Geodetic_system#Local_east.2C_north.2C_up_.28ENU.29_coordinates
[/url]
Once you are in ENU, then delta-dist = (deltaE^2 + delta^N^2 + deltaU^2)^0.5
Speed = Delta-Dist/Delta-time ... That last part (dT) might be variable with a dropout.
(And You need the "U" term for the true speed versus the ground speed.)

But why don't you just use wheel speed?
Or plot the number of satellites, as that goes to zero (or usually 0,4,5...), never 1 or 2. So when the nSats <4 then make the speed invalid.

and to calculate time in HH:MM:SS from a value of seconds after midnight.

HH = ???(sec/3600)
The "???" is whatever gives you the whole number of the division fix, trim, truncate - something like that.
MM = ???(Sec - (HH*3600))
SS = sec - (HH*3600) - (MM*60)

or
SS = MODULO(seconds/60) , Assuming that there is a MODULO function... MODULO returns the remainder of Sec/60.
MM = MODULO((Seconds - SS),3600)
HH = (seconds - SS - (MM*60))/3600

[GaryA]

Thanks for the reply,

We don't measure wheel speed as we fit our test gear on different vehicles and have limited time to instrument the vehicle, so unfortunately we don't have access to the ECU (Although this should change soon, which will reduce our reliance on GPS data).

I do have 'satellites used' logged, but because there was a lot of tree cover (we were driving through tropical rain forest near Cairns) i think the signal was 'scattered' as the speed approaches 344Kph when i have 10 satellites and there are a few occasions where number of satellites is >5 and the data seems to be rubbish however there are also quite a few times the satellites used is around 4 and the data seems to be good (go figure), so i thought I'd use a different criteria i.e. rate of change of lat and long which seems to work well.


I will try and convert the time today and let you know how it goes.

I have got the time now, i used 4 maths channels,
HH = time in seconds/60
MM = HH-round_downHH/60
SS = MM-round_downMM/60

Time = round down(HH) * 100 + round down(MM) + SS/100
the output is HHMM.SS


Thanks again for your time!

[GaryA]

I have noticed another issue, when the GPS drops out the distance does not increase (as you would expect). One method of compensating for this was the idea of calculating the distance between GPS points. I think this would be fine if the GPS data did not drop out as i could use a time_shift function to calculate the change in GPS position
eg. delta lat = lat - time_shift(lat 0.05) the GPS is logged at 20Hz.

I have thought of an idea to fix this which would be to use the last valid value so that the equation would be:
Delta lat = lat - (last valid value of lat) - however it is the last bit i dod not know how to do.

Also i could make a rough approximation of speed if i could use the last valid value method and use that speed to calculate distance.

I was thinking i might need to add a 'latch' style function using the external maths function but have no idea on how to do it, is there any infomation on how to develop external maths functions?

Cheers
Gary

[JamieA]

As for your request to have google earth images overlaid in i2, we have discussed this, and it is on our request of 'stuff to do'. It is important, and will eventually be added, but just isnt quite there yet.

Keep an eye out on the i2 updates and we should have something in a while.

regards,

Jamie