Author Topic: Observation  (Read 3731 times)

Offline ChrisTheC

  • Full Member
  • ***
  • Posts: 132
  • Karma: +3/-1
Observation
« on: December 11, 2013, 07:06:55 am »
Now that this groovy plugin has it's own forum area, I thought I would post an observation in it's own topic (because I can ;D )

First, let me say, the draggable home location is pretty handy. It allows us to more finely pinpoint the target location.

I've been using the later versions (currently v1.69d) of iPhone Locator that allow distance calculations for driving, walking, etc. One thing I noticed (this is not a problem with the plugin, but with Google Maps IMO)

When I switch to the Distance Mode: driving, I get many more Present/Away notifications. Over a period of one hour while I was sleeping, I left home, & arrived home 5 times last night.

My Home Range setting is at 0.1. Obviously, all I have to do is increase my Home Range setting to compensate for GPS/WiFi location drift. Or, maybe I'll switch the setting back to direct and lose the superior distance & eta calculations (which I haven't figured out a good use for yet anyway) That's an internal question for myself to answer.

Anyway, thanks again for a great plugin

Offline amg0

  • Moderator
  • Sr. Hero Member
  • *****
  • Posts: 2949
  • Karma: +192/-8
Re: Observation
« Reply #1 on: December 11, 2013, 07:53:17 am »
same for me , I think 'driving' mode induces drastic changes on "distance" reported with only a small gps position change. probably wise, when working with driving mode to increase home range distance, while keeping it small for direct mode. I think I personally prefer direct mode. Not to mention that the Range Circle displays on the map, make no sense in driving mode for instance

Offline duiffie

  • Full Member
  • ***
  • Posts: 126
  • Karma: +2/-0
Re: Observation
« Reply #2 on: December 11, 2013, 10:58:15 am »
Another option may be to use PLEG to automatically set the mode to walking while you're very close to home, you could use the Distance variable to trigger such an action.
Vera2 (1.5.622)

Offline mikee123

  • Hero Member
  • *****
  • Posts: 1521
  • Karma: +18/-11
Re: Observation
« Reply #3 on: December 11, 2013, 11:46:04 am »
I find my distance variable is very unreliable. The map shows me as 10 miles away, but when i check on my PLEG distance is being reported as 0.006, i am not sure why this is the case. My polling maps is getting the right distance, as the polling interval is right for the distance. Just when i get a status from PLEG, the distance sometimes takes a long time before being updated. Or so it seems.
That is one reason i try not to rely too much on it in my logic.

Offline Ryanoc75

  • Jr. Member
  • **
  • Posts: 57
  • Karma: +0/-0
Re: Observation
« Reply #4 on: December 13, 2013, 03:16:28 pm »
One other small issue I've noticed in direct mode is that speed calculations are very erratic.  I was watching the dashboard as my mother was driving home yesterday to observe the triggers in action.  On one poll it reported her speed as 30km an hour while she was seated in a restaurant.  Next she was reported to be travelling at 17km per hour in an 80 zone (she would have been averaging 90 - 95)  The very next poll reported her travelling at 245km per hour!! She can be a bit of a lead foot at times but I am more than confident that she was NOT travelling that fast lol.  I still had it in direct mode at that point with a custom polling map.  I've since switched to driving mode with a custom poll but haven't been able to observe accuracy.
Vera3 2gig Thermostat, Kwickset door lock, DSC Alarm Panel wiznet - IT100, 7 DSC motion Sens 4 DSC 3way sen, 5 DSC door cont,  Aeon HEM pwr mon, Aeon 4in1 multi sens, 8 GE dimmers, 8 GE switches, 15 Intermatic plugs, GC100 Wifi, 2 F/F Sens, 5 HA07 RFremotes, Siri Homekit Bridge 10 Cams, 4 SmrtSwitch

Offline amg0

  • Moderator
  • Sr. Hero Member
  • *****
  • Posts: 2949
  • Karma: +192/-8
Re: Observation
« Reply #5 on: December 13, 2013, 06:03:45 pm »
One other small issue I've noticed in direct mode is that speed calculations are very erratic.  I was watching the dashboard as my mother was driving home yesterday to observe the triggers in action.  On one poll it reported her speed as 30km an hour while she was seated in a restaurant.  Next she was reported to be travelling at 17km per hour in an 80 zone (she would have been averaging 90 - 95)  The very next poll reported her travelling at 245km per hour!! She can be a bit of a lead foot at times but I am more than confident that she was NOT travelling that fast lol.  I still had it in direct mode at that point with a custom polling map.  I've since switched to driving mode with a custom poll but haven't been able to observe accuracy.
yes speed is very erratic because it is calculated like  diff ( distances ) / diff ( times ). but diff (times) is based on polling times from the VERA ( which we know & control ) , while diff (distances ) is made of position reported by iCloud. remember that we do not get position from iDevice directly but from iCloud, and we do not know exactly at what point of time the position was acquired by iCloud, therefore we are caculating speed based on a set of distance and a set of time which do not necessarly correspond well , hence this result.

Offline mikee123

  • Hero Member
  • *****
  • Posts: 1521
  • Karma: +18/-11
Re: Observation
« Reply #6 on: December 14, 2013, 05:29:01 am »
I discovered yesterday that i can use Homewave to monitor my iphonelocator and display some values. So while on the train home i was watching updates on distance, iconcode, last poll, speed and poll frequency (I am using custom poll maps depending on iconcode)
My observation on it. Distance updates a lot better than i thought. Iconcode stayed on 0 until i was nearly home (changed from 0 to 70 after 45 mins of travel). The last poll and polling frequency did not update for nearly 45 mins, although distance was updated. This could be down to Homewave ? Speed was very erratic. Stayed at 0 most of the time, then jumped to 176,then to another extremely low figure. Looking at previous thread, that ties in with these findings. I was also lookibng at my wifes iphone at the same time, she was at home not moving. Her last polling seemed to update ok.
This is just my observation, and the first time i was able to watch the iphonelocator while traveling. Very interesting.

Offline ChrisTheC

  • Full Member
  • ***
  • Posts: 132
  • Karma: +3/-1
Re: Observation
« Reply #7 on: December 14, 2013, 07:50:35 am »
I discovered yesterday that i can use Homewave to monitor my iphonelocator and display some values. So while on the train home i was watching updates on distance, iconcode, last poll, speed and poll frequency (I am using custom poll maps depending on iconcode)
My observation on it. Distance updates a lot better than i thought. Iconcode stayed on 0 until i was nearly home (changed from 0 to 70 after 45 mins of travel). The last poll and polling frequency did not update for nearly 45 mins, although distance was updated. This could be down to Homewave ? Speed was very erratic. Stayed at 0 most of the time, then jumped to 176,then to another extremely low figure. Looking at previous thread, that ties in with these findings. I was also lookibng at my wifes iphone at the same time, she was at home not moving. Her last polling seemed to update ok.
This is just my observation, and the first time i was able to watch the iphonelocator while traveling. Very interesting.

Interesting. Which device type do you create in HomeWave? Do you use "Multi value display"?

Thanks,
Chris

Offline amg0

  • Moderator
  • Sr. Hero Member
  • *****
  • Posts: 2949
  • Karma: +192/-8
Re: Observation
« Reply #8 on: December 14, 2013, 12:02:03 pm »
There is hope to get better accuracy of the actual "speed" of the device as the iCloud information returns a timestamp, but cf my question here http://forum.micasaverde.com/index.php/topic,18237.0.html , I am not a programmer of MAC/IOS and do not know how to decode that timestamp value.

EDIT: turns out it is pretty simple  timestamp/1000 is the same unix/vera timestamp so version 1.69f uses this timestamp and I hope we see improvement in speed accuracy ( therefore eta prediction in direct mode )
« Last Edit: December 15, 2013, 05:57:02 am by amg0 »

Offline mikee123

  • Hero Member
  • *****
  • Posts: 1521
  • Karma: +18/-11
Re: Observation
« Reply #9 on: December 15, 2013, 08:42:41 am »
Yes i used 2 multi value displays for one iphonelocator in Homewave. In each multi value you can spcify 3 different values to display, so i have 6 values from the iphonelocator displayed. Very useful to track and see what its doing. Just not sure yet if the updates on Homewave are quick enough, or are as quick as they would be if you look at the UI. But it helps me evaluating and making imore accurate for my specific needs

Offline ChrisTheC

  • Full Member
  • ***
  • Posts: 132
  • Karma: +3/-1
Re: Observation
« Reply #10 on: December 15, 2013, 12:30:16 pm »
Yes i used 2 multi value displays for one iphonelocator in Homewave. In each multi value you can spcify 3 different values to display, so i have 6 values from the iphonelocator displayed. Very useful to track and see what its doing. Just not sure yet if the updates on Homewave are quick enough, or are as quick as they would be if you look at the UI. But it helps me evaluating and making imore accurate for my specific needs

Thanks, think I'll fiddle with that myself.

C