Filtering Points

Questions and discussions about operating Tru-Traffic

Moderator: bullock

Filtering Points

Postby Kyle.Williams.KHA » Thu Jun 23, 2011 9:24 pm

I am in the process of manually filtering points for my travel time studies. I have a couple questions regarding the points highlighted in the attached image.

1) Why is there such a drastic change in time between collecting points?
2) Clearly this is a full stop. Does the GPS stop collecting points while stopped?
3) If I remove these points, how does Tru-Traffic treat the missing time and am I eliminating a stopped condition from the analysis?
4) In this case, will my stop be from 54 seconds to 64 seconds (below 5 mph) and does the time in between 54 to 64 seconds and/or eliminated points not matter?

Thanks for the help.
Attachments
Screenshot.jpg
Screenshot.jpg (74.24 KiB) Viewed 5275 times
Kyle.Williams.KHA
 
Posts: 1
Joined: Thu Jun 23, 2011 9:12 pm

Re: Filtering Points

Postby bullock » Thu Jun 23, 2011 11:20 pm

Excellent questions, Kyle.

  1. In recording trip logs, Tru-Traffic ignores essentially redundant points. When there are three consecutive points at approximately the same location and with speed practically zero, Tru-Traffic discards the middle point, keeping only the outer two. As a result, the points at a full stop will often have gaps of > 1 sec. These have no effect on the analysis. Keep in mind that each point has a time-stamp associated with it, so the overall time does not depend on the number points, but rather, the difference between the time-stamps of the first and last points. The first and last points of the stop are retained, and possibly a few in between, but many of the interim points are discarded as they add nothing new, and the duration of the stop is determined by the time difference between the first and last points only.
  2. Tru-Traffic continues collecting points, but when it determines that a newly collected point obviates a previously collected point, it discards the previous one, replacing it with the new one.
  3. If you delete the points, the condition of a stop and the stopped time both remain the same (there is no "missing" time, as the duration of the stop is still the time difference between the first and last points). By default, a "Stop" is counted when the speed falls below 5 mph after having exceeded 15 mph, and these three highlighted points are all below the 15 mph threshold, so they don't affect the tally of stops. What you'd be deleting are a few interim locations showing some negligible details of a random GPS noise-drift while stopped. In this case, the only thing this would affect -- and only slightly -- is the overall travel distance, since that noise-drift adds about 20 - 30 feet to that travel distance, and changing the path of the drift to a straight line (by removing those points) would probably make a difference of only a few inches. This figure below illustrates this.
  4. I think you're confusing the Point # column and the Time (sec) column. Your stop is from roughly the 56s point until roughly the 89s point. (Tru-Traffic linearly interpolates to the near neighboring points to better estimate the exact time the trip log crosses the 5mph & 15mph thresholds). Removing the essentially redundant points at 65s, 72s, and 80s has no effect on this stop, which is still from the ~56s to ~89s markers, or roughly 33s in duration without or without the intermediate markers.

In this case, I think it's harmless to leave the intermediate points in place. In cases where the noise-drift follows a more erratic path, adding appreciably to the overall travel distance, I recommend filtering out the noisy points. But in your case, the noise within the drift is probably adding no more than 2 inches to the (unavoidable) 20 - 30 feet of the overall drift, since the noise-drift path is already pretty close to a straight line.

If this isn't clear, or if you have further questions about this, please let me know.

Regards.
Greg
Attachments
ScreenShotwithComments.png
ScreenShotwithComments.png (377.48 KiB) Viewed 5272 times
bullock
Site Admin
 
Posts: 218
Joined: Thu May 06, 2004 6:51 pm
Location: Pacific Grove, CA


Return to Program Operation

Who is online

Users browsing this forum: No registered users and 5 guests

cron