Mats,

There is not any meeting tonight so you will not be missing anything.  I am still waiting for estimates of cost from Redflex and the police department for the 1 hour of video from the Allen and Lombard intersection. The police department is responding to your request for crash data so they will be contacting you separately.

 

Your request dated November 12, 2013 is actually for two pieces of information.  Each will take staff about 30 mins to provide at a cost of $35 apiece.  The first is traffic light timing which will contain information about how the various green phases change with time of day along with the yellow and red phases which do not change.  There are not any underlying calculations to support these values since they are based on engineering judgment and observations of signal efficiency over time.  Since all discussion to date has been about the duration of the yellow and all red phases, I am pretty sure that this is not what you want but let me know if I am incorrect.  Please also indicate your acceptance of the $35 cost.

 

What I assume you are looking for is the duration of the yellow phase of the signals in question and more importantly calculations justifying those values.  I cannot provide any calculations for the values since they are based on regional practice and consistence with ODOT published recommendations. 

 

Scaled drawings for the other three intersections that are similar to those provided at no cost for the Allen/Lombard intersection should be available.  The cost to get copies of all three is estimated to be $35 so let me know if this is agreeable.

 

peter

 

 

 

Peter and Bill,

 

You are lucky today because I will not make it to the council meeting this evening. I will respond to your below message next week. However, I'm attaching another public request to find out what you are basing your traffic light timings on in your four intersections with cameras. Please see attached request. In the meantime, please give me an update on the video request and the accident report documentation.

 

Thanks,

Mats Järlström
13520 SW Hart Rd.
Beaverton, OR 97008
USA
503-671-0312 Tel.
503-671-0454 Fax
mats@jarlstrom.com
www.jarlstrom.com

 

 

From: Peter Arellano [mailto:parellano@beavertonoregon.gov]
Sent: Thursday, November 07, 2013 6:19 PM
To: 'Mats Jarlstrom'
Cc: Dan Gill; Eric Oathes; Randy Ealy; cwillis@kgw.com; Jabra Khasho; Bill Kirby
Subject: RE: Allen and Lombard Intersection

Mats,

I have responded to a majority of your comments below.  I will not be able to complete the response until I hear back from ODOT and determine the city’s cost of complying with your request for the Redflex videos.

 

peter

 

 

 

Peter,

 

You were present at the City Council meeting last night so you already heard some of my responses to your comments below.

 

We have addressed the northbound SW Lombard yellow light timing error which you say you now have changed. As you saw in my October 8th report the northbound Yellow light timing did not meet neither ITE nor ODOT's values. You have also informed me that the timing of the Allan and Lombard traffic light controller has not been changed since the camera system was installed so the short northbound yellow light error has been in effect at least since 2001. At the meeting I submitted my first Public Request (see attachment) to investigate the 21 accidents with injuries in the SW Allen Blvd and SW Lombard Ave intersection to see if specifically the northbound short yellow light interval caused any of the accidents (see attached Beaverton 2013 report page 4). The City attorney and police department will be addressing this request.  My next question and request will be, who had the jurisdiction over this intersection during the 2001 to current time period, ODOT and/or Beaverton? The city has always had jurisdiction over this signal.

 

Regarding ODOT’s “Minimum Yellow Change Intervals” document, I’m certainly not confused. The ODOT document does not reference any errors or tolerances. In fact no ODOT document does this. Who did you talk with at ODOT? I have already informed you that Rebecca Thoreson at ODOT said that ODOT would re-time the traffic light controller so the visible yellow light would be the MINIMUM recommended time or more. I have contacted ODOT to get an explanation of the discrepancy in information that we are getting from them.  I will provide a more complete response once I hear from them.  Again, this is basic engineering. Why are you not willing to fix this error since it is a safety issue? And if your latest data after the lamp replacements is correct we are talking about 0.1 seconds of error correction to make it right.

Also, the ITE formula calculates 3.2 seconds for the Yellow change interval in the east and west direction of Allen and Lombard and 1.8 seconds for the All-Red clearance time (see Oct. 8th report attachment). On what do you base your decision to pick ODOT’s minimum 3.5 seconds and 0.5 seconds when it should be 3.5 seconds (ODOT minimum) for the Yellow and 1.8 seconds for the All-Red? It seems like you are just picking numbers out of thin air and hope for the best! Do I need to file a public request to see your calculations (if any) for this intersection too?  The decision to adhere to ODOT minimums based on the value of having regional consistency was made years ago by the city’s former transportation engineer, a decision that the current transportation engineer supports.  I have confidence that the regional discussion about this policy was based on many years of combined experience in traffic engineering and to insinuate that it was no more than picking numbers out of thin air and hoping for the best is insulting.

 

I reviewed your attached spread sheet data but I do not know how you got your numbers and what tools you used. The provided data looks strange since all recorded visible yellow and visible red timing data pairs end with the same decimal. I would like to do a public request to get copies of the two video’s you used so I can verify your data. Attached is this request.  The software I used was AVS Video Editor which was recommended by you.  You would have to contact a video expert or an AVS representative for the answer to that question but I believe that the answer may be sampling the video data on set intervals.  I will not be able to get you an estimate of our cost to supply the videos until at least Tuesday next week as I am out of the office tomorrow and our offices are closed on Monday.

Again, the citation grace period is to compensate for the errors in the overall traffic light and red light camera system. We already know that traffic lights have timing variations and errors but it is not safe and fair for drivers if some gets a longer yellow light and some a shorter. We also do not yet know the error in the Redflex camera system. The information you provided below says that Beaverton takes the data from Redflex Traffic System at face value with no checks and verifications. My questions to you is how do you know the timing data is accurate in the citation pictures? How do you know that the video is recorded and played back at correct speed and not tampered with? Do you know the overall tolerances Redflex have in their system? I have already pointed out that my wife’s citation speed has a 40% error compared to the Redflex video data.  The validity of the citation is based on the location of the car relative to the crosswalk stripe at the point that the light turns red.  If the car has not crossed the first crosswalk bar when the light turns red and then proceeds to cross through the intersection while the light is red, a citation is issued provided that the evidence supports the fact that a violation, as defined by state statute, has occurred.  Neither the speed of the vehicle or the speed of the video recording or playback play any role in determining if a violation has occurred.  I have already responded to your concerns regarding the vehicle speed as reported on your wife’s citation.

The grace period should be the sum of the timing variance of your traffic light timing errors and Redflex errors. Since we do not yet know the later I can not give an answer what the total number should be but your traffic light system is adding 0.2 seconds (3.52s-3.40s=0.12 seconds visible yellow light variance rounded up to the next tenth of a second) if your current data is correct. Doesn't the error/variance number of 0.12 seconds sound familiar? Yes, it is the same time stamp my wife triggered the Redflex camera system with! Maybe you are now starting to see WHY I'm going to every city council meeting to fix the red light citation process and your erroneous traffic lights!

You claim that "It is not possible for the Redflex equipment to report the length of the yellow light since the city does not share that signal with Redflex". Of course it is, you already do! One simple solution to add the yellow timing to the citation documentation is the Redflex video itself which you used to check the yellow light timing yourself… The yellow lights on-time NEEDS to be checked for every citation and if the visible yellow light is less than the minimum 3.50 seconds no citation should be issued. If you on the other hand fix the current timing error as I suggest so the visible yellow light will ALWAYS be more than the 3.5 seconds (E/W Allan and Lombard) you will have valid citations all the time. I'm not going away until you make this simple change especially in all your intersections with Redflex cameras.  One of the positive outcomes of this interaction with you is that we now know that we have the capability to periodically verify that the signal equipment is operating within the acceptable tolerances established by ITE.  While this process could prove to be a more cost effective tool to periodically ensure the integrity of our equipment, I do not think that the added cost associated with more accurately verifying the length of the visible yellow phase for every citation is a good expenditure of taxpayer resources.

 

Thanks,

Mats Järlström
13520 SW Hart Rd.
Beaverton, OR 97008
USA
503-671-0312 Tel.
503-671-0454 Fax
mats@jarlstrom.com
www.jarlstrom.com

    

 

From: Peter Arellano [mailto:parellano@beavertonoregon.gov]
Sent: Monday, November 04, 2013 7:54 PM
To: 'Mats Jarlstrom'
Cc: Bill Kirby; Dan Gill; Eric Oathes; Randy Ealy; cwillis@kgw.com; 'Joseph Rose'; Jabra Khasho
Subject: RE: Allen and Lombard Intersection

Mats,

One item that appears to be confusing you is that ODOT’s “Minimum Yellow Change Intervals” as shown in Table 1 of the attached document are the minimum values that they program into their controllers and are not guaranteed minimum durations of visible yellow light.  I have confirmed this fact with ODOT at both the Region 1 and State levels and have been told that this information has been shared with you.  This is the same policy that the City of Beaverton uses when programming its intersection controllers so variations in the duration of the yellow phase are expected provided that they are within the tolerances established for the signal equipment that is available.  There is nothing in the information that you have provided nor in the video that I have viewed from Redflex that is outside of the cumulative tolerances of the equipment.  This is the policy that the city intends to follow at all city maintained signals until such point there is a region wide change in policy or, in the opinion of the City Traffic Engineer, there are safety concerns that warrant longer yellow or red phases.

 

That said, when viewing video of the cameras at the Lombard and Allen, I noticed that there were short gaps between the visible green and yellow light which I attributed to delays in the yellow lamps becoming visible.  To address this issue, I had our crews replace the yellow lamps and then viewed video of the intersection after they were replaced.  I have attached the results of this change which covers a one hour period for both cameras located at this intersection.  As you can see from the data, the shortest duration of the yellow light was 3.4 seconds and the average duration of the light was 3.48 seconds.  This is compared to a minimum of 3.36 seconds and an average of 3.44 seconds before the lamps were replaced which addresses two thirds of the error in the average visible yellow light time.

 

You have asked about the city’s efforts quality control of the citation process.  The police department does verify that a violation occurred and the information presented to them from Redflex in still photos, vehicle make, driver photo, license plate, and violation are clear and accurate.  Once the photo red-light officer  verifies the information is accurate she watches the entire video of the light being run, to make sure the person did not come to a stop after hitting the inductive loops.  The city signal staff verifies that the program values in the signal controller match the values that the traffic engineer has posted in the controller box at least once a year.  These values are also verified when requested by a citizen that has received a citation.  The city does not inspect or maintain any of the Redflex equipment, just the end product that they provide.

 

I have added some responses in red below in an attempt to make sure that all your questions have been addressed.  Please let me know if I have missed anything.

 

peter

 

 

Peter,

 

Thanks for your response. Why don't you program the northbound SW Lombard Ave to 4.2 seconds to compensate for errors we have measured in the yellow phase actual outputs and at the same time adjust or program the remaining east, west and south yellow phase times to 3.7 seconds? If you and Beaverton will do this you will comply with the standards in our area which is ODOT's minimum recommended yellow change interval times and actually stay within the recommended range by using error correction. I address this in the first paragraph above.  This is the right thing to do and if you would like to make it perfect we should add the red light camera citation grace period of 0.5 seconds before a citation can be issues. As I have stated previously, I cannot recommend that the city not cite people who run red lights when they are in possession of evidence that a vehicle entered the intersection on red.  When all these steps are in place you will have elevated the red light camera intersection and system to become safe, fair and ethical. In my opinion we have reached this milestone since the only people who receive a citation are ones who clearly enter the intersection after the light turns red as shown on the still photo and video evidence but I respect that you disagree.  You can also present this updated information and compliance on your website and the public will respect the City of Beaverton and they way the red light cameras are used. And so will I.

 

Police Chief Spalding was present at the meeting in September and the Mayor should delegate the questions to the correct department so I expect a swift response to my questions below.

 

I do not know how Redflex is measuring the speed, if it is the loop detectors or if they have optical sensors or even use the video information. Regardless, my point is that 27 MPH is very low speed in a 30 MPH zone and the data that is provided by Redflex needs to be checked and verified by Beaverton before it is used in the Municipal court room. I do not even trust my own measurements before and I have verified my data several ways. The bottom line is I wish to know how Beaverton does this check or verification and not just take the data from Redflex for granted. This is the most important question I have asked the City of Beaverton when it comes to your relationship with Redflex Traffic Systems and your citation process.  We have been told by Redflex that the vehicle speed is measured by two inductive loops per lane in the pavement just outside of the intersection.  This speed is used to determine the likelihood that a vehicle will stop for the red light and to arm the camera if it is determined that the vehicle is not likely to stop.  The speed that is displayed on the citation photograph is the vehicle speed measured by these loops.

 

You provided the scaled drawing of the Allen and Lombard intersection and I don't know if you actually looked at page 2 of the "0542536_RedflexVehicleSpeedError.pdf" document where you see the 68 feet distance marked. FYI, I measured from the stop line to the first line of the next crosswalk, not between the two crosswalks. And I agree, my wife probably stepped on the gas through the intersection and I think we can also see at least one of her brake lights come on in the sunlit video after she has passed the second crosswalk. The speed is not used by Redflex, as I understand, but it is added as data in the citation pictures.  I addressed this in a previous email.

 

One last major missing piece of information is the yellow lights on-time in the citation process. This information should be documented with each citation but it is not. As you know this is the key critical information that show if the lights were working properly at the time of the citation. The red light is just a matter of fact - the yellow is the MOST important information we drivers have go by in the traffic light intersections.  It is not possible for the Redflex equipment to report the length of the yellow light since the city does not share that signal with Redflex.  Anyone who has been cited can request that the city verify the programming in the controller, something that we periodically do. 

 

Thanks again Peter for your response, I hope you will follow through and make Beaverton a great place to live, drive through and just visit.

 

Mats Järlström
13520 SW Hart Rd.
Beaverton, OR 97008
USA
503-671-0312 Tel.
503-671-0454 Fax
mats@jarlstrom.com
www.jarlstrom.com

 

  

 

From: Peter Arellano [mailto:parellano@beavertonoregon.gov]
Sent: Tuesday, October 15, 2013 3:45 PM
To: Mats Jarlstrom (mats@jarlstrom.com)
Cc: Bill Kirby; Dan Gill; Eric Oathes; Randy Ealy; cwillis@kgw.com; Joseph Rose
Subject: RE: Allen and Lombard Intersection

Mats,

My comments in red below.

 

peter

 

 

Peter,

 

I never responded via email to your response below so attached is my PDF response and also last weeks city council meeting presentation in PDF format. As you saw from my latest data, Beaverton needs to either reduce Northbound SW Lombard Ave approach speed into the SW Allen Blvd intersection to 30 MPH or re-time the yellow light phase to 4.0 seconds ASAP.  We have confirmed that the yellow phase of the northbound approach to the Lombard Ave. and Allen Blvd intersection is  currently programmed for 3.5 seconds which is 0.5 seconds less than the ODOT recommendation that we generally follow.  Since changing the speed limit is a timely process that ultimately has to be approved by ODOT, we have decided to increase the programmed value to 4.0 seconds so that it is consistent with this recommendation.  We are currently reviewing safety data at this location and may also do a speed study to determine whether a speed limit reduction on Lombard Ave. south of Allen Blvd. is appropriate.

 

I am also attaching my presentation material from September 10th which includes an important question that I'm still waiting for an answer on: I want to know how Beaverton know that the citation data and information is correct from Redflex Traffic Systems? Remember you are using this information as evidence in your Municipal Court.  The Beaverton Police Department manages the red light camera program so you will have to contact them to get information about the review that they do of the photos and video prior to the issuance of a citation.

 

Here is also an example of Redflex provided citation data - Citation vehicle speed 40% data error: I measured and calculated my wife's vehicle speed using her Redflex citation video and the intersection drawing of SW Allen Blvd and SW Lombard Ave you provided as seen in the attached "0542536_RedflexVehicleSpeedError.pdf" document.  I believe that Redflex uses a pair of induction loops to approximate vehicle speed before it enters the intersection.  They appear to use this information to determine if it is likely that a vehicle will run a red light so that the camera can take the shot.  I do not know if the accuracy of the displayed speed has any material effect on the validity of a citation.  That is a matter that you will have to take up with the court.

 

The method that you are using estimates vehicle speed through the intersection so I am not surprised that the numbers differ, especially if the vehicle accelerated to clear the intersection.  For the record, the actual distance between the crosswalk lines in question is 64 feet and I measure an elapsed time of 1.268 seconds resulting in a speed of 34.4 mph instead of 37.9.

I looks like there is a lot of work left to do since this is only one intersection that I have checked for accuracy and I have found so many errors...

 

See you tomorrow at 6:30 PM!

 

Thanks,

Mats Järlström
13520 SW Hart Rd.
Beaverton, OR 97008
USA
503-671-0312 Tel.
503-671-0454 Fax
mats@jarlstrom.com
www.jarlstrom.com

 

From: Peter Arellano [mailto:parellano@beavertonoregon.gov]
Sent: Tuesday, October 08, 2013 1:43 PM
To: Mats Jarlstrom (mats@jarlstrom.com)
Subject: FW: Allen and Lombard Intersection

Mats,

 

Please let this serve as my formal response to issues that you have raised regarding the duration of the yellow light phase for the signal located at the intersection of Lombard Ave. and Allen Blvd.  As you are aware, you provided video of this signal that captured a total of 9 cycles of the yellow light which varied in duration from 3.48 seconds to 3.36 seconds and averaged 3.41 seconds.  Since we last talked, I have secured and reviewed two hours of video at the same location from Redflex.  This video represents over 100 cycles of the same light with the duration of the visible yellow light varying from 3.36 second to 3.52 seconds and averaging 3.44 seconds (summary attached).  Using these values it appears that the actual visible yellow phase averages 0.06 seconds shorter than the programmed value and at most is 0.14 seconds shorter.  In all cases however, the yellow light phase is longer than the minimum (3.2 seconds) as described below.

 

The city programs the yellow phase time and the all red time in its controllers based on the values recommended by ODOT unless, in the opinion of the city’s traffic engineer, a longer period is warranted.  In the case of the east and west bound legs on Allen Blvd. these values are 3.5 seconds for the yellow phase and 0.5 seconds for the all red phase since Allen Blvd. is posted for 3o mph and there have not been any safety issues that warrant larger values.  This value for the yellow phase is greater than the 3.2 seconds derived from the formula recommended for use by the ITE (Institute of Transportation Engineers) and is considered to be a minimum by the city.

 

In response to your inquiries I have talked with technical staff at the city and at ODOT in an attempt to better understand the operation of signal equipment in Oregon.  I have learned that the acceptable tolerance of actual phase time for the combination of new controllers and load switches is +/- 0.1 second compared to the programmed value, a tolerance that meets ITE specifications.  This tolerance is not the result of inaccuracies in the controller’s internal clock as I originally reported, but is instead an acceptable overall tolerance in the combined controller, software and load switches when bench tested while new.

 

This tolerance does not take into account the acceptable tolerances for new red, yellow and green lamps that are mounted in the signal heads.  Per ITE, the lights must reach 90% of full illumination within 0.075 seconds of the application of power from the controller and must fade to black just as fast.  In the worst case scenario, if the green light went black instantaneously and the yellow light took the full 0.075 seconds to illuminate, and the controller only provided 3.4 seconds of yellow phase, the resultant visible yellow phase could be as short as 3.25 seconds.  Obviously this is an unlikely event, but it does demonstrate the limitations of the equipment that is available and should provide insight as to why the city does not program its controllers to the minimum value recommended by the ITE.  One item that I want to stress is that all the preceding discussion applies to all intersections in the city independent of the existence red light cameras. 

 

While it is true that the duration of the yellow phase measured by yourself and Redflex is within equipment tolerances, greater than the minimum values recommended by the ITE, and most importantly, adequate in the opinion of the city’s traffic engineer who is empowered by the ITE to use professional judgment, I do believe that there is room for improvement.  As you have previously noted, there is often a delay between the green light going dark and the yellow light becoming visible.  This may be indicative of the capacitor in the yellow lamp drying out so that there is less charge available to illuminate the yellow lamp quickly.  Based on my observations of the video, replacing the yellow lamps could result in an increase of between  0.03 and 0.06 seconds in the duration of the visible  yellow light so this is something we plan to do in the near future.

 

Another issue that has come to light as a result of your observations is the information that the city provides about signal timing on the Photo Red Light Enforcement FAQ’s web page.  As you have pointed out, the information on the web page reports the programmed values for the yellow light duration which could be interpreted as guaranteed minimums.  I will be recommending revised language on this page that will more clearly inform the public about the variance between programmed time and actual time.

 

Finally, you have requested that the city consider following what appears to be Virginia state law by providing a 0.5 second grace period and not ticketing individuals that have entered the intersection less than half a second after the light turns red.  If the city were to make this adjustment, cars could lawfully enter the intersection the moment that the opposing lights turned green since the all red phase is 0.5 seconds throughout the region.  In order to adjust for this the city would have to increase the all red phase at these four signals to give these cars time to clear the intersection which would make these signals operate fundamentally different than all others in the region.  This is contrary to the city’s goal to operate these signals exactly like the others in the city and throughout the region.  Adding this grace period would also put the city in the position of ignoring evidence of unsafe behavior that could jeopardize the safety of the public.  This is not consistent with the underlying goal of the red light camera program which is to improve intersection safety.  For these reasons I am not willing to make this recommendation to the mayor or council.

 

I want to thank you for bringing these concerns to our attention.  Ultimately you have given the city a way to more easily monitor the performance of the signals that are located at the red light camera intersections as well as given us valuable feedback on the information that we are providing the public.  Please feel free to contact me should you have any questions.

 

Sincerely,

 

Peter Arellano

Public Works Director