Cyclist killed in Oakland

← Back to Forums


Ahlir
Participant
#

The school zone idea could be a good approach, but as @Vannevar points it would require executive action. It’s worth a try.

In the meanwhile, could we just, maybe, try to enforce the existing speed limit? All the legal detail is already worked out. Even the signs are up. All we need is for the city and the police to do their job. What am I missing here?


StuInMcCandless
Participant
#

That local law enforcement has to use 1937 technology to enforce speed limits, by state law.

The ironic side to this story is that someone looking out a window a few floors up can do a better job of determining vehicle speed than LEO at street level. Mr. Bauman’s tool shows every vehicle at every moment.

That’s why I think all the cops need to do is grab the plate numbers off a street level video camera. I’m sure there is plate recognition technology on the street somewhere. All we’d need to do is hook the two together, and have the police mail out warnings (at first), by the truckload. Mail it to everyone going 26+. Send 1,000 a day, if it’s automated. Make sure *everyone* knows about this.

Then, after a few weeks of this, have LEO actually employ that 1937 technology to cite what motorists they can, and haul some people into court. If certain people showed up on the warning list 15 times, and were on the high side of 40 for five of those, that should help the magistrate with his/her decision on issuing a fine.


reddan
Keymaster
#

In the meanwhile, could we just, maybe, try to enforce the existing speed limit? All the legal detail is already worked out. Even the signs are up. All we need is for the city and the police to do their job. What am I missing here?

Seriously. I’ve been wondering the same thing. Are there actual impediments (policy, conflicting state-versus-local laws, other) or simply a lack of will/manpower?

Who would be best to talk to to find out, in an authoritative manner, what obstacles need to be overcome? (Note that I don’t mean “what reasons can *we* come up with that might be true?”…while entertaining, such speculations do little to foster change.)

I know there’s pending legislation to allow local PA cops to use radar (House Bill 71/Senate Bill 535), but I’m not sure where that stands…seems to have been referred to the Transportation committee early this year, with no motion since.


MaryShaw
Member
#

@reddan and @Vannevar — thanks for bringing @Vannevar ‘s details into the thread. I searched pretty thoroughly on multiple sources, but this blog didn’t get swept into the net.

I found the K-12 trap: it’s in the definition of “school”, not the definition of “school zone”. PA Code section 212.1 (Definitions) says

School—A public, private or parochial facility for the education of students in grades kindergarten through 12.

School zone—A portion of a highway that at least partially abuts a school property or extends beyond the school property line that is used by students to walk to or from school or to or from a school bus pick-up or drop-off location at a school.

http://www.pacode.com/secure/data/067/chapter212/s212.1.html


MaryShaw
Member
#

There was a bike-ped meeting in Oakland on November 4 in response to the change.org petition. Lots of city folks gave updates on their plans, but there was a remarkable lack of willingness to take concrete action in the short term. The community police officer from the local zone was there (sorry, I didn’t get his name).

Speed enforcement was discussed. It was reported that there are quarterly speed enforcement blitzes with Pittsburgh City, Pitt, and CMU Police cooperating. Except that it has been several quarters since this was done. Here and at other meetings there has been real reluctance to do speed enforcement.

So yes, we should press hard on this. The next opportunity is tonight (Th 11/19) at the Oakland Green Team (i.e., bike/ped) meeting — 6:00 pm at the Oakland Career Center, 294 Semple St, 15213.

If we could get school zones, I’d put Schenley Drive in the vicinity of Tech St on the list.

As for the Bauman data, it’s great. Looks like it’s data on Forbes between the Cathedral and Schenley Plaza — cars coming from the stoplight at Bieglow headed for the stoplight at Dippy. They’re coming out of Oakland, where there is a stoplight almost every block. To get up above 35mph at the location Bauman studied, you have to be seriously lead-footed coming out of the Bigelow intersection. Are they accelerating trying to beat the light at Dippy? If so, let’s get that light re-timed so racing for it doesn’t work.


byogman
Member
#

I like the idea and it’s great if doable. I don’t know how light timing changes propagate. Let’s say they can’t change it because of XYZ change in the traffic pattern.

So, police can only be there so much, and can’t use reasonable technology to police speeds. What about stoplights?

The idea is this (I’ve proposed this once before, but I think it’s worth a repeat in this context).

When someone exceeds a threshold speed where they’re creating a hostile and dangerous environment, wherever we set the bar (I’d prefer lower, but the current 25mph speed limit would still be a great improvement), and have that AUTOMATICALLY trigger a move to the red phase, say, 3 seconds before they’d reach the intersection. Combine with red light cameras in case someone is feeling motivated to double down on stupid.

Have these signs throughout Oakland at these intersections making it very clear how it works. Make it so idiot motorists triggering the reds get an earful from OTHER MOTORISTS. If it works, roll out in other pedestrian/business heavy corridors.


MaryShaw
Member
#

@byogman I think the lights in Oakland are timed for about 22-23 mph, or at least that’s what it seemed to be when I was driving through regularly.

How many of the drivers know this, or even understand that driving a steady and correct pace through a series of timed lights gets you there just as quickly as racing for the next red light?

Would an education campaign on this topic help? Rational drivers would presumably adopt the speed that’s designed into the timing. But what fraction of the drivers are rational and what fraction just want to beat out any car they can pass in some sort of competitive frenzy?

If they’re mostly rational, the less drastic measure of providing information would help. If they’re mostly irrational and competitive, it would take your approach.

A great place for this would be Schenley Drive — put a detector at the top and a stoplight at the bottom that forces speeders to wait until 25mph would have brought them to the light, with a sign saying so. This doesn’t require any speed measurement — just note when cars enter Schenley Drive and set the light to turn green an appropriate number of seconds later. Schenley Drive traffic might be a little to heavy for this to work, but it might.


mbauman
Member
#

My impression is that the light at Schenley Drive Extension (by Dippy) is the last of the timed lights. Bellefield is almost always green to go straight (unless a Ped hits the walk button; not sure if that’s synchronized). And then Craig is definitely not synchronized. Just changing that would probably help out.


StuInMcCandless
Participant
#

I can’t make that meeting tonight. Can someone please bring up my idea?


byogman
Member
#

Matt, I share your impression that those two lights are problematic. I can feel my stress level rise on Forbes from the point I pass Schenley Plaza.

Unfortunately I doubt think fixing the timing is enough. Mary, I think fact that driver’s don’t “get it” now, is a pretty clear sign that they won’t “get it” in the future because of some nice signs. It’s just easier as a semi-unconscious driver (and let’s be honest here that’s what most are) to pay no particular attention to speed and then just slow down when there’s a red staring them in the face than it is to hold the line on speed. The only way you get the latter in practice is if the next light is pretty close and the driver toward the front of the green wave or if they’re just stuck behind somebody.

So we do need to wake them up, and Stu, I think your idea has merit, but I don’t think it’s as practicable as penalizing with lights since it requires manual intervention and putting together information from two perspectives. If they’ve delayed several quarters doing the routine stuff when is THIS going to happen? They’ll definitely fall asleep in the interim.

I’ll also admit that I’m perhaps a bit in love with the idea of other motorists cursing out those who speed badly the way those who speed badly curse out everyone else, and perhaps giving them an earful at the light. But the bigger factor is that it works directly against the motivation for speeding, immediately, and is always on.

Unfortunately, I can’t be at the meeting to voice support for either, but good luck to everyone!


alleghenian
Member
#

In England, they simply put a big sign on many main downtown roads that says “speed camera” along with a camera icon and the speed limit. They don’t try to be sneaky and “catch” the speeders with secret cameras. Everyone goes slow in these business districts. Even someone who has never been to England before knows exactly what it means. Instant behavior changer. No police necessary. No ACTUAL CAMERAS are even necessary really.


reddan
Keymaster
#

I’ll also admit that I’m perhaps a bit in love with the idea of other motorists cursing out those who speed badly the way those who speed badly curse out everyone else, and perhaps giving them an earful at the light.

Unfortunately, motorists already curse each other out and yell at one another at lights, and it has no perceptible effect on behavior (other than to jack up tempers further). Trying to encourage drivers to rage more, even as an attempt to encourage peer-group shaming for purposes of behavioral modification, is not something I find appealing.

That said, I really do like the idea of some form of direct, personal, and immediate consequences of speeding.


jonawebb
Participant
#

I’m doubtful that changes to the light timing will happen without careful engineering study, which costs $$$. My friend, the traffic engineer, spends a fair bit of time doing traffic simulations of Oakland (compute intensive, BTW), as part of his company’s BRT work. I suspect the city will insist that this sort of thing is necessary for any change to the Forbes / Fifth corridor, including, alas, changes to the striping, narrowing the lanes, etc. Unless there’s a way to persuade CMU and Pitt to care about the people who work and study there, and to invest some money in their safety, I don’t think it can happen without Federal funding.
Enforcement seems like an easier get, but it’s also unlikely to do more than shut us up. After it’s over, of course people go back to treating the wide open spaces like invitations to speed, as they do everywhere else.


byogman
Member
#

It’s not a bad thing that light timing changes require study. I will stake the tiniest bit of hope in the idea that my red light as punishment ~might~ not require quite the same degree of scrutiny since there’s good reason to believe that those conditions will start rare and become vastly rarer once people “get it” and the overall effect on traffic flow should be to make it much smoother.


mbauman
Member
#

I’ll drop by before the Oakland Green Team meeting tonight. It’s unfortunate that it coincides with the OpenPittsburgh/Students for Urban Data Science meeting (6:30 at CMU); I had promised to give a short intro to the project there… and I’m hoping to find folks who will help me.


reddan
Keymaster
#

Quoth @byogman:

I will stake the tiniest bit of hope in the idea that my red light as punishment ~might~ not require quite the same degree of scrutiny

I would venture to guess that it will require more scrutiny than simply changing the timing. Changing the timing of lights has plenty of historical precedent, and, more importantly, hard data regarding results.

However, it’d be really easy to implement on a temporary basis, with no systemic changes needed. Station cops with a radar gun (not for purposes of enforcement, but just to monitor speeds), and give them override control of the lights in question. Whenever they see a vehicle traveling over X speed, trip the yellow-to-red cycle.
Putting up a few prominent “new traffic patterns ahead’ signs would probably be a good idea, though.


byogman
Member
#

The unknown always triggers suspicion so you’re doubtless right, but devil’s advocate question. Why would stationing cops to do this not require a traffic study but programming in the behavior require it? It’s not like it would be hard to turn off the behavior. The only difference is the degree to which it relies on expensive and generally more fallible humans.


reddan
Keymaster
#

The only difference is the degree to which it relies on expensive and generally more fallible humans.

The only difference is how quickly you could test your hypothesis in real-world conditions. Using existing resources is generally faster than building new systems.

My point re: utilizing LEOs was not to discourage flow analysis; it was that an initial manual implementation for real-world testing *after* initial analysis would be fairly trivial, compared to implementing a new automated traffic control system.

Once the results from the real-world test have validated the analysis, then it’s time to talk about automation.


AtLeastMyKidsLoveMe
Participant
#

@byogman: your scheme is novel, but also problematic. As already noted, people in cars already have enough anger toward each other. But more importantly, enforcing laws is the job of the police, not me in my car… in 2015… when everyone is gun-happy… and crazy… etc., etc.

Inciting driver-on-driver violence is no solution.

(Although it just occurred to me that with your idea, being so hyperbolic, you were possibly trolling. If so, please ignore my comments.)


Ahlir
Participant
#

On reflection…

Speed limits should be enforced. But in the end punishment is not the right way to change human behavior. The world needs to be arranged so that the right behaviors happen simply because the environment make it the rational thing to do. Otherwise people simply focus on beating the system (I would).

Unfortunately people have to notice the contingencies. The lights on Forbes can in fact be traversed without stopping; I agree it’s 20-25mph (and btw Liberty westbound is ~25-30mph). But there’s nothing there to tell you that.

So, actually, the stuff that counts as “calming” is probably the better solution. I could go for tables. Robotic machine guns on the Cathedral would be a bit much. Well, ok, maybe for cars going >40mph…


courtney
Participant
#

Just wanted to respond to some comments about the driver in this crash. They are being charged criminally. 4 counts.


sparkles
Member
#

courtney do you have a link to an article?


courtney
Participant
#

No. Susan lived with me. I was told by her family earlier this week. I don’t know all four of the charges, but one was vehicular manslaughter.

I hope that is helpful to the discussion.


J Z
Participant
#

@courtney thank you for posting what you know. I hope that the individual who was responsible for this is held accountable for their actions. condolences to you, I’m sorry.


jonawebb
Participant
#

Wow. This is the most remarkable thing I’ve read on this message board.


sparkles
Member
#

@courtney I’m so sorry. Thanks for letting us know.


RustyRed
Member
#

courtney, thanks so much for updating us. I’m so sorry for your loss. I hope some real change will occur as a result of this senseless act.


paulheckbert
Keymaster
#

Essential Pittsburgh radio program discussed Matt Bauman’s analysis of traffic speed on Forbes Ave, relevant to Susan Hicks’ death. 10 min. http://wesa.fm/post/using-informal-look-traffic-speed-oakland-starts-conversation


Ahlir
Participant
#

Observations about the traffic lights in that area (made while I’m stopped on a cross-street):

At some point in the cycle the Forbes lights at Bigelow, Schenley Park Ext, Bellefield and maybe Craig are all green. For a noticeable amount of time.

I think it would be natural for a driver, approaching Bigelow from Bouquet, to see this as a straightaway and simply start speeding up. Maybe hoping (irrationally) to make it through all the lights in one shot. So we get speeding.

What if the cycle were altered so that cars see something different? (Yes, timing traffic cycles is its own field and it’s all more complicated. But maybe there’s a workable solution.)

I like that idea of sensors that trip when (any) car goes >25mph and cause the next lights to go red. This contingency could even be signed so people learn what they should do to avoid it.

Maybe there’s a way to fix that stretch without having to build stuff.

[yes, I’m waiting for dinner to cook.]


Benzo
Participant
#

I love the solution to mitigate speeding by forcing a light change when speeding is detected.

Put up lots of signs indicating this is what happens. People will stop wanting to speed.

If enough people decide they don’t want to speed, it will slow down the ones who want to by reducing their opportunities to speed.

Seems like you could add that as a workable parameter on top of the Surtac system.


Drewbacca
Participant
#

I’m all for some sort of active sensor… the light timing, as is, doesn’t work. It may work in an ideal scenario in which only one car is involved, but when you have a line and it takes a minute before everyone is going when the light changes from red to green, then the lights fall out of sync for a bunch of drivers. Those drivers in turn get frustrated and act aggressively.

I don’t think it’s a need to see the road as a highway when a driver sees all the green lights, so much as a desire to get through the lights before they change… again, and again, and again. It’s conditioning, in my opinion.

Whatever approach is taken, it needs to be active instead of passive. It needs to be able to correct itself for variations in traffic patterns or else good intentions can easily turn into unintended responses. Traffic doesn’t tend to respond as expected… so it would take a lot of tinkering to get right.

I really like the idea of lights all turning red when an excessive speeder is detected. It’s certainly a worthwhile idea, to a point. If only we had a world class university or two near by that could put students to work on such a scheme.


edmonds59
Participant
#

I have a rant stewing about the whole Surtrac situation, but this is probably not the appropriate thread for it.


Mick
Participant
#

Timed lights – if you go through a light that just turned red, you can speed for quite a bit before you start hitting red lights.

Maybe we should consider paying people to enforce the law? Wouldn’t that be unusual?


paulheckbert
Keymaster
#

But Pittsburgh is short on traffic cops (they’re understaffed), and labor is expensive. As with many things, automation is the answer. At least an important part of the answer.


StuInMcCandless
Participant
#

Ultimately we want to get to where enforcement is not necessary because street design causes people (in cars, on bikes, on foot) to do the right thing, so that everyone is safe by default. We aren’t there yet.

What we need first is some sort of traffic calming measures so last-second veers into cars in adjacent lanes don’t kill people. As just stated, we don’t have the manpower to place cops at every corner, and even if we did, we’re just plugging up the magisterial system with time consuming paperwork. I don’t think strict enforcement is a viable option in the short term, either.

Traffic lights cost a lot of money, easily a quarter million dollars. Retiming them runs into tens of thousands of dollars and months of time at the least, in figuring out what to do, getting the necessary clearances, and actually making it happen.

If we want real results right away, we need to think outside the box. Implement my suggestion. I bet with the right help from the police, Mr. Bauman, and other bright minds, we could have a functional system in place in a matter of weeks.


byogman
Member
#

Apologies, but I have to rant for a sec. Why is it an article of faith that traffic lights, pedestrian signals, etc… are a mind boggling expense? I swear the more I hear that the stupider and more pathetic it sounds. A shielded system for switching between lights is trivially easy and nowadays programmability in a tiny form factor with tiny power consumption, using standard interfaces, and in standard programming languages is also dirt cheap. What is THE REAL PROBLEM here and can we kill it and move onto more interesting problems with a better toolset already?


reddan
Keymaster
#

No need for faith…just read pricetags. One’s personal definition of mind-boggling may vary, of course, but hard numbers (such as “$55K-$75K for a new pedestrian signal at a crosswalk” are easy to come by.

See, for example, http://www.itscosts.its.dot.gov/its/benecost.nsf/SingleCostTax?OpenForm&Query=Arterial%20Management for various cost summaries. Interesting stuff, such as the cost of retiming averaging about $3000/intersection, or the cost of adding adaptive signal control being between $20K and $80K per intersection. (Adaptive signal control, wherein the timing of the lights changes dynamically based on demand, would seem pertinent to the discussion here…as, effectively, what’s being discussed is adding speed sensors and some logic to an adaptive signaling network.)

I can’t speak to what portion of those expenses is warranted, versus pork, bloat, kickbacks, etc.

[ETA:] Also, costs for a new signal, rather than updating an existing one, range from $250K-$500K, per various sources (such as the Washington state DOT).


byogman
Member
#

There are definitely plausible looking starting points at alibaba.com in the low three digits not the low to mid 5 digits. Doubtless minus programmability, but that’s so, so trivial now. Even if you have to hack it apart, spice a couple wires, maybe even run an extra wire between signals, for a quarter mill a pop? Stop making excuses and git r done. I mean, we’re bringing the control software responsibility to CMU and SurTrac anyway. Why the hell couldn’t we ask it to be built to a standard interface do with commodity hardware?


Ahlir
Participant
#


cost of retiming averaging about $3000/intersection, or the cost of adding adaptive signal control being between $20K and $80K per intersection.

Well, gosh; sounds like a lot.
But maybe we should consider some context.

What’s the cost, per-incident, of services dealing with each dead (or dying, or crippled) pedestrian or cyclist? Don’t forget to include the victim’s lost productivity. I’ll give you a pass on the pain-and-suffering; but feel free to add it in.


reddan
Keymaster
#

@ahlir: personally, I think a couple of grand to retime an intersection sounds like a frickin’ bargain. Retiming the entirety of 5th Ave in Oakland for what, $30K? That’s worth it just to TRY a new timing sequence, even if it turns out to suck and we need to revert.

As far as an adaptive signal system goes, I don’t know enough about them to have an opinion as to their value. But, if they can be demonstrated to improve matters, I’d advocate for ponying up the cash where it makes sense.

@byogman: I don’t doubt that there are ways to economize. My point in introducing real numbers is to encourage people not to assume, as an article of faith, that change is simple and easy until proven otherwise. Without understanding the existing costs, it’s not realistic to hand-wave them away.

← Back to Forums

You must be logged in to reply to this topic. Click here to login.

Supported by