off-topic : Traffic signal wait time in Bangalore

[Opendro]

Advanced apologies for posting an off-topic. This is the only forum where I can reach more than two people ;-)

I wrote a complaint/suggestion to http://www.bangaloretrafficpolice.gov.in/Complaints.htm. I wrote a similar feedback in fb when I was in fb and they switched off the traffic signal in some areas which made the movements very smooth. Now, they have made long signal waits everywhere. This is mindless and Tugluk management. Could you guys please write to the following sugegstion/feedback in fb as well as the website? Thanks.

I have been observing that the traffic signal wait times had been double since around June 2016. This is not doing any favor to anyone. While it is true that it is efficient to clear one side when the traffic momentum is there, it makes commute time longer for the commuters as they will eventually wait too long at some or other junction.

Take for instance, the signal near Bangalore International Academy 7th block Jayanagar. There will not be any traffic accumulating there without any signal. You have switched of the signal sometimes and you would have seen that this junction does not need signal at all. Even if you put a signal, this junction does not need more than 15 seconds green time for each side. The sooner it clears off, the lesser crowding around the school. Instead, you have made the signal too long making people wait or jump the red in every morning school/office time. Please do some research before increasing the wait time. Even if you increase, please increase smaller amount of time and observe if it is helping or not and please find an optimum time.

Another problem with long waiting big junction is that they will suddenly release huge number of vehicles. That puts pressure on the next unsignaled junction. So, it is best if the traffic movements are released in bits.

I go to office by bicycle. But I do drop my child to school in motorcycle. Please help me and everyone abide by the law making the system more efficient.

Another suggestion is that Bangalore traffic police should spread some awareness about "right of way" instead of spending too much money of drunken driving awareness, which people already know. Right of way is needed because we cannot install signal in every junction and no one in India even seems to know about it. This is a very important awareness.

[Kartik Rustagi]

I have my self thought about this so many times and agree that shorter wait time / quicker switch is generally more efficient. This is based on just observation and not actual statistics. I see the same concern when a traffic signal is turned into a manned one where typically they apply the same strategy of clearing one side and then switching. This again, IMO, detiorates the overall situation.

Will file the suggestion!

[Karthick Gururaj]

I disagree on this point. IMHO, the efficiency of the traffic signal - in terms of number of vehicles crossing (regardless of the direction) per sec - will improve with longer wait times. This is because there is a cost to "context switch" flow of traffic from one direction to other. Many vehicles take time to accelerate from stand-still to their "cruise" speed - so the average speed of such vehicles while crossing the junction is lesser (than if they are just zooming by). And for safe crossing, we need a buffer of few secs after the traffic from one direction stops, to when the other direction can start. Slower speed plus time lost due to safety buffer implies lesser efficiency.

When a junction becomes a bottleneck, it makes sense to tradeoff increased wait times to gain higher efficiency of traffic flow via the junction.

Of course, for a specific commuter, it can be very frustrating (unless you are "zooming" by).

- Karthick

On Tue, Jan 10, 2017 at 5:09 PM, Kartik Rustagi <kartik...@gmail.com> wrote:

I have my self thought about this so many times and agree that shorter wait time / quicker switch is generally more efficient. This is based on just observation and not actual statistics. I see the same concern when a traffic signal is turned into a manned one where typically they apply the same strategy of clearing one side and then switching. This again, IMO, detiorates the overall situation.

Will file the suggestion!

--
biking conversations on the world famous "Bangalore Bikers Club" :)

are you a part of the bicycle racing scene?
Visit www.bangalorebicyclechampionships.com for more details
---
You received this message because you are subscribed to the Google Groups "Bangalore Bikers Club" group.
To unsubscribe from this group and stop receiving emails from it, send an email to bangalore-bikers+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Prashanth Chengi]

I think arbitrarily fixed time limits are the problem; there really has to be real-time capture of traffic density, and adjustment of the duration accordingly.  I've seen lights changing, even when there is a sufficiently large line of vehicles backed up, for no real reason, even when other lanes are relatively empty. On top of this, there is the switching overhead, which causes snarls to get worse.  

/Prashanth

[Naresh V.]

I think BEL (or some other non-privatised company in India) already has traffic light systems that can sense how many vehicles are backed up in a particular road and adjust accordingly. I saw this near the Police grounds near Garuda mall during one of the Bangalore Traffic Police week exhibitions in January three or four years ago.

Similar to how the Python 5000 pothole patching machine was brushed aside by the luddites (or opportunists) in the system, maybe these traffic lights never saw the light?

Also, what Karthick says makes sense. Think about the slow, small, heavy vehicles (TATA Ace, etc.) hogging up the right lane taking their own sweet time to catch up on the space in front of them and compare that to the "evil" autorickshaws and two wheelers and even smaller private cars and cabs trying to move into those gaps as if it was some kind of a tetris game and we can see why it's rather useful to have longer signal wait periods than shorter ones.

On Wednesday, 11 January 2017 12:17:51 UTC+1, Prashanth Chengi wrote:

I think arbitrarily fixed time limits are the problem; there really has to be real-time capture of traffic density, and adjustment of the duration accordingly.  I've seen lights changing, even when there is a sufficiently large line of vehicles backed up, for no real reason, even when other lanes are relatively empty. On top of this, there is the switching overhead, which causes snarls to get worse.  

/Prashanth

On Wed, Jan 11, 2017 at 11:54 AM, Karthick Gururaj <karthick...@gmail.com> wrote:

I disagree on this point. IMHO, the efficiency of the traffic signal - in terms of number of vehicles crossing (regardless of the direction) per sec - will improve with longer wait times. This is because there is a cost to "context switch" flow of traffic from one direction to other. Many vehicles take time to accelerate from stand-still to their "cruise" speed - so the average speed of such vehicles while crossing the junction is lesser (than if they are just zooming by). And for safe crossing, we need a buffer of few secs after the traffic from one direction stops, to when the other direction can start. Slower speed plus time lost due to safety buffer implies lesser efficiency.

When a junction becomes a bottleneck, it makes sense to tradeoff increased wait times to gain higher efficiency of traffic flow via the junction.

Of course, for a specific commuter, it can be very frustrating (unless you are "zooming" by).

- Karthick

On Tue, Jan 10, 2017 at 5:09 PM, Kartik Rustagi <kartik...@gmail.com> wrote:

I have my self thought about this so many times and agree that shorter wait time / quicker switch is generally more efficient. This is based on just observation and not actual statistics. I see the same concern when a traffic signal is turned into a manned one where typically they apply the same strategy of clearing one side and then switching. This again, IMO, detiorates the overall situation.

Will file the suggestion!

--
biking conversations on the world famous "Bangalore Bikers Club" :)

are you a part of the bicycle racing scene?
Visit www.bangalorebicyclechampionships.com for more details
---
You received this message because you are subscribed to the Google Groups "Bangalore Bikers Club" group.

To unsubscribe from this group and stop receiving emails from it, send an email to bangalore-bike...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--
biking conversations on the world famous "Bangalore Bikers Club" :)
 
are you a part of the bicycle racing scene?
Visit www.bangalorebicyclechampionships.com for more details
---
You received this message because you are subscribed to the Google Groups "Bangalore Bikers Club" group.

To unsubscribe from this group and stop receiving emails from it, send an email to bangalore-bike...@googlegroups.com.

[Opendro]

There are advantages of the context switch. The most significant one being a staggered flow thereby allowing unmanned/unsignalled junctions handle the right of way. And also being able to handle low traffic morning and night situations very well.

What is the benefit of cruising? Hardly any. It is like make everyone wait till the bus is full so that the bus does not have to pick/drop any more passengers on the way. But in reality, each passenger end up waiting longer as the bus will eventually wait too long before it could even move or in the next connecting bus. Similarly, those on cruising in one signal will eventually wait too long in the next signal. It is like disposing the garbage bin only when it is full. It saves the handlers time, not the travelers. If the road was only one way without any junction, it could have helped. But never in a network of roads.

Just to give an extreme example. We can open the signal from Madiwala checkpost to Hosur green for half an hour in one direction. We will have enough volume to fill for half an hour continuously. Will that help? It will help if everyone on this stretch was traveling to Hosur. But that is not the case.

If one side of the jn has more volume than others (road size being different), then the bigger road can have proportionately shorter wait time, but that is only relatively speaking. None of the other sides should wait for too long.

It is plain illogical and I have seen the results myself. Today, Bangalore International Academy signal was switched off again. It was off for two months when I complained last time. It never had jam there as vehicles came trickling and staggered. But the volume was same. Travel at any hour of the day and you will know what I mean. However, such unsignalled jn or short wait time jn will again fail  if there is a nearby jn with long wait time, because they will release vehicles suddenly creating an artificial high volume like in a dam.

- Karthick

To unsubscribe from this group and stop receiving emails from it, send an email to bangalore-bike...@googlegroups.com.

[Prashanth Chengi]

It's a classic scheduling problem, and as with all scheduling problems, you need to pay attention to the algorithm, to ensure there is no starvation (like the extreme half hour opening of one road that you mentioned). That said, I still feel that arbitrary times are silly. It definitely helps if successive signals can be synced with each other, but it should be a bonus factor used for optimization, and not the main driving logic behind the signal. Opening and closing of signals should indeed be traffic density based, which also makes your point about low traffic morning and night hours, moot. Signals should only change if there is a need. I've seen in parts of the US that lights don't change at all, if there is no traffic in any lane, i.e it stays green on the last used lane, till such a time when a vehicle approaches from any other lane. And when there are vehicles approaching from multiple lanes, it should surely be based on density of traffic, which in my opinion would also result in reduced pollution (fewer vehicles waiting for the light with engines running).

/Prashanth

To unsubscribe from this group and stop receiving emails from it, send an email to bangalore-bikers+unsubscribe@googlegroups.com.