DIYcity Challenge #3

According to this article, NY Waterway has been tracking their buses with GPS since 2003. We just need to persuade them to make the data available to the public. Then a little GPRS transceiver bolted to the bus stop pole...

In other news, a building on Roosevelt Island is trying to install a bus tracking system - but won't make the data available to bus passengers who don't live in that building, because their buildings won't contribute.

That NY1 article mentions that the MTA already installed tracking equipment on some buses:>> Under a contract awarded in 2005, tracking equipment was installed >> onboard 185 Manhattan buses
would there be any way to take advantage of that existing infrastructure?  Either with the permission of the MTA (difficult to get) or without?

While it's not full DIY, I think there is technology to handle this: there is cell phone coverage everywhere--in all those canyons, etc. Integrate a sensor into the bus and track it thru the cell system--no satellite drop-outs. Requires MTA and cell company to cooperate, though.

And a list of all buses and a budget to pay the old folks and a team to distribute it all...?

If so you could do it Monday morning.

You've got a bunch of people wanting to know when their bus will come. In the absence of sensors on the bus, you need a bunch of people telling you where the buses are. They could be the same people.

Barry Busrider wants to catch the uptown M15 at 42nd Street. He gets to the bus stop at 3:20, pulls out his phone and loads the Bustracker app. The app gets his location as 42nd and Third and tells him that he can get the M42 or M15. He chooses the M15 and it tells him that the next M15 is scheduled to arrive at approximately 3:25. Because the cloud is incomplete, it doesn't have any further information.

The bus shows up at 3:27 and Barry gets on. After paying for the ride, he pulls up the Bustracker app and hits the button that says "I'm on." When he gets off at 86th Street, he clicks the button that says "I'm off - and not transferring to another bus."

Shari Straphanger is waiting for the same bus at the corner of 72nd Street. She gets there at 3:35, loads the Bustracker app and tells it that she's waiting for the M15. It tells her that there's an M15 scheduled to arrive at 3:40, but thanks to Barry's information it knows that the bus is running late and has only gotten to 60th Street. It tells her that and estimates the arrival time at 3:44.

The bus comes at 3:44, Shari gets on and hits "I'm on." When she gets off at 96th Street she hits "I'm transferring to the M96."

This system doesn't require cooperation from the MTA, but it has lots of vulnerabilities. It gives an incentive for people to run the app, but none for them to turn it off. It also requires a certain number of people to have phones capable of running an app like this. Its ability to get location data is uncertain and possibly dependent on cell carrier cooperation; if that is not available, people have to put in their location data every block, which can get tedious after a while. I still think that bus sensors are a much easier way to go.

Being a newbie, I only saw/read the other solutions after I wrote my non-grassroots solution, so I now understand what you intended. The autonomous cell phone idea has some merit, but it is hard to see how it would work consistently and precisely on a bus-by-bus basis. You’d like multiple handsets per bus and you need to be sure which are on the bus and not in nearby cars. Once you involve humans saying where they are and which bus they are on, you introduce new forms of error and sources of unreliability, including potential data vandalism. Furthermore, you still need the mapping, analysis, and decision software plus all the signage, so you have only approached the location problem in a grassroots manner. I think you need to solve this with GPS that works in urban canyon.

I've been thinking about this a bit lately since reading a relatively recent thread in OpenStreetMap's lists:

How about making an iphone app where people can just type in "I just saw the 555 bus go past"? After a few samples you have a timetable.

It deals more with creating a timetable than with real-time data, but I think that if we enough data on, say, the subway, we actually could have a really reasonable timetable (which is better than what we have now). We might do this simply by having an app that has all of the stations or bus stops on the line you are on (starting with the one that it knows you are on since it may know your location as you step into the station), and you have a way of indicating that you have got to the next stop. In buses, this might be automated to some degree using GPS. On the train, it could be interesting if motion detection could be used to help some. Like I said, this could be used to create a more true-to-life timetable, which could give more accurate predictions of when the next bus/train will be arriving nearby. When someone is uploading very recent data, this could be used to improve predictions.

I know, often the trains are messed up thanks to service changes or holidays, and that there would be the requisite vandalism, but it's nothing that can't be overcome.

Unfortunately, there isn't much incentive for giving data. It is a bit closer to mapping with OpenStreetMap, where it takes a good deal of initiative to record the data and upload it. Maybe one way of getting people to use the app would be by offering a kind of web-311 where one could offer feedback on their current train or bus or stop. Maybe we could create emotional maps of our transit systems ("I'm always scared when I have to wait in this station for more then a few minutes." "The tilework in this station makes me happy."). Maybe we could create walls for parts the systems where people can freely communicate and leave notes for each other (and--who knows?--maybe it could be more than a location-aware Missed Connections).

Just a few thoughts. I know they're a bit late to the party, but I'm excited to find a group of people interested in something so similar to something I'm seriously considering implementing.

I understand that the canyon effect causes gps-tracked bus locations to get scrambled, but what about times when the buses aren't in areas with higher buildings - for example, most of Brooklyn and Queens? Could we tolerate a fuzzy information system - accurate location or estimated time of arrival when possible, something less accurate when not.

Right now, there's no info at all. It would be neat to know that my bus started its route on time and seemed to be running on time til downtown Brooklyn when the tracker lost contact. Maybe there's a correlation between building height and likelihood of delay. Having some information some of the time would be ok if we knew that the dead spots would be consistent.

I'm not convinced we need to make it as complicated as an app. But I'm just a plain old person and not a transportation strategist or a developer.

MY IDEA: This obviously only pertains to urban areas with decent coverage, but can't we just encourage existing Brightkite users and other folks with geo-locational phone apps to ping from their locations and then encourage the use of hashtags like #sfmuni47? The post would appear with location, min last posted and bus number. From here, if you're a bus 47 rider you just follow the feed: http://hashtags.org/tag/sfmuni47.atom

I know the hashtags feed works on following other events like conference comments, but it requires a fair amount of herding.

Dana
Villagers with Pitchforks

Ok there's few issues on the real time bus tracking that we should break down first of all:
-We need to find a technology to track bus in real time
-The technology must be as efficient and as cheap as possble
-The information needs to be available in a variaty of format: WEBSITE, MOBILE WEBSITE, VOCAL SERVER, TEXT MESSAGES
------------------------------------------------------------------------------------------
-We need to find a technology to track bus in real time

So far, the most common technology to track a moving object anywhere on earth is GPS.
We can find example form france to germany where buses are equiped with GPS and feed back their location to their central.
Browsing a little bit on the web, on you'll find out that the only way to track a car or a bus efficiently is GPS.
Alternatives method such as user input, are great but required user participation, which isn't realist as a system in itself, it is ony realist as a "sidekik", but this information cannot be operated effectively in a peer-to-peer mode, but in a centralised mode.
As far as I know GPS is the only maybe the cheapest way to track a vehicule.
The way GPS technology works, correct if I'm wrong, is that there's a transmitter in a devise that sends its data to a satallite which in turn pings back the latitude and longitude.
So a GPS devise in a bus will send get pings back from a satalite wiht its longitude/latitude.
Thats tells you teh position of the bus, but you need to enter then, the position of the bus stop as well, and you will need to give a unique ID to the bus stop as well as to the bus, so that both position can be mapped.
The last thing is to get the map of teh city patched on the GPS positioning.
that's how a car GPS works, and how it is telling you how far you are from your objective, and where you should go.
This compare to any other method requires no interaction from users, and little interaction from the bus driver.
------------------------------------------------------------------------------------------
The technology must be as efficient and as cheap as possble

- i think the technology is efficient enough since it seems to work pretty wel in others countries, but the cost is probably high in terms of software and equipement.
a) Every buses needs to be equiped with a GPS,
b) and will need to have a central server with a database to collect real time data and feed it back to whatever system
-------------------------------------------------------------------------------------------
The information needs to be available in a variaty of format: WEBSITE, MOBILE WEBSITE, VOCAL SERVER, TEXT MESSAGES

that's says it all, in france anyway, one needs to call a vocal server, tell the the stop id, and the server reply with the corresponding info, here the time the bus wil arrive.
The idea I think is to have a set of databases and a program which in simple terms will feed back the position of the bus and its approximated time of arrival to a variety of medium where the user in a way or another will query his Bus Stop ID through a vocal server, text message server, web site, or wap website.

One cool thing I think is to have a text messaging services, where you send your bus stop ID to a no, and you got a reply wiht the approximate time, with the option of having this text messsage send to you every morning for example, much like an airline remaiding you of you flight via text message.
So I think so far there's no beating around the bush to find out what technology to use, but rather finding the most efficient way to fund it, and how to implement it.Whihc is where crowdsourcing could be put to use.

So my questions is:
How sould we implement the system:
Should we for example have companies specialised in real time transit system to implement it,
or should it be built by freelances in a semi-open source manner at the risk of not succeding,
or should the Bus Transit authority/department hire professionel itself and implement the project itself.
-any other sugestions?

How should we fund the system:
Could we ask every resident of a city to wire monney to an account linked to the project, and select individuals to manage the funding?
Could we co-fund the project, by funding either the hardware or software?
-What incentives could we give to the people who fund the project?? Bus ticket reduction for a set period??reembursement of the sums given by everybody wiuht interest/ or no interest.
-set up a special stock of the public transportation and get people to buy shares to fund the project, with the ability to resale their shares??

This is exiting and I hope I'l get replies..
-

As far as I understand some of the issues that have stymied this system
in the past have been because there hasn't been enough fidelity with a
purely GPS based system in NYC. Some folks here at TOPP Labs have
recently built a proof of concept iphone app and map server interface to
show that it can be done using a cellphone that provides more location
detection than just GPS. Devices like iPhones actually have 3 different
methods to determine a location (GPS, tower triangulation, wifi hotspot
triangulation), so they provide pretty significant accuracy. We're
continuing to develop this proof of concept, but it's yet to be
determined how the technology can go from an iphone app to cheap
full-scale deployment. I'm personally somewhat skeptical of a purely
crowdsourced model providing sifnificant enough support for the tracking
infrastructure, but I absolutely think it should be part of the system -
either for error correction, bootstrapping the project (or drawing
attention to it), and for providing a parallel independent layer of the
stack to build services on top of. Also, never underestimate the power
of complaining as an incentivization model. People waiting for a bus
will want to complain about a late bus just as people on a bus will want
to complain if they're stuck in traffic.

Topp Labs work on the bus tracker app is still in early development, so
there's not a whole lot of documentation, but we'll try to keep the wiki
page for it updated and then point to more of a real home for the project:

http://ideas.topplabs.org/wiki/Real_Time_Bus_Tracker

phil

jrdesvernay wrote:
> comment 'GPS on all Buses' by jrdesvernay
>
> Ok there's few issues on the real time bus tracking that we should break
> down first of all:
> -We need to find a technology to track bus in real time
> -The technology must be as efficient and as cheap as possble
> -The information needs to be available in a variaty of format: WEBSITE,
> MOBILE WEBSITE, VOCAL SERVER, TEXT MESSAGES
> ------------------------------------------------------------------------------------------
>
> -We need to find a technology to track bus in real time
>
>
> So far, the most common technology to track a moving object anywhere on
> earth is GPS.
> We can find example form france to germany where buses are equiped with
> GPS and feed back their location to their central.
> Browsing a little bit on the web, on you'll find out that the only way
> to track a car or a bus efficiently is GPS.
> Alternatives method such as user input, are great but required user
> participation, which isn't realist as a system in itself, it is ony
> realist as a "sidekik", but this information cannot be operated
> effectively in a peer-to-peer mode, but in a centralised mode.
> As far as I know GPS is the only maybe the cheapest way to track a
> vehicule.
> The way GPS technology works, correct if I'm wrong, is that there's a
> transmitter in a devise that sends its data to a satallite which in turn
> pings back the latitude and longitude.
> So a GPS devise in a bus will send get pings back from a satalite wiht
> its longitude/latitude.
> Thats tells you teh position of the bus, but you need to enter then,
> the position of the bus stop as well, and you will need to give a unique
> ID to the bus stop as well as to the bus, so that both position can be
> mapped.
> The last thing is to get the map of teh city patched on the GPS
> positioning.
> that's how a car GPS works, and how it is telling you how far you are
> from your objective, and where you should go.
> This compare to any other method requires no interaction from users,
> and little interaction from the bus driver.
> ------------------------------------------------------------------------------------------
>
> The technology must be as efficient and as cheap as possble
>
>
> - i think the technology is efficient enough since it seems to work
> pretty wel in others countries, but the cost is probably high in terms
> of software and equipement.
> a) Every buses needs to be equiped with a GPS,
> b) and will need to have a central server with a database to collect
> real time data and feed it back to whatever system
> -------------------------------------------------------------------------------------------
>
> The information needs to be available in a variaty of format: WEBSITE,
> MOBILE WEBSITE, VOCAL SERVER, TEXT MESSAGES
>
>
> that's says it all, in france anyway, one needs to call a vocal server,
> tell the the stop id, and the server reply with the corresponding info,
> here the time the bus wil arrive.
> The idea I think is to have a set of databases and a program which in
> simple terms will feed back the position of the bus and its approximated
> time of arrival to a variety of medium where the user in a way or
> another will query his Bus Stop ID through a vocal server, text message
> server, web site, or wap website.
>
>
> One cool thing I think is to have a text messaging services, where you
> send your bus stop ID to a no, and you got a reply wiht the approximate
> time, with the option of having this text messsage send to you every
> morning for example, much like an airline remaiding you of you flight
> via text message.
> So I think so far there's no beating around the bush to find out what
> technology to use, but rather finding the most efficient way to fund it,
> and how to implement it.Whihc is where crowdsourcing could be put to
> use.
>
>
> So my questions is:
> How sould we implement the system:
> Should we for example have companies specialised in real time transit
> system to implement it,
> or should it be built by freelances in a semi-open source manner at the
> risk of not succeding,
> or should the Bus Transit authority/department hire professionel
> itself and implement the project itself.
> -any other sugestions?
>
>
> How should we fund the system:
> Could we ask every resident of a city to wire monney to an account
> linked to the project, and select individuals to manage the funding?
> Could we co-fund the project, by funding either the hardware or
> software?
> -What incentives could we give to the people who fund the project?? Bus
> ticket reduction for a set period??reembursement of the sums given by
> everybody wiuht interest/ or no interest.
> -set up a special stock of the public transportation and get people to
> buy shares to fund the project, with the ability to resale their
> shares??
>
>
> This is exiting and I hope I'l get replies..
> -
>
> View original:
> http://diycity.org/discussions/diycity-challenge-3#comment-376
> Post reply: http://diycity.org/comment/reply/242/376, or you can post
> a new post by e-mail: discussions@diycity.org
>
> Note that if you reply to this mail delete the quoted text and do not
> modify the subject field or the message will not reach it's destiny.
> --
> You are subscribed from the group 'Discussions' at DIYcity.
> To manage your subscription, visit http://diycity.org/og/manage/58.

Again, not DIY, but haven't companies like Clever Devices, located on LI, solved this? I think they use a combination of odometer readings/distance traveled measurements, knowledge of the route, and GPS to determine position. They had it in Pittsburgh (not reported in realtime, but in-bus "next stop" signs) and it worked fine... (BTW, I'm not affiliated with Clever Devices).