Data Integration, Feasibility Studies, Geospatial Data, IoT Development, Pervasive Location™, Prototyping, Requirements Definition, Solution Design
Offering services linked to a very specific location is meant to be one of the next big things. But what if the location moves? Sounds odd, but services based on a moving location are just as relevant as those for fixed services. For example, when you’re on a train, you might want to know which way to head for the nearest toilet, or to find a free seat. This is true also when you are waiting for the train to arrive at a platform — which carriage do you get on to get a seat?
The obvious problem with services based on a moving location is that you cannot use GPS. You therefore have to rely on other technologies, such as beacons (iBeacon, Eddystone, etc.) or Wi-Fi. If you have appropriate beacons installed in each carriage, then your smartphone can be used to detect which carriage you are in and then you can link that to data telling you what facilities there are on the train. But what about before the train turns up — you cannot scan the train before it arrives. Can you?
It turns out in the UK that there is no simple way of finding out what carriages go to make up a train service. Although train operators plan this on a daily basis, it is not unheard of for a driver to take the wrong train out of a depot, or, dare we say it, for train services to be disrupted in any way. And then there's the thorny issue of which way round the train is running. If you've got a seat reserved in carriage 'A' you could stand at one end of the platform only to find that carriage 'A' is at the back of the train, not the front.
Operators' daily plans include the list of carriages that go to make up each train service, but these are not exposed by any of the public rail data feeds in the UK. Add to that the lack of publicly available facilities information, as well as the opportunity for trains to change, and even operators do not know what facilities might or might not be available on a particular service. Under normal circumstances, experience helps; commuters will know the best carriage to get on, where the bike space is, etc. However, not all of us are hardened commuters.
In a project funded by the RRUKA, we have been working with the University of Surrey and Loughborough University to work out if location-based technologies can be used to automatically detect what carriages and facilities go to make up a train service.
Our work has been to build a service which, through a series of deployed scanning devices, can detect passing trains from a fixed location, then report the information back to a central web service to collate the results, work out which carriages are in which train service and link this with facilities information to present the information to passengers.
There are several ways in which carriages can be detected. Perhaps the most reliable is to use RFID tags attached to each carriage and then have track-side devices scan carriages as they go past. However, while reliable, deploying any new infrastructure in the rail industry is expensive. So, what if we could use technologies that already exist on a large number of trains?
In the project, we have been trialling the use of Wi-Fi scanners positioned next to the track to detect passing trains — the majority of which have Wi-Fi installed in every carriage. Unlike RFID, Wi-Fi scanners can be placed further away from the track, so such scanners could be positioned at stations, or indeed be mobile and carried by a train guard, so reduce installation costs. The drawback is that the data is much noisier (think of all those Wi-Fi devices across the UK).
In a live trial with an operator over four weeks, we have deployed an Android app onto a series of Orange Pi boards connected to a Mi-Fi device to continuously scan for Wi-Fi access points and send the data back to the web service. Android was chosen because of its support across a range of platforms, and so that the same app could be used on a smartphone for mobile scanning.
Scanning was achieved using the Pervasive Location SDK, which, as well as being able to geofence using GPS, beacons and Wi-Fi on iOS and Android, can also continuously scan for Wi-Fi signals on Android. The only hard part left for the app to do was filter the data and send it to the web service.
Digital processing is at the heart of innovation. It may be that you have a difficult or complex problem. Pervasive Intelligence have the expertise to cut through the complexity to give you a solution you can use in the real world.