Movipolis: A proactive, personal, multi-media interpreter for the sensory impaired
Luis Alcantara
LUIS ALCANTARA: First of all, excuse me for my English. It is not my mother tongue, so I hope you understand.
I tried to be as practical as possible. I come from a small company, so we have to be practical, because at the end of the day what we want is to make money. What we want to do is to make things used by the people. We know technology and we know it's very useful, but we have to make it useful.
This is what I wanted to talk today. First, why we are doing this project, and then how it works, just a little bit, not getting into technical details - if you want to, we can talk later on that - what it offers for people and, finally, how to make a profit, or how to make it financially viable, because that is our ethos. Finally, if you want, if we have time, we can speak of who we are and why I am here, but it's not so important, so I will leave it for the end.
Why we are talking about this project here. In fact this project or this platform called Movipolis was not developed for the sensory impaired people. It was not developed for handicapped people. It was developed as a much wider element.
The tourist sector is very important to Spain, and we wanted to have a mechanism to reach tourists, as many tourists as possible when they visit the town, and offer them a guided service. We developed the system. It's up and working now and we are trying to find new uses for it. One that came to our mind was for the sensory impaired people. What are the needs as we understand of these kind of sensory impaired people. By sensory, I mean deaf and blind!
We wanted to guide in the environment. When I say "guide" I am not really talking about guiding them to walk around a London street. That's quite difficult, and again we are practical, and we know we cannot do that in a short time. There are mechanisms for that. GPS has been explained before.
We want to guide people as when somebody explains to them when they ask them at the corner of the street, "What is here? I want to reach a shop", not just taking them by the hand and leading them. We also wanted to provide interaction with the environment, what they cannot see or hear, and also to allow them to interact, which means people in this society can participate in an environment as any other people would do that.
How to do that? Basically we provide them with a device that will render all the information they cannot perceive by their less operating senses. This device is then like an interpreter of the environment. For example, if there is something in the environment that they cannot see, like a banner with some signs, we can make that banner be read by a device to them, so they can hear.
Now, to do that, we want to detect them in the special points, and for that we use Bluetooth technology. We want to be proactive. We cannot expect people to ask for information because they don't see the banner. If they don't see the banner, they cannot ask what is written in it. So we have to tell them, "There is a banner and this is what is written on it."
For the interactivity, we want to make people interact in the environment. For example, an automatic door that you could possibly open with your hand, but you cannot see the handle. You could somehow put a device in the environment to make the door open when someone wants to open it. We will come to that later.
This is the concept of what we want to do. Basically, we have the environment. This is a street, a building, doors. There is lots of information here - visual information, audio information, SMS, everything - and the person with all five senses can sense that, but somebody who is blind cannot take it all in. So what we do is, with that information that is missing, that we know is going to be missed, we put it into another format and render that to the user using a device.
We will speak now how it works. There must be an entity or an institution that's going to run the service, to operate the service. This is a key part. So far today, it has not been spoken, not that much, that there has to be an entity that's going to run the service, to run this intelligence somehow. Say it is the RNIB or the Government, they will have to install the hardware, the infrastructure, to make this operate.
We use beacons - or we call them gateways, because they are really a bridge - transmitting in both directions from the user to the server and back to the system, and the other way round, and they are the devices that detect the people. The operator - and I am talking about the institution - has to configure that infrastructure to tell it how to guide the people, how to contact the people, and what it has to tell people.
Once everything is up and running and configured, then you provide the user with the device, you register the user, you tell the system what the user's characteristics are - for instance, a blind person or a deaf person - and then that's it. After that, the system will contact the user directly. The phone will vibrate or will make a sound. You just pick up the phone and you hear something which is telling you what is around you, and it will contact you. Also, you can basically act on the device with the interface to get more information from the system, so it's a two-way communication.
This is a sketch of the architecture of an existing platform. This was a research project three years ago; now it's a ready-made product. This ready-made product consists really of three pieces of software that have to be installed in three types of devices.
The most important device perhaps is what we call the gateway. It's really a small computer with a WiFi interface and a Bluetooth interface. We also have the user device which the user will be using, just a mobile phone. We didn't want to use the PDA, and we are practical, so we wanted it to be used by as many people as possible, and we wanted it to be in time something of a mass service. We know that people go around with a PDA, like the people in this room, but that is not normal!
Finally, the other important part of the platform is the server. This is the intelligent part of the system, and it runs an expert system that knows at all times what is going on in the area. It knows where the users are, where they have been, what is the path they followed, what information they have queried so far and what information has been transferred so far. Then the server gets all the events that are going on in the area.
For example, somebody gets into a Bluetooth area or somebody gets out of a Bluetooth area. Somebody asks for information and then, according to a set of rules that we configure, it would decide what message it has to send. Therefore, it is fitted with a lot of information you want to put in, like general information of what's around, say for example, the toilets, the bus stops, whatever. Also, why not commercial information or services, shops, cinemas, and so on? All that information has to be put into the proper format for all the people you are going to support. It can be audio format or visual format. It can be everything that a mobile phone can display, of course.
Finally, all the information that is produced by this is collected, so, finally, you have where blind people go, how long they stay in each place, all they ask for, and so on. They offer an image of how a deployment can be optimized in a town. We don't want to provide whole Bluetooth coverage in a town. It is too expensive to deploy such a network.
Bluetooth range, because of the original implementation and because of limitations in power, it can be around 20-30 metres; no more. We wanted to install those gateways or beacons at special points where we know that people need to be contacted by the system, for example, in bus stops, the entrance to the underground, and also in the places where people are expected to ask for information. In the rest of the area, obviously we don't want the people to be lost, so they would have GPRS access. They always will have some kind of mechanism to contact the server.
Those beacons can be located everywhere, so what we do for that - we cannot expect to have a line connection at the bus stop - we use WiFi for that. We deploy WiFi areas as an umbrella covering those areas in town where we are going to install the gateways.
As I said, everything is controlled by an intelligence system. That system has a set of rules which you can configure and each rule has an associated event. For example, when somebody passes near the entrance of a house, it has a set of conditions. For example, if the person has already passed through a certain square, a specific message is received, the time, etc. And finally a consequence that will occur when all the conditions are met. The consequences normally are to send a message.
You also fit the system with a set of messages. You must be aware that the messages are not uniquely associaited to places: you can get a message in many hot-spots, and you can have different messages at the same hot-spots - you will not always send the same message at the same place. It will be determined by the context. The context is built dynamically by the system.
That's how it works. Now we talk about what it offers to people. What are the advantages? First of all, for the disabled people, it provides freedom. I don't think we need to speak any more about that. We have talked plenty today. It is proactive. We are not waiting for people to ask for information because they don't know what the information is going to be. We are telling the people directly what information is around them. That's why the phone will actually ring as if somebody was calling them.
It is intelligent, in the sense as I said before about the context and how it knows the whereabouts of the people. It knows where they have been and it knows what information there is already.
It is interactive, which is a very important thing. Those beacons that we are installing are not just radio beacons that are there and the device knows them. Actually the device is quite passive. The beacons act as a bridge, so all the information transferred to the mobile phone is transferred over a Bluetooth.
It transfers in both ways. So, the user, for example, can click on links to ask for information. That information will be delivered over a Bluetooth. There is no cost for the user. It's convenient. We try to make things very intuitive, and that's always a problem, because at the moment we are limited when developing a user interface by the software system that the device is based on. Although we know, in Spain at least, some equipment has display- to-audio translators.
Basically any kind of information that you can render in a mobile phone can be transferred using this system. We are not using SMS. We are not using common multimedia messages. We are using something more powerful than that. In fact it's kind of individual web pages that will be sent. Basically, when you transfer information, you send out a kind of web page over Bluetooth with the audio images etc.. If you want something else, you just click on the links and you browse over those contents or new pages.
What it brings to the operating party, to the actual institution, we talked a bit before, which is really the enabling part of this kind of service. The installation of the infrastructure, there is the cost of installation and deployment, but after that the information is over the WiFi and Bluetooth and there are no GPS costs. It's flexible. We are not only using standard communication, but all the pieces of the infrastructure are movable and you can change the contents.
You know how it's working because it produces statistics. It has a return channel, so you can have feedback from users and you can act on the environment depending on what you want. As was said in a previous presentation, you can book a ticket for a cinema, for example, over the system, because you have a channel for that.
Very important, you have a promotion channel. You can be a kind of Bluetooth operator, so you can deliver information that is not only general information but it's also commercial information - ads, for example - and it opens a door for convergence among different players in the area and agencies, like the government, shopping centres, organisations, blind associations. In theory, they all could not run, but configure their own rules into the same platform because it's all web based.
Let's get back into the more economical part of things; the system and the costs. What does it cost to operate the system? There will be a set of areas of costs. There is a software code, but there are things that are not related to us. For example, you have to produce the contents, and before you produce the contents you have to decide what is the service you want to offer. That's the first part. You have to do an analysis and describe the design and the rules.
The technical part is our part, but the financial analysis has to be done by the operating party, and you have to add the costs of the installation. But, once it is installed and configured, the costs are very small, because the costs of operation are just the personnel costs. Normally you have people that go to the website, enter the administration password and configure the rules and the contents.
You have to pay some communication costs because of the WiFi hot-spot basically (they need an Internet connection), and also because at some point the user will ask for information in places where there is no Bluetooth coverage. In that case, you pay for that. You have licences and rights - for example, you install a gateway at a place which is not yours - and the maintenance.
Another important point is the revenue. The institution that runs the service may want to make it financially viable, and to do the most immediate things, to make a service fee for the users, say blind people, deaf people, whoever. That's a basic model. But you don't need to do that. For example, you get the money by advertising. Of course, that could lead to a very intrusive service if it is not properly designed.
How do you configure it? The idea is that you not only include the commercial information in the database in the contents, which is the normal passive model, but you can also send information proactively according to the user context. So, for that matter, you could charge the third party, the party who is being advertised, a higher charge than you normally could if you just put a banner on the street.
But further on, the idea here is that you could potentially include all those third parties, like cinemas, shopping centres, transport entities, into the platform by offering them to install their own gateways. They can be part of the infrastructure. Finally, because you have installed the network and you have a network that is intelligent, you can use it for other things. You can use it for third party access. You can run a service or you could offer it to book hotels, or whatever. You could even offer a free basic service but you could charge for extra services, like downloading contents.
And that's it for the Movipolis. I think I had better answer questions from you, because that's the most important thing. Any questions?
NEW SPEAKER: We have some members who are partially sighted or blind. One of my concerns would be how are you going to make it accessible to people who have other impairments as well, which is 80% of disabled people. In particular, are you going to have a voice activated version or tone of voice problems? Are people going to be able to attach the device to a larger keyboard?
LUIS ALCANTARA: This is a matter we are considering. This product has been developed for, in principle, people without a disability. It is in the commercial phase. We also have now started the contents for developing the system for disabled people. Actually we are in contact with the blind organisation in Spain, but we have not developed yet the interfaces. So, this is up and running, but it doesn't have, for example, the mechanisms for you to interact if you were blind. At the moment, you have the mobile phone and it contacts you. If you cannot see, you just pick up the phone, you press the OK button and you hear, and that's it.
Obviously if you are blind you can't see the links on the screen and you can't go to a link, because that part has not been developed yet. But in theory, you can have an audio interpreter also running in the phone so you could hear what's in the screen, and the same for all disabilities. So, the stage where we are in development is that the system works for people without disabilities and we have to develop the interface for people with disabilities.
NEW SPEAKER: I am a student from Buckinghamshire University. I wanted to ask, if this phone is being used by the blind and visually impaired person, what is the guarantee for the accuracy of the information and also the quickness of the transfer of the data?
LUIS ALCANTARA: I said before, this cannot be used as a guidance system. You cannot trust this system, for example, to avoid people falling on a hole in the street, because there is a time between you detecting the person and the time when the phone rings and you read or hear the information.
At the moment it's pretty good somehow, and if you get inside the Bluetooth beacon, say 20-30 metres, in about two or three seconds the mobile phone will ring. Depending on the contents, about five or ten seconds later you will have the full information, so the phone will ring again and you will hear the information. The first ring is to make sure that people stop! Otherwise, if you are too fast, you go away from the area, so this is too late for information on Bluetooth. You have two, three, four seconds time lapse there, and you have to be aware of that, from the proximity of the area until the user is informed. You have to design the service bearing in mind it's not an emergency system, it's an information system.
NEW SPEAKER: I believe there is a similar system in Birmingham for the guide dogs for the RNIB. Are you connected with that in any way?
LUIS ALCANTARA: No. I didn't know that. I will look into that. If you want to ask any further questions, you can just contact me later. Thank you.
JOHN GILL: Thank you very much.
APPLAUSE
| Previous | Contents | Next |