x Abu Dhabi, UAESunday 23 July 2017

Blind, but able to 'see' traffic

A UAE University professor is leading a project to create a wireless system using GPS to help visually impaired pedestrians.

Illustration by Fred Matamoros for The National
Illustration by Fred Matamoros for The National

Al Ain // Hesham Kamel, an engineer who cannot see, doesn't dare cross the street alone. And given the speeding culture on UAE roads, he reckons even sighted pedestrians need to watch out.

The UAE University professor is leading a three-year project to create a system using wireless communications and GPS to help blind people such as him "see" their environment better.

The Pedestrian Alarm and Security System (Pass) should help them detect moving objects - particularly cars speeding down a street they intend to cross - and let drivers know when a blind person is near.

Existing systems work for stationary objects such as pedestrian signals and shopfronts. But until now, providing fast and accurate information about moving objects has eluded engineers.

"Anything less than perfectly and you can kill a person," said Dr Kamel, who lost his sight at the age of 23 after an eye procedure. "Here, the pedestrians watch out for the vehicles, and not the other way around."

There are three main challenges: the system must provide accurate, reliable information; it must give users simple, effective information; and it must make commercial sense for manufacturers and governments.

The Pass system aims to provide information in real-time and "interface" with users as simply as possible. After a year of design work, the team is making prototypes and plans to deploy a trial version in 2014.

Blind pedestrians and cars would carry devices that continuously transmit their location and receive data about others within one kilometre - all within tens of milliseconds, according to Abderrahmane Lakas, a co-principal investigator on the project.

For distances up to 1km, cars and pedestrians would exchange information directly. For longer distances, they would exchange GPS data with a central server, via internet-connected roadside units. These offer more bandwidth than the regular 3G/GSM mobile phone network, and do not require additional internet gateways.

The devices could tell pedestrians about others carrying activated GPSs, and indicate useful features such as pedestrian crossings or shops, based on a digital map. Crossings, bridges and subways would also be fitted with transmitters.

With so much information available, the challenge is to put it all together and give pedestrians information they can use.

Anything dangerous - a speeding car - will take priority. And if they are not crossing the road, the system will concentrate on pedestrians in their path rather than cars.

It will be conveyed through a combination of speaking, sounds and vibrations. Voice synthesis can supply the most information ("Crossing 50 metres ahead") but takes longer. Beeps and musical notes are quick, but may go unheard on a noisy street. Vibrations avoid that, but convey only basic information.

"The interface has to be really simple, [yet it] is not easy at all, because you have to account for a lot of different cases," said Dr Kemal.

For drivers, a dot on a GPS map could represent visually impaired people, with a spoken warning about their distance.

Richard Long, a professor of blindness at Western Michigan University, believes the Pass project is a step forward in pursuit of making the streets safer for blind pedestrians.

Direct, peer-to-peer transfer can be practically instantaneous, while the combination of GPS and internet communication is slower and less precise, especially in cities where buildings can block out satellites.

"We do have the ability to communicate very quickly and efficiently between vehicles and pedestrians and infrastructure. There's no room for error - you're talking about an approaching vehicle," he said.

Still, the interface requires fine-tuning - even in the case of a simple beep for an approaching car. It needs to include an instant judgment of whether the pedestrian plans to cross the road, the speed of the car, and how far ahead an alert is needed.

"It makes a lot of sense intuitively, but believe me, there are 1,000 very challenging questions underneath that beep," he said.

The question of moving cars has become especially urgent in recent years with the rise of all-but-silent hybrid and electrical vehicles. To counter the absence of engine noise as an aural cue, researchers in the US have been working with carmakers on vehicles that make a unique noise when they near a crossing.

That kind of simplicity is key. Make a system hard to use, and people won't use it, said Bill Crandall, a scientist at the Smith-Kettlewell Eye Research Institute in California, which develops technology for the blind.

And if people aren't going to use it, governments won't back it, especially in times of austerity. Dr Crandall notes that even basic systems such as the one developed at his institute, which emits spoken messages near crossings or bus stops, have been a tough sell. "Any of these systems require a critical mass to be useful," he said.

Still, the technology used in Pass and other systems developed initially for the blind might have broader uses, which the UAE University team is exploring.

Dr Kamel points out that speech synthesis used in GPS and other applications originated to help the blind. "It's not really true that research and work that's done for physically challenged people stays only within the physically challenged community," he said.

Pass devices could be carried by other disabled or elderly people or even ambulances, to tell cars ahead of them to move out of their way.

The in-car units could help improve traffic flow - an application being pursued in a separate two-year project by Dr Lakas.

They would gather information about local road conditions and share it instantly with other cars. If a snarl lay ahead, the unit could look for clearer routes.

"As soon as you turn on your car ... the device starts talking to all the cars in the neighbouring area, and those cars to their neighbours, and you collect information about the traffic, you process it locally, and you create a map of traffic around the path you're taking toward your destination," Dr Lakas said. "And this process is done continuously."

And he is keen to emphasise the project's origin.

"The project was first driven by a need to give more independence to people who are visually impaired," he said, but in the end "there are many applications".

 

chuang@thenational.ae