Putting high-tech sensors to work is the remit of the aptly-named Everyday Sensing and Perception working group at Intel, writes Karlin Lillington
PHONES WHICH know who is holding them; remote controls that sense when you are speaking too fast during a presentation; hand-helds that tell your house when you walk into a room and get it to adjust the music to your favourite playlist.
Those are just a few glimpses of possible intelligent devices in the near future thanks to embedded sensors, according to Andrew Chien, Intel's director of research, who was in Ireland last week to address an Intel-hosted conference on innovation in Leixlip.
As often happens, sensing technologies are not quite being used as was initially predicted, he says. Much research initially focused on creating sensor networks made up of millions of minute sensors - often referred to as "smart dust" - that might be embedded in roads or walls.
"You'd basically tile the world with sensors and they'd be able to tell where you were in a house or who's in a room," he says. "A great idea, but it introduced all sorts of interesting research problems."
Key among those were making them cheaply enough, deploying them and then creating the networks to utilise them. Also, it was hard to justify such a massive investment and raise capital for something with no immediate commercial application.
"If those markets do develop, they will be slow to develop," he adds, meaning that they are of no immediate interest to Intel.
However, mobile devices interest Intel quite a bit - and as anyone who has flipped an iPhone sideways or used the Wii gaming platform will know, embedded sensors can do some pretty cool things in a device.
The advantage of thinking about sensor applications is that people buy and use many devices already and sensors can be incorporated without building out full networks to support them, says Chien.
The idea of sensors in devices isn't even startlingly new or strange - a lot of devices we carry around already have a lot of sensors in them, "but we don't do anything interesting with that data".
Consider the average mobile phone: it may contain a camera, a microphone and a radio frequency device. A few have accelerometers, the motion sensors inbuilt in iPhones and the Wii.
Putting those and other sensors to productive work is the remit of Intel's aptly-named ESP - Everyday Sensing and Perception - working group, launched last year. It is one of the largest efforts within Intel, Chien points out.
ESP marries sensors with mobile technologies ranging from phones to laptops. The group's goal is "to find out how to do user-centric computing" to enable devices to work more intelligently for a user in everyday environments, all day long.
"This will happen in small devices you don't mind having with you all the time," Chien notes. One such device might be a small "egocentric video" camera worn on your shoulder, which quietly videos what you are seeing all day.
"If you had such a device, it could recognise where you last saw an object," he says. That might be handy for finding the keys, but more seriously, could be useful for documenting activities through the day and serve as a memory aid later.
An accelerometer - a motion sensor - could have many applications. "An accelerometer in a remote control could know who is holding it," Chien says. "You could build a controller that would know a given person's hand."
Such a controller could limit a child's access to films or television programmes, or automatically deliver a user's favourite channels or stored programmes.
Chien notes that Intel's research lab in Leixlip, Co Kildare, which does work on geriatric medical applications, already uses sensors in devices that measure gait, heartbeat rate and acceleration - whether someone is falling - applications in a device that might send an alert to a relative's mobile phone if a parent has taken a fall, for example. He also suggests the possibility of a "safety phone", which could draw upon data about neighbourhoods and crime statistics to alert a user that a particular area - a car park or bus station, for example - is unsafe.
"Then there are systems that could actually recommend behaviour," he says. An increased heart rate or sweaty palm is typically a sign of stress.
If someone giving a talk, who may be prone to speaking too quickly when nervous, has a sensor device measuring pulse rate or increased sweat - perhaps that device could be built into the remote for a slide projector - it could buzz gently or flash a message on the computer reminding the speaker to slow down.
If those sound like wacky and weird applications, Chien notes that we have transformed our thinking about mobile phones in just over a decade of their being commonly used.
"Frankly, I don't know what a phone is any more," he says, noting that it now does numerous things once delivered through other devices. "And you have people using Bluetooth headsets; you've totally disaggregated the phone further into its parts."
A major challenge for the ESP group is not coming up with applications so much as making them workable. Sensor devices can be very sensitive, for example.
"Weird things matter," says Chien - like altitude. Certain sensors give completely different readings at altitude, which might ruin an application for a frequent flyer.
Other things on the ESP want list? "We want sensing systems to work across multiple devices - right now they don't talk well to each other."
And finally, they need reliability and consistency of operation if they are to be on all the time. Chien stresses these are problems Intel aims to solve, not simply explore. Intel goes into markets where it feels there are eventual sales to be made.
So who knows - before long, that little voice you hear may be your mobile telling you to calm down, your heart rate is too fast.
"Before long, that little voice you hear may be your mobile phone telling you to calm down, your heart rate is too fast
