Imagine, if a smartphone could detect when the driver is too tired and yawning, and send an alert to his supervisor. Or, if data transfer could be automatically scheduled in accordance with the strength of the signal, in order to save battery.
These are some areas of research Venkat Padmanabhan, Principal Researcher at Microsoft Research India, is working on. He is one of the recipients of this year's Shanti Swarup Bhatnagar Prize for Science and Technology.
He won the award “for his pioneering work on indoor localization, smartphone-based sensing, and mobile communication, which has resulted in academic and industry-wide impact and also technology transfer to Microsoft’s products,” says the citation.
In an interaction with The Hindu recently, Dr Padmanabhan gave some insights into the ongoing work.
Road safety
It's the high rate of accidents that prompted the study on making use of smartphones to ensure road safety. Data is gathered from sensing devices: the front-camera of the phone that looks at the driver, the rear camera that points to the road ahead, the phone’s GPS and inertial sensors; besides the OBD-II (on-board diagnostics) scanner.
The data is aggregated and analysed at the backend to provide valuable real-time information about the vehicle and the driver. Says Dr Padmanabhan, "The camera could capture the driver yawning or talking on the mobile phone or whether he is driving without seat belt. Similarly, from the camera pointing ahead, it's possible to find out if the driver is going too close to the vehicle ahead." The challenge is in designing detectors that are both efficient and effective.
One of the applications is in a fleet of vehicles, like cabs or buses or trucks, wherein the data can be overseen by a supervisor. The solution is called HAMS (Harnessing AutoMobiles for Safety), and an initial version of it is has been deployed on a dozen office cabs at Microsoft Research India in Bengaluru. Data from over 10,000 km has been gathered.
Battery drain in smartphones
Phones running out of battery is a common problem. A major reason for it is the cellular radio communication. It has been found that the energy consumed per bit is six times higher when the signal is weak than when it is strong.
A Microsoft team is working on a system that would help cut down the energy used during data communication. To realise that goal, applications have to communicate when signal is strong, by either putting off non-urgent communication or by advancing anticipated communication to coincide with periods of strong signal, Dr Padmanabhan said.
The experiments have been performed on four cellular networks across two large metropolitan areas, one in India and the other in the U.S.
Flexible applications, such as email syncing, photo sharing, and on-demand streaming, lead to saving of energy. Experiments have shown that up to 10 per cent energy saving can be achieved for email sync and 60 per cent for on-demand streaming.
Future work is focussed on better approaches to signal strength prediction and design of appropriate APIs.
