The ability to utilize GPS has revolutionized the mobile phone industry. But GPS is not without its limitations–chief among them that the service doesn’t work well indoors. Now computer engineers at Duke University have come up with a way to compensate for GPS’s shortcomings. Using the microphones, cameras, and accelerometers in a smartphone, the researchers were able to identify a person’s location in the end all and be all of indoor mazes: the shopping mall.
Each mall business has its own visual and acoustic fingerprint, the researchers reasoned, e.g. a mall Starbucks at 3:00 in the afternoon has a unique look and sound. By combining the visual and audio input with the movements measured by the phones’ accelerometers, the researchers were able to pinpoint a person’s location in over 50 mall stores with 87 percent accuracy.
Of course, without preexisting baselines for these indoor geographic fingerprints, the researchers have to start from scratch with each location. Still, the program, dubbed SurroundSense, could eventually prove useful for applications aimed at mobile phone users in key locations, such as an airport or convention center.
[Image, Romit Roy Choudhury; via Duke University]