Feb. 22, 2012 at 1:18 PM ET
The images of nearly every major stretch of road taken by Google's Street View team and the snapshots we capture with our smartphones may soon be all we need to navigate the world, according to an Australian researcher.
That is, we can ditch the expensive satellite and computer technologies that power modern GPS systems and rely on low-resolution pictures instead, Michael Milford, an engineer at the Queensland University of Technology in Brisbane, explained.
The system was inspired by his research on how rats get around.
"Rats (and many insects) are amazingly capable navigators with quite 'poor' vision, much 'worse' than a human," Milford wrote to me in an email. "While we may see up to 100 megapixels, a small insect or rodent might only see a few thousand pixels with their eyes. The system uses similarly small images."
How it works
Milford envisions building a visual guidance system that combines a database of images stored on a local hard drive and some mathematical algorithms.
The system starts by testing thousands, or millions, of hypotheses about where it is and then tests them again and again by comparing what it sees to the images in the database.
So, say you're in the kitchen at your office. The system would look around and see a microwave and fridge, noting you're in a kitchen. If you then step outside into the hallway, it would match the images to an existing image of the corridor in the hard drive and test the hypothesis you're at the office.
"It would also test the assumption that you're in the car park, in a restaurant, etc. … millions of other locations. As you move around the environment a bit, the number of plausible hypotheses will drop like a stone until there is only one left," Milford explained.
How it compares
Gadgets such as the iPhone have brought GPS to the masses. As many of us have learned, we can be just about anywhere and with the tap of a screen find out almost exactly where in the world we are.
The behind the scenes wizardry that makes this happen is a combination of location signals beamed down from satellites and some assistance from data contained in the cellular network that allows our phones to work.
One advantage to this system is we don't, usually, have to move around for the satellites and cell towers to tell our smartphone where it is. But all this technology costs money and is sometimes unreliable, especially in big cities where tall buildings can scramble signals, according to Milford.
His approach, dubbed SeqSLAM (Sequence Simultaneous Localization and Mapping), doesn't need access to external data to work.
"The beauty of using small images is that you can store the entire world's road network on a standard hard drive, without needing access to the 'cloud,'" he said.
What's more, he added, combined with so-called "dead reckoning" navigation (how sailors plotted their course through uncharted waters), would allow "some ability to keep track of where you are in a place with no existing data."
"And that data would then be uploaded to the central database for the next person to use."
John Roach is a contributing writer for msnbc.com. To learn more about him, check out his website and follow him on Twitter. For more of our Future of Technology series, watch the featured video below.