Imagine smartphone GPS that uses 99.96% less power

How is such a concept possible? If you guessed the cloud, you are on the money. Microsoft Research is working on a project which, in broad strokes off-loads the high-power consuming tasks to the cloud. It is being called, Cloud-Offloaded GPS (CO-GPS).

We all know that GPS is a power-hungry feature and unless you are packing a charging cable or have a device with a ginormous-sized power cell, it is not practical to use navigation with impunity on a road trip and still have enough power for a functioning device.

In the current environment, this limits the usefulness of GPS in many applications.  GPS location acquisition is the giant power grab every time. There are two reasons for this. The first is that satellite information is sent at a woefully slow rate, as low as 50-bits-per-second. The second is due to the amount of signal processing that is required, and the inherent weakness in the signal. That requires the GPS chip, and the CPU to work continuously.

Researchers are now using a number of already available collections of information to create a post-processing environment. They have also built a system called CLEO (Cultivating the Long tail in Environmental Observations) to use as a sensing device to evaluate how efficient their solution may be. Their 14 page report has extraordinary detail, but the bottom line is that CO-GPS solution can help devices obtain a GPS fix in only 2-milliseconds, and subsequently results in power consumption from 1 Joule of energy to about 0.4 – millijoules, 99.96% more efficient. Apparently that number is not good enough as the researchers are looking to make things even more efficient by experimenting with signal processing and compression methods.

CO-GPS does not require long continuous connectivity because most of the data is accessed “offline” from the satellite feed. What this means is that as this technology and other solutions gain ground to make GPS more energy efficient, there will be born a whole new set of services and applications.

sources: Microsoft Research via MIT Technology Review and The Verge