It has long been known that there are two kinds of software (and hardware), the military use and civil use. For example, it is known that the military, at least in the United States, have systems that give more precision than can give schemes for civilians. A GPS data for the US military at best occurs in decimeter accuracy. One for the ordinary citizen, you may have an uncertainty even meters. In fact, the GPS cell phones give an accuracy of 2-3 meters at acceptable conditions, but in many cases the accuracy becomes an uncertainty of 10 meters, which is not much but in the end is not desirable.
However, as regards GPS, it seems that the teams for civilians could give a more precise positioning, but this larger antenna is required . The reason for this is that the problem is the multipath reception, which reduces the accuracy. This spoils the information because the satellite signal takes a series of paths to reach the receiver. Greater directional antenna and sophisticated processing can reduce multipath reception and achieve centimeter accuracy.
This is not reasonable for most applications, but now has an algorithm to extract , with greater accuracy, without using specialized hardware. A team of researchers from the University of Texas at Austin has taken the hardware that is in cell phones and software testing only has tried to improve the accuracy of GPS. Note that in some cases the computer uses hardware upgrades such as a reference oscillator temperature compensation, but even if this, the results are very interesting.
The key to the process is a sophisticated software filter that uses the random motion of the receiver. The simple GPS receivers demodulate the signal and extract the message. This is a package of data tells the GPS position fixed by a couple of meters. But if we look at the problem from which sends the message (the carrier), it is possible to improve the position and give a figure with uncertainty in centimeters, so accurate even down to 2mm of uncertainty, at least in theory. The method developed makes the filter attempt to remove the errors caused by multipath reception. This filter includes a term for the position and velocity of the antennas. A surprising result is that the estimated improvement
Here’s a demonstration.
By the time all processing is done on a computer that is separate from the phone, but in principle it is possible phone’s processor could do the task. Eventually, a better receiver costs about $ 50 today. The applications of this new precision fall in robotics, cars that run themselves and 3D scanning
References:.
Centimeter Positioning with a Smartphone-Quality GNSS Antenna by Kenneth
Pesyna M., Jr., Robert W. Heath Jr. and Todd E. Humphreys, University of Texas at Austin
No comments:
Post a Comment