Use of the Term "channels" in GNSS Receivers

Articles about GPS, GLONASS, GALILEO, WAAS and other satellite navigation systems
jimh
Posts: 11672
Joined: Fri Oct 09, 2015 12:25 pm
Location: Michigan, Lower Peninsula
Contact:

Use of the Term "channels" in GNSS Receivers

Postby jimh » Fri Oct 21, 2016 1:40 pm

Many GNSS receivers mention a specification using the term "channels" to describe the capabilities of their receivers. This is a confusing term, and a bit of background on GNSS reception is necessary to understand it. Let's limit the discussion to the US Air Force GPS system to keep things simple.

A recreational-grade marine GPS receiver utilizes the civilian signal from GPS, called the L1 Coarse Acquisition (L1 C/A) signal. All satellites transmit this signal on the same frequency. The term "channel" is often used to designate assigned frequencies, but in the case of GPS, the frequency is the same for all satellites. They all transmit at 1575.42-MHz. There are no other frequencies used by L1 C/A receivers; they just listen on that one frequency.

Although all the signals from GPS satellites are sent on the same frequency, they are modulated by different digital modulation schemes. The basic modulation scheme is the Coarse Acquisition code, which is a 1.023-MHz digital signal of 1,023 bits. These digital modulation patterns are carefully chosen in order that there is very little correlation between the codes. Thus each GPS satellite transmits its signal on 1575.42-MHz with a different digital modulation pattern of 1,023 bits. These patterns are known as psuedo-random noise or PRN, and are referred to by a number. For navigation satellites, the PRN numbers range from 1 to 32, typically.

A GPS receiver looks for signals by tuning itself to the L1 frequency. (To digress a bit, the receiver tunes to the L1 frequency plus or minus the calculated Doppler Shift of the signal that results from the apparent motion of the transmitter to the stationary receiver.) It then searches around the calculated frequency, looking for signals from the satellite it is trying to acquire. As the receiver steps through small frequency increments, it applies the the same digital modulation pattern used by the satellite to the received signal. The actual signal strength of the GPS signals is very low; they are below the noise level. Due to the digital modulation method, however, the demodulation process results in a very substantial gain in signal-to-noise ratio. This is called correlator gain. When the receiver is closely matched to the transmitter's digital modulation sequence, a signal pops out of the noise. Because the receiver's digital generator is not locked to the satellites digital generator, the receiver then tries to vary the timing or phase of the digital demodulation to get the best signal. A GPS receiver will have several of these correlator processes going at once. This are what is meant by the term "channels."

Let's get back to the satellites for a moment. In GPS there will likely be at the most only ten satellites in view. It would seem that a receiver with more than 10-channels would not be necessary. But, as always, there is more to it than you think.

In order for the receiver to know if it has precisely matched the phase of the original digital modulation and gotten the best signal, the receiver uses three correlators at once. The timing between the three correlators is staggered. The middle correlator output is compared to a correlator that is slightly ahead in time and to one slightly lagging. By this process, the correlator adjusts the timing until the middle correlator signal is peaked--that's how it knows it has properly tuned in the satellite's signal.

Each satellite to be tracked needs three correlator processes or "channels." If there are ten GPS satellites in view a receiver would need at least 30 "channels." Every satellite only remains in view for a limited time, so a GPS receiver must also be looking for satellites that are rising in their orbits and coming into the receiver's view of the sky. The receiver is very smart--it has an ephemeris that tells what satellites should be visible at what time. So a receiver needs to be looking for new satellites to acquire as it is about to lose satellites that are setting out of its sky view. That means a receiver needs a few more "channels." Let's say we need to be hunting for two new satellites all the time, so add six more correlators or channels.

If the receiver is going to get a position fix augmentation signal from the FAA's WAAS satellites, it will need three more correlators for each WAAS satellite it tries to track. So add at least nine more "channels."

Thus a modern only-GPS satellite receiver should have about 45 "channels" to take full advantage of the signals available. If you throw in other global navigation satellite systems like GLONASS, there can be even more satellites in view. Perhaps seven more satellites, so three correlators for each, and now we are up to 66 correlators or "channels."

Is a GNSS receiver that has 66-channels better than one with 32-channels? It might be in the sense that it can track more satellites in view at once. But if the receiver is a GPS-only model, 32-channels should be sufficient.

For a more technical description of the process, see

https://en.wikipedia.org/wiki/GPS_signals

jimh
Posts: 11672
Joined: Fri Oct 09, 2015 12:25 pm
Location: Michigan, Lower Peninsula
Contact:

Re: Use of the Term "channels" in GNSS Receivers

Postby jimh » Fri Oct 21, 2016 1:54 pm

A very good analogy of how a GPS receiver looks for signal was given in a university course I took from Stanford University by the professor who literally "wrote the book" on GPS technology. He likened the process of satellite signal acquisition to something we are more familiar with. The analogy went like this:

Imagine you are driving your car in a strange city you've never been to before. You want to tune in some classical music on the car radio on FM. You will use a process very similar to a GPS receiver. You probably will tune the radio to the low end of the FM broadcast band, and listen for signals. You tune to 88.1-Mhz and see if you hear anything. If you do hear a signal, you stop scanning and listen to it for a moment to see if you recognize the type of music you are looking for. If you don't hear anything, or if you hear something and it is not classical music, you turn the dial to move the receiver up in frequency by 0.2-MHz (the channel spacing of the FM broadcast band), and repeat the process. This is analogous to the action of a GPS receiver. It hunts for satellites by scanning across a range of frequencies around 1575.42-MHz plus or minus the calculated Doppler Shift in small frequency increments. At each scanning interval, it applies the matching PRN code and checks to see if it can hear the satellite it expects to find. It then increments the phase of the PRN code and checks for match. If it finds a signal at all, it then tries to optimize the matching of the local digital string to the signal, resulting an acquisition of the signal or "lock." Once locked on, it then can compute the range--which is another story in itself and won't be described here.

Note that if a GPS receiver does not know where it is located, it cannot make any assumptions about what satellites will be in view. This means it has to search over a wide range of frequencies (to account for satellites approaching or leaving its position) and for all possible PRN codes. A process like this can take as long as 30-minutes or more. A receiver with more correlator processes (or "channels") can throw all of those processes into finding one satellite, and this will greatly speed up the outcome. If we have a receiver with 66 channels, we can use them, three-at-a-time, to search 22 combinations of frequencies and demodulator signal phase at once. It would be like having 22 FM receivers in your car and 22 people tuning them looking for that classical music station. You are very likely to find the desired station much faster using 22 receiver simultaneously than with just one.

jimh
Posts: 11672
Joined: Fri Oct 09, 2015 12:25 pm
Location: Michigan, Lower Peninsula
Contact:

Re: Use of the Term "channels" in GNSS Receivers

Postby jimh » Fri Oct 21, 2016 2:04 pm

The receiver "channels" or correlator demodulators are all implemented in digital circuits. Thus with more powerful microchips comes more correlator processes. Newer receivers tend to be made with more powerful chips, so they have more correlator processes. The development of modern GNSS receivers and the microcircuits that implement them is really quite amazing. The first GPS receivers where huge devices that required a soldier to carry a heavy backpack. Now they're just a chip in your mobile telephone or in your dog's collar.

Early miniaturized GPS receivers only had 12-channels or 16-channels, barely enough to track the three or four satellites necessary to develop a two-dimension (latitude, longitude) or three-dimension (latitude, longitude, altitude) position solution.