Sect_4.1.3

4.1.3 GPS Receiver Design:

GPS SIGNAL PROCESSING TECHNIQUES

The antenna preamplifier of a GPS receiver generally converts the incoming signal (see Figure 1 below) to a signal of a lower frequency. This intermediate frequency is obtained by mixing the incoming signal with a pure sinusoidal signal generated by the local oscillator (the quartz "clock"). The frequency of this beat frequency is the difference between the original (doppler-shifted) received carrier frequency and the local oscillator. The intermediate or beat frequency is then processed by the signal tracking electronics.

Figure 1. The incoming signal.

Frequency 1 = L1 carrier + P* code + C/A code + Navigation Message
Frequency 2 = L2 carrier + P* code + Navigation Message
(* Changed to Y code under AS)

There are several classes of signal processing techniques that can be employed to make pseudo-range or carrier phase observations, as well as several proprietary implementations of tracking technologies. However, it is beyond the scope of these notes to discuss in detail the electronic circuitry that is required, the reader is referred to such engineering books as KAPLAN (1996). At a conceptual level it is only necessary to deal with:

the two most commonly used signal processing techniques: "code-correlating" and "codeless" approaches, and
some details of the processing architecture within the receiver, and in particular the characteristics of the "tracking channel(s)".

Code-Correlating vs Codeless Receivers

Code-correlating receivers employ tracking loops to extract the necessary measurements and Navigation Message data from the beat signal. A typical GPS receiver contains two types of tracking loops:

the delay-lock, or code-tracking, loop, and

the phase-lock, or carrier-tracking, loop.

The delay-lock loop is used to align the PRN code sequence (C/A or P code) that is contained in the signal from a satellite with an identical PRN code generated within the receiver. A correlator in the delay-lock loop continuously cross-correlates the two code streams, time shifting the receiver generated stream until alignment is achieved. The time shift is then the pseudo-range observation. Once the code-tracking loop is aligned, the PRN code can be removed from the satellite signal (section 3.2.2). The stripped signal then passes to the phase-lock loop where the satellite message is extracted. Once the local oscillator is locked onto the satellite signal it will continue to follow the variations in the phase of the carrier as the satellite-receiver distance changes. The integrated carrier beat phase observable is obtained by simply counting the whole elapsed cycles (by noting the "zero crossings" of the beat wave) and by measuring the fractional phase of the locked local oscillator signal (section 3.2.3 and section 6.1.1).

The carrier beat phase can be measured by another method besides this code-tracking / phase-tracking combined technique. The basic "codeless" or "squaring" technique takes the incoming signal and multiplies it by a copy of itself (section 3.2.2). (Other variations of this also exist. The reader is referred to VAN DIERENDONCK (1995) for further details on tracking channel terminology.)

The organisation of a code-correlating tracking channel and that of a squaring tracking channel is shown schematically in the two Figures below (and discussed in, for example, WELLS et al, 1987).

Figure 2. Code-correlating tracking channel

Figure 3. Squaring (codeless) tracking channel.

There are several advantages and disadvantages to GPS code-correlating and codeless receivers:

Only code-correlating receivers have access to the satellite Navigation Message. Either the P or the C/A code is required to despread the arriving signal without destroying the message, hence squaring receivers need an external satellite almanac in order to operate the tracking channels, and require externally-provided satellite ephemeris data for subsequent data processing.
Codeless receivers must be synchronised with each other and UTC (or GPS Time) before observations begin. Code-correlating receivers can synchronise themselves with GPS Time at a point with unknown coordinates after acquiring signals from a minimum of four satellites, and carrying out a pseudo-range navigation solution.
Code-correlating receivers enjoy a significant signal-to-noise advantage because of the inherent insensitivity of spread-spectrum signals to noise in a decoding receiver.
Code-correlating receivers are capable of both pseudo-range and carrier phase measurements. Non-military code-correlating receivers use the C/A code, which is only available on the L1 frequency. Dual-frequency ionospheric delay correction of either the pseudo-range or carrier phase data is only possible if measurements can be made on the second (L2) frequency.
Codeless receivers can access both the L1 and L2 frequencies without knowledge of the PRN codes. This is particularly important under the Anti-Spoofing regime that exists at present, unless special tracking techniques are developed to overcome this handicap.
Cycle slips in pure squaring receivers may sometimes be harder to repair as the wavelength of the phase data is half that of the carrier frequency for certain GPS receivers.
Single frequency C/A code receivers are the cheapest GPS receivers on the market.

Nowadays the tracking channels are entirely digital in design, making GPS hardware smaller, lighter, and less expensive to build. Using analogue-to-digital conversion techniques a modern receiver usually consists of multiple digital tracking channels fed by a single receiver frontend (see Figure 4 below). The characteristics of tracking channels, both the varieties used in earlier GPS receivers and those used in presentday equipment, are briefly discussed below.

Figure 4. Digital GPS receiver design. (CLARKE, 1994)

Hardware and Software Components of Tracking Channels

A GPS receiver may be described as "continuous" or "switching", depending on the type of tracking channel(s) it uses.

A continuous-tracking receiver consists of dedicated hardware channels (Figure 5 below). Each channel tracks a single satellite and maintains continuous code and/or phase lock on the signal. Each channel is controlled and sampled by the receiver's microprocessor with input/output operations being performed fast enough so that tracking is not interrupted. Continuous-tracking receivers may enjoy a signal-to-noise advantage over switching receivers in that the satellite signal is continuously available and may be more frequently sampled. They are therefore particularly suited to high dynamic (kinematic) applications. A further advantage is a potential redundancy capability for should one of the hardware channels fail, it may still be possible to obtain sufficient data from the remaining channels. One disadvantage of a multi-channel receiver is that the differences in signal path delay in the channels, the so-called inter-channel biases, must be well calibrated (though in reality this is not a significant problem). This is the most common channel architecture used in survey-type GPS receivers.

Figure 5. RF configuration with continuous tracking channels.

A switching receiver has one or more hardware channels. Each channel samples sequentially the incoming signal from more than one satellite (see Figure 6 below). Code and/or carrier phase tracking for the individual signals is controlled by software (or, more correctly, the "firmware") within the microprocessor. As a result, greater demands are placed on the microprocessor in a switching receiver and its programming is necessarily more complex -- in effect hardware complexity is exchanged for software complexity. (The term pseudo-channel is sometimes used to identify software tracking channels.) Depending on the application and the environment, a switching receiver may be more susceptible to cycle slips than a continuous-tracking receiver. If the cost of hardware components is a significant factor in determining the sale price of a GPS receiver then, in principle, the cost of a switching receiver should be less than a continuous receiver. The reality is that multi-channel GPS receivers can also be found in almost all low-cost GPS receivers.

Figure 6. RF configuration with switching tracking channel(s).

There are three basic types of "switching" receivers, distinguished by the time required to sequence through the signals tracked by a particular channel.

A multiplexing channel is one for which the sequencing time to sample all satellites assigned to the channel is equal to 20 milliseconds, the period of one bit in the satellite Navigation Message (section 3.3.1). The sampling is organised in such a way that no message bit boundary is spanned by any tracking interval. In this way, the messages from all satellites phase tracked by the channel can be read at the same time as the measurements are made. A multiplexing channel can be used to obtain both L1 and L2 data by alternating between the frequencies every 20 milliseconds. The early Texas Instruments TI-4100 GPS receiver had a single multiplexing hardware channel that was used to sample both L1 and L2 signals.

If a channel switches between signals at a rate which is asynchronous with the message bit rate, the channel is referred to as a sequencing channel. A fast sequencing channel is one which takes the order of fractions of a second to perhaps several seconds to sequence through the signals. A slow sequencing channel may take many seconds. A single sequencing channel would lose bits in a particular satellite message during those intervals spent sampling the signals from other satellites. Consequently, sequencing receivers may have an extra hardware channel just for message decoding. Alternatively, the Navigation Message must be decoded before the receiver starts the tracking cycle for real-time positioning (remember that the message only changes once an hour).

Figure 7 below illustrates the tracking coverage for different types of channels.

Figure 7. Tracking pattern for different types of channels.

The configuration of channels is selected so that, in the case of a navigation receiver, a minimum of four satellites can be tracked at the same time. This may call for a single switching channel, or two switching (between two satellite signals) channels, or even five channels (one for calibrating the other four primary tracking channels, and decoding the Navigation Message). Nowadays even low-cost GPS receivers typically have 12 channels.

In the case of GPS surveying receivers it is now recognised that it is very advantageous to track as many satellites simultaneously as possible (the so-called "all-in-view" tracking capability is preferred). To increase the number of simultaneously tracked satellites but still limit the number of channels (each requiring a "board" in the first generation of analogue GPS receivers), a combination of continuous and switching channels were sometimes used within the same receiver. An example of a hybrid geodetic receiver was the early Wild WM102, which was able to track six satellites on both frequencies. However, with the cost of electronic circuitry falling dramatically (due to the fabrication of digital tracking channels on VLSI chips), the latest generation of geodetic receivers have many continuous-tracking channels, some code-correlating and some employing a codeless tracking technology, or some variation of these. Configurations of 12,16, 24 and even 48 channels are now used.

Back To Chapter 4 Contents / Next Topic / Previous Topic