SK43793A3 - Gps receiver - Google Patents

Abstract

The receiver is used in the NAVSTAR satellite global positioning system, intended for radio navigation and time transmission, where the transmission from each of a plurality of satellites consists of coded signals in the C/A code, these signals comprising a code carrier frequency and modulated data, the receiver comprises means for receiving coded signals in the C/A code transmitted from at least four satellites and means for processing these signals, which includes an open-loop signal parameter estimator for phase-coherent estimation of several signal parameters, including code delay, carrier frequency, carrier frequency phase, carrier frequency phase acceleration, data value and data delay.

Description

The invention relates to a radio receiver of a global positioning system, in particular a receiver for use in a NAVSTAR Global Positioning System.

Previous qt ^ V tegbolkY

Global Positioning The NAVSTAR system is a satellite-based global navigation and time-based radio navigation system using the US Department of Defense satellite installation. This constellation consists of at least 21 satellites, located on many orbital orbits at predetermined distances above the Earth, and arranged so that at any time and at virtually any position on Earth there are always at least four satellites above the horizon. Furthermore, a satellite control station is provided which, inter alia, controls the very accurate clocks in the satellites, synchronizes those clocks, determines the orbits of the satellites, and supplies orbital orbits to those satellites for retransmission to the users.

Each satellite of the whole constellation carries two unambiguous broader spectrum messages in a straight sequence in phase quadrature on each of the two L-band frequencies. The described receiver only processes messages coded in C / A (Coarso / Aoqulmltlon) on the carrier L.1 (1575) , 42 MHz); however, the principles described also apply to the coded P signals (accurate) on the carrier frequencies L1 and L2. The coded direct-sequence signal having a frequency of 1023 Mcps / s and a coded time repetition rate of 1 kHz modulates the carrier frequency in a binary phase shift (BPSK) writing manner. This wide signal is further modulated by a binary phase shift SOdat signal. The data includes information that allows the receiver to measure the distance between the receiver and the satellite, i. j. data that allows spacecraft modeling of orbits and timing of information relating to accurate satellite clocks. The receiver clock will always have a constant deviation from the satellite clock, so these distance measurements are known as pseudo-distance measurements. Four pseudo-distance measurements are required to perform position dilution - to solve the four x, y, z variables and the local permanent deviation of the clock. To further increase positioning accuracy, multiple measurements can be made - over a period of time (several measurements from the same satellites), for a larger set of satellites, or for a wider range of satellite signal variables such as phase and phase velocity.

The position reporting system using the satellite signals of the NAVSTAR global positioning system is described in our co-pending Australian application No. 6. 63995/90.

Many factors influenced the basic design of the receiver of the global positioning system of the invention. In addition, according to the special requirements of customers, several special proposals were registered.

First, the design of the receiver should be quite flexible for several applications. This is achieved by a modular design in which the individual modules are interchangeable. In addition, the proportionality of the operation of the receiver performed by the Software should be optimized by the fact that the Software is written in a modular form, preferably in a high-level language.

Second, the receiver design should be suitable for many potential environments. This includes the use of suitable limits and densities in the hardware embodiment. More important is the ability to process signals for operation under various dynamic conditions, such as dynamic driving conditions, or vibration and alternate blocking of satellite signals.

Thirdly, a high-performance and inexpensive receiver has minimal manufacturing costs that can be achieved by minimizing hardware, either maximizing the functionality performed by the software or using high-density compression techniques, or both. In addition, the receiver is adapted to use a C / A code, as opposed to a much more complex P code.

Fourth, using a C / A code can be a teaer receiver as accurate as a P code receiver when used in a differential mode. This requires systematic correction of pseudo-distance errors to the receiver at a precisely positioned base station. The receiver must measure the pseudo-distance (and some otherSle parameters) for the distribution and accuracy that are required for an accurate solution, so the measurement and distribution error must be less than the systematic errors that are eliminated.

Fifth, for portability and some comfort, the receiver must be designed to have a minimum weight and volume. Moreover, since it is mostly used as portable, it is equipped with batteries so electricity consumption must be kept to a minimum. Also in this case, the technique of high-density densification is important.

A known Global Positioning System (GPS) receiver of the feedback type uses two loops: a code-controlled feedback loop that selects a code-delay estimate (pseudo-distance) and a phase-controlled feedback loop (usually a Costas loop) for data retrieval. The Cootase loop can also be used to make other measurements of carrier frequency parameters such as phase or phase velocity. In the code-controlled feedback say, a duplicate of the satellite code is used to narrow the received signal, and a data-modulated duplicate of the carrier frequency is used for the coherent demodulation of the narrowed signal. The resulting energy of each of the earlier and later channels is equalized to match the duplicate 3 with the received signal. In the Costas loop, the data is demodulated by coherent deaodulation.

This phase-controlled approach is only approaching optimization. Another problem is sensitivity to loss of fuse and circular shifts in the low-frequency signal in relatively high noise ratio (SNR) situations, such as the presence of high noise or vibration, or under deliberate interference conditions. The code-controlled feedback loop also suffers from poor performance, but is less significant in terms of overall receiver performance. This problem prevails in dynamic conditions. However, even in low dynamic conditions, as with high-precision surveys, the probability of circular displacement becomes essential because the measurements take a long time.

Another type of Global Positioning System receiver uses an open-loop signal parameter estimator. In particular, a generic correlation receiver may be used for the maximum probable estimation of signal parameters. Estimates are made by selecting signal parameters that are processed to resolve the position.

The dynamic performance of an open-ended version is more accurate than a feedback-type version, as the open-ended version is not subject to the circular shift phenomenon and loss of carrier frequency protection. This is because the open-ended receiver has more freedom to change the parameters, to adapt to the requirements and to the different applications. Further, the embodiment of an open-ended receiver can be much cheaper due to the modulation of the basic processing and because the designer is much less constrained and in terms of allowable processing delay.

A highly dynamic receiver of the global positioning system is described in U.S. Patent 4,578,678, issued to W.J. Hurdovl. In contrast to the Hurd receiver, the receiver according to the invention operates in phase coherence. This means a significant improvement in situations with a high noise ratio. Further, without a phase coherent approximation, the carrier frequency phase parameter cannot be measured. This also makes it possible to increase the number of parameters to be used in positioning, which leads to an increase in accuracy. While Hurd used the P code, the receiver of the invention uses the C / A code, which leads to considerable simplification of hardware and processing software. The Hurd design is based on the use of Specific Hardware and Software. Because processing is performed in the Software, the receiver of the invention is much more flexible in use in many different situations by simply changing the algorithms of the Software of the receiver. A further advantage over Hurd's remote is that the processor control has a generator over the full frequency of the local code that allows optimization of the measurement made from the signal parameters.

The global positioning system receiver of the invention makes estimates of code delay, carrier frequency size, carrier phase phase, carrier phase phase acceleration, data values, and data delay. These estimates are made using an open-loop correlation technique. Using multiple estimates of optimal parameters, the receiver is very accurate. An open loop receiver design means that the receiver is relatively immune to high vibrations or high dynamic environments, and can operate reliably with intermittent signal blocking. In addition, for most signal processing performed in software, the receiver is cheap and versatile.

SUMMARY OF THE INVENTION

The above drawbacks eliminate a receiver for use in NAVSTAR's global positioning system for radio navigation and time transmission, wherein transmission from each of a plurality of satellites consists of coded A / A coded signals comprising a carrier frequency of 5d and modulated data , according to the invention, which consists of means for coding signals in the C / A code transmitted from at least four satellites and means for processing these signals, comprising a signal paraset estimator, with open sliders1 for phase coherent estimation of several signal parameters including code delay, carrier frequency, carrier phase, carrier phase phase acceleration, data value, and data delay.

Receivers based on the principle of the invention can be designed for many different applications, i. in different environments and under different dynamic conditions.

BRIEF DESCRIPTION OF THE DRAWINGS

BRIEF DESCRIPTION OF THE DRAWINGS The invention will now be described in more detail with reference to the accompanying drawings, in which: FIG. 1 shows a functional block diagram of one receiver according to the invention, FIG. 2 shows a coarse pseudorange 1 on a correlation diagram; and FIG. 3 is a block diagram of a Kalman filter.

DETAILED DESCRIPTION OF THE INVENTION

For ease of estimation of signal parameters and for avoiding systematic errors, three parameters are not estimated directly, but as a difference between the selected satellites. Before making the Four Estimates of a Pseudorange, the shortest pseudo-distance is nominally denoted by zero, and then the values are marked with a value that exceeds the shortest pseudo-distance. This is shown in order to avoid introducing any errors in positioning, since one of the variables for which the solution is required is a permanent deviation of the clock. This estimation method efficiently adjusts the value of the permanent deviation of the clock to the value of the shortest pseudo-forwarded1. This approach not only eliminates the bends common to each channel, but also reduces the number of bits required to describe the pseudo-distance by more than one quarter, which reduces the size of the message that can be sent to the base station when using aferential positioning. Differences are also estimated for phase, frequency, and phase acceleration parameters. This eliminates estimation errors from various satellites due to the variation and phase of the reference oscillators.

HARDWARE fig. 1 illustrates the construction of a receiver of this embodiment. The function blocks are divided by aedzl hardware and software executed by a digital signal processor. The hardware performs expansion, downward conversion, and filtering functions, all of which are normal functions of the clay performed in the hardware. The downward conversion is a pseudo-one step, i. the carrier frequency is reduced only slightly to the frequency (due to the size of the carrier frequency and the signal bandwidth) above the downlink conversion. This permanent deviation makes it possible to self-detect the resulting values and guide them to the local oscillator where they are to be filtered. Since the permanent deviation is very small compared to the signal bandwidth, it is necessary to create components in the façade and in the quadrature, i. with a phase shift of 90 °.

In this embodiment, encoding and embedding are also performed in the master processor hardware. At this point, another downward conversion is also performed using a hard constrained complex local oscillator. The frequency of the local oscillator is controlled by the processor to allow dri ftu and predictable Doppler shift, thereby increasing the effective range of carrier frequencies that the receiver can process. The mixing is carried out digitally, in a binary manner, the reception signal being severely limited, i. is represented by binary 1 when it is positive, and ”0 when it is negative. A severe limitation results in a loss in high noise conditions (SNR), but a simplification that will allow the hardware to overcome this disadvantage.

The code mixing results in a signal having a bandwidth roughly twice the frequency of the code, i. - 2 MHz until the codes are perfectly matched. This signal is sampled at 5 MHz and then filtered in a low pass filter and sampled? down to 9768 Hz, which is low enough to use processing in Software. This descending sampler filter is designed as an Integration and Write, integrating (summing) many input samples and creating individual output samples. The integration and padding filter has a decreasing functional frequency, according to which it has its first zero at the sampling frequency. That is, there are a substantial number of pass bands above the Nyqulst frequency. Thus, the signal and noise frequency components in this region pass as noise relative to the optical illusion. This problem can be easily solved if so many input samples are integrated twice for the same output sampling rate. This means that the output string is the sum of the previous two integrations and printouts. This second sum, known as the optical illusion filter, is performed in hardware.

about

MAIN PROCESSOR

Up to this point, the functional blinking of the receiver was in operation either continuously or at sampling, where each sample had no particular significance in terms of its location over time. In the main processor, all algorithms are based on the time known as integration period 25 as. Certainly, each Integration Period is performed. otherwise, the unit, algorithms are executed once in 3 or 54 integration periods, as shown in FIG. First

Also, several algorithms ον of the main processor are performed according to the data collection status of the receiver, i. if the satellite was recorded or not. The input signal is converted to the FFT frequency range. Each FFT memory file is validated to size. If no trapped file exceeds the threshold, the satellite signal is considered unrecorded, and the main processor waits for the next Integration period. If the threshold is exceeded, then the memory set at which the maximum is reached is converted down to a zero Hz memory set and a narrow infinite echo (IIR) low pass filter is used in the time domain. The signal at this stage is the carrier wave, extended by 50 Hz.

The expanded phase of this signal is determined and differentiated during the integration period. This makes it possible to detect the transmission of data bits (180 ° phase transfer). Once the transmission chain memory locations have been found, the data can be decoded. This operation retrieves repeated basic information which, once reserved, removes the ambiguity of polarity, and the rest of the data can be decoded. The purpose of this decoding is to create the time of transmission of the data transformations and to select the ephemeris data that describes the satellite orbit with a high degree of accuracy.

Based on these detected transformations, the signal is narrowed by multiplying the duplicate data found. The detection optimum is guaranteed by performing several of these narrowing operations around the detected transforms and by searching for the peaks of the correlation function for the data delay parameter. Once the signal is narrowed to the data, namely by leaving only the carrier frequency, the phase measurement is performed again. Linear adjustment is made to this phase function. This produces an estimate of the lowest squares of the phase frequency <during the integration period. The signal phase is then reversed using a transformation by which a linear line · fit · line is transformed into a real axis. The sum of the real values of this inverted signal is the phase coherent correlation of the signal. One of these correlation values is generated for each integration period. The set of correlation thresholds may also detect that the satellite has not been recorded.

To perform a pseudo-distance measurement, it is necessary to estimate the code delay. In another embodiment of the invention, the pseudo-distances are not measured directly, but as differences of Bedai satellite®. One satellite is marked as zero and the other satellites are marked as Bedzl differences by this pseudo-distance and the pseudo-distance in question. Measurement of differences makes the receiver simplified so that many accurate clocks are not needed and the connection to the base station is easier.

The code delay is estimated by setting the code generator to three different phases for three successive integration periods. These three phases are blocked (i.e., local and satellite codes are aligned), part of the chips earlier and part. Chips LaterSle. From these three code phases, the correlation function can be reconstructed. The optimal way to perform the correlation function reconstruction and peak search is to correlate the measured correlation function, applied at three points, with the expected correlation function, which is measured in advance. Using this double correlation technique produces rough estimates of pseudo-distances. The use of this second correlation is necessary for optimizing the use made of the three integration times contributing to the measurement. Each of the three correlation samples is generated by observing the signal for only one Integration Period, but the total measurement time is three Integration Periods. Noise on samples is independent between sample · 1. Therefore, the sampled correlation curve is optimal in part. Correlation, or correlation with the expected correlation function, maximizes the signal-to-sum correlation ratio, thereby restoring process optimality and allowing sample interpolation.

In order to meet the requirements for accurate measurement of pseudo-range differences 1, phase differences, frequency differences and phase acceleration differences, the signal must be observed for a much longer time than the integration period (the length of which is set by the signal detection requirements). In other words, rough estimates must be combined in several optimal ways to produce a more accurate estimate. The method of combining the estimates is to use the Kalman filter. The input data to the filter are estimates of pseudorange, frequency and phase difference, and correlation pseudorange from base station. This filter models the receiver position, satellite and receiver movement, and drift of the receiver oscillator. From these models and input data, the filter not only creates just aligned, but incorrect pseudo-transmissions for transmission to the base station, the lie also estimates the position of the receiver relative to the user. These estimates are presented to the user in a Standard Reference System that needs a geometric transformation.

CORRELATION

The Maximum Probability Estimator (MP) is a correlation detector. The signal parameters in question are the code value, the signal delay (measured using the code delay and the data delay), the carrier frequency size, the carrier frequency phase, and the data value. The receiver must efficiently perform the six-dimensional correlation and look for its peak to determine optimal estimates in all parameters, to derive the optimal position resolution.

Two other parameters that arise in the estimation process are the local oscillator drift and the code generator frequency. The oscillator drift is important because all measurements performed at a given time in the receiver are performed in relation to one oscillator that is used for the local oscillator and the sample clock. The output of the Kalaan filter is used to make an estimate of the frequency of the oscillator that is used to correct the frequency setting of the code generator. The code generator frequency correction allows close tracking of the satellite code, optimizing the corre- sponding shape and value, and maximizing both 1 and 20 m noise situation (SNR) situations. It is not estimated as a separate parameter (i.e., the correlation is not performed against it) because it is directly proportional to the carrier frequency of the satellite signal, which is already estimated.

I

CODE VALUE

Selecting a code value simply means selecting a satellite code to be used for each channel of the receiver. Each satellite has a timing code of 1023 chips, which is repeated for each time frequency of 1 kH2. In the receiver, each satellite channel creates a duplicate of the satellite code from which the data is received. The locally generated code is multiplied by the received satellite signal. If the local code is extended so that the codes are aligned, the received signal is narrowed from the extended 2 MHz bandwidth spectrum to 100 MHz data bandwidth.

DATA VALUE ESTIMATES

The data value is determined by searching for data conversions. The correct data polarity is created in the data decoding algorithm. For signal processing, the value of the data is estimated to be inverted or non-inverted, relative to the actual data, with 1 or ** 0 ** being determined at a later stage. The transforms are placed over time by filtering the carrier frequency phase of the modulated data using a transform-adapted filter. Peaks in the comparison filter output occur in the data conversion area. Comparative filtering is the optimum way of placement of transformations and is accurate even with a small number of samples. For independent placement of the transformations given by the initial estimate from the comparison filter.

code periods can be used. Since the comparison filter places transforms into several samples (i.e. hundreds of microseconds), and the data transformations always correspond to code period variations that occur every 1 ms, it is possible to unambiguously place the transforms if the code period bounding in the integration period is known. This method is used in the receiver.

When data transformations have been placed and decoded, the information in the data contributes to further measurements. The time information in the data is used to set the weekly time and allows the receiver to calculate the edge transmission time used to measure the pseudo-distance. Ephemeride data is used to accurately position the satellites given the transmission time.

PHASE, PHASE SPEED (FREQUENCY) AND PHASE ACCELERATION ESTIMATES

The frequency estimation in each satellite channel is simply the sum of all downlink transformations used for coherent signal detection. Namely, there are four such downlink transformations - in analog hardware (via * local LO oscillator), in digital hardware, and two in the main process (one upstream of the IIR filter and one phase reversal path, both in the time domain). Each of these transformations is nominally smaller than the previous one. The measured value of each is then proportional to the LO frequency. This is because the downlink frequency performed in the hardware and in the processor are all at a sampling frequency that is derived from the same source as the LO. In other words, drift LO causes multiple errors in this measurement. However, by estimating the frequency differences, the permanent deviation of the first LO frequency is eliminated, and drift only multiplies the downward variables performed in the Software.

Optimization adjusted to frequency estimation only affects the last two downward transformations - the first two are constant in rotation. The FFT transformation is a frequency correlation, so estimating the frequency for the first downlink conversion in the main processor through the FFT transformation provides an estimate of the MP. Phase processing using averaged frequency estimates and the phase can be represented as an optimal interpolation of complex FFT transformation estimates.

If the phase is measured within the integration period, the measurement is also optimal. Phase contribution by the first LO is eliminated using phase differences. Both the downlink in digital hardware and the downlink FFT contribute to the zero phase of the signal within the integration period. This is due to these downward oscillators, which have integral cycles in the integration period and are initially zero. This means that the phase estimate is the average value estimated at the phase processing. To eliminate phase ambiguity, the phase estimate shall be expanded relative to 2.

about

Phase acceleration is not optimal because it does not use all available Information. It uses only earlier and later integration periods, both of which will have a substantially lower noise ratio (SNR) than the point.

LEAST ESTIMATION (PSEUDOVENCY)

The estimation of the delay is much, but most is the most basic cat. As mentioned above, only differences between pseudoodhadaml are estimated. The estimate requires input both from a code delay estimate (which has an ambiguity of 1 ms but can be estimated more easily because the code is known) and a data delay (which is not ambiguous but unknown). The data conversion occurring in each channel is used for timestamping the pseudo-distance measurement. This performs placing the data conversion into one sample as described for estimating the value of the data. Data conversion always coincides with code transmission more complex than other estimates, to determine the location of the reception period. Interpolation between samples is made possible by sampling or sampling the phase of the code at the beginning and end of the Integration period, for sub-chip accuracy. When it is determined that the local code frequency is blocked to the received code frequency to a degree that can be estimated, the point between the samples where the conversion occurs may be interpolated. If the local code was accurately compared (in delay) with the received code and there was no noise in the system, this interpolated pseudo-distance measurement could be accurately corrected.

Maintenance requirements; blocking code and execution

However, even an accurate estimate, despite the noise in the system, determines that two additional processing steps are required. The first involves finding the correlation peak of the delay parameter, and the second involves combining a series of measurements made at the observation period to ensure that accuracy requirements are met.

Pseudo-distance estimates made using the above method were called gross pseudo-distance !. These estimates are not optimal because there is no correlation between these estimates and the correlation in the delay parameter. This is addressed by making three such estimates in the three subsequent integration periods, each estimation being made for a different phase of the code. One of the integrations is performed with the code set at the best estimation of the code phase (i.e., most closely estimated to be blocked). The other two are performed at the moment of a portion of the earlier chip and a portion of the later chip, thereby forming the measurement set shown in FIG. 2. To find the exact peak of the correlation, and thus to correct the pseudo-distance measurement made at best estimate, it is necessary to use these three points to reconstruct the shape of the correlation function. An optimal way of reconstruction is to correlate three points with the expected correlation shape (shown by the dotted in Figure 2). This pseudo-distance estimation was called "interpolated pseudo-latency1, and is an optimal estimate because correlation15 was used (twice, because two estimates were made - one correlation shape and one correlation peak) when making the estimation. However, despite the optimality1 of each estimate, the pseudo-distances are not yet accurate. Several (namely at least 18) of these interpolated pseudo-distances are fed to the Kalman filter, which produces final estimates of pseudo-distances and position. Phase and frequency estimates, such as inputs to the Kalman filter, also contribute to improving estimation of pseudorange.

KALMANOV FILTER

The function of the Kalman filter is to monitor the state of the linear dynamic system by processing the observations of the least squares repetition. The state of the system is described by a vector of system variables that must be related to each other by a set of known linear equations. The state vector usually consists of one or more sets of derivatives, and therefore the equations that concern them are differential equations describing the dynamics of the system. Observations are made by measuring observable quantities, which must be known by linear functions of the state variables.

Fig. 3 shows the Kalman filter function. The observation vector Y consists of three pseudo-distance differences and three phase differences. The state vector X consists of the receiver position, the speed and the acceleration in three coordinates, and the rate of change of the permanent deviation of the local clock. Hat shapes ** are estimates or predictions.

The conversion matrix O is multiplied by X to obtain a state vector estimate for the next time interval. The loop is known as the state transformation loop. Every N conversion and observations are performed and a correction is made to X. Thus, between observations, the transformation matrix models the receiver's dynamic behavior prediction.

Corrections are made by comparing the predicted observation vector Y with the observation vector Y, and this results in the processing of the predicted error vector. The error vector is multiplied by the. The Kalman gain matrix K. The Kalman gain matrix K is calculated by a random modeling technique including a state covariance matrix P, a covariance matrix g of observed errors, a covariance matrix Q The sum of the system, and an observation matrix M. The observation matrix M shows the relationship of the measured vector to the state vector.

The presence of designation of the matrices defining their model (i.e., 0, Q, and M) indicates that these matrices must be periodically renewed because they are not time constant. These matrices O, Q and M depend on the positions of the satellites and the receiver. In addition, the matrices Q and R can be estimated from observations in different ways, allowing the filter to adapt to changes in the dynamic behavior of the receiver. Another way to adapt the filter is to multiply it by a scalar factor relative to the estimation curve related to the predicted error. The scalar factor is 1 + et, where <t is linearly related to the matrix curve of <(Y-Y) (Y-Y) T> - (MPMT + R).

Observation in the context of this Kalman filter is input from the interpolation algorithm of pseudo-distance and input from frequency and phase estimation devices. After the required amount of observation has been made, the estimated position value can be directly output from the predicted status vector to the user. Similarly, estimates of pseudo-distance differences 1 to be transmitted to the base station in differential operation can be directly selected from the predicted observation vector. When pseudo-distance corrections are available from the base station, they may be used to determine a more accurate position by feeding them into the filter as shown in FIG. Third

POSITIONING ALGORITHMS

The positioning algorithms used in this receiver are performed in the Kalsan filter.

It will be appreciated that the invention is not limited to the described embodiment. Many changes and modifications can be made within its broad concept. The receiver can be upgraded in any way.

The generic concept described above concerns receivers with four or more channels, known as multichannel receivers.

T

With little effort, this concept can be extended to applications such as for three-channel ship receivers or single-channel multi-receiver receivers. For single-channel receivers, the important difference is that absolute measurements of pseudo-range 1, phase, etc. are needed rather than differences.

Further changes and variations can be made to ensure special use cases and special environmental conditions.

Claims (13)

1. A global positioning system receiver for use in a satellite-based global positioning system NAVSTAR, intended for radio navigation and time transmission, wherein the transmission from each of a plurality of satellites consists of coded signals in a C/A code, said signals comprising a code carrier frequency and modulated data, characterized in that it comprises means for receiving coded signals in a C/A code transmitted from at least four satellites and means for processing said signals, which comprises an open-loop signal parameter estimator for phase-coherent estimation of a plurality of signal parameters, including code delay, carrier frequency, carrier phase, carrier phase acceleration, data value and data delay. 2. The receiver of claim 1, wherein the open-loop estimator is a correlation detector. 3. A receiver according to claim 1 or 2, characterized in that it comprises means for selecting at least four satellites from which signals encoded in the C/A code are to be received. 4. A receiver according to claim 3, characterized in that at least one of the signal parameters is estimated as a difference between one and the remaining selected satellites. 5. A receiver according to claim 4, characterized in that the difference estimate is made from the code delay. 6. A receiver according to claim 4 or 5, characterized in that the difference estimate is made from the magnitude of the carrier frequency. 7. Receiver according to claim 4, 5 or 6, v y z n and oh u- ing with t I think that estimate the difference is made from phase bearing frequencies. 8. P r i t i c a according to claim 4, 5, 6 or 7, in y from on a- characterized by the fact that the difference estimate is made from the phase acceleration. 9. A receiver according to any one of the preceding claims, characterized in that it comprises means for extracting information from signals transmitted from a satellite to enable the receiver to measure the distance between the receiver and said satellite. 10. A receiver according to any one of the preceding claims, characterized in that it comprises means for extracting information from signals transmitted from a satellite to enable the receiver to measure the permanent clock drift in the receiver and to refine the clock in said satellite. 11. A receiver according to any one of the preceding claims, characterized in that the signal processing means comprises a digital signal processor. 12. Receiver according to claim 11, characterized in that the functions of the receiver are performed in hardware and software. 13. The receiver according to claim 3, characterized in that the receiver has at least four channels, each channel corresponding to one of the selected satellites. 21. 14. A receiver according to any one of the preceding claims, characterized in that the receiver is of modular design, the individual modules being interchangeable. 15. A receiver substantially as described and illustrated in the accompanying drawings.