Patent Grant
10921459
This invention relates to GNSS navigation receivers, and, more particularly, to GNSS receivers that use complex Code Shift Keying signals.
Global navigation satellite systems (GNSS) signals are often used in many practical applications to precisely determine a position. Progress in the field of satellite navigation requires navigation receivers capable of receiving signals in ever wider frequency bands, the signals being more complex in processing. Among such complex signals, Code Shift Keying (CSK) should be considered [1]. These signals benefit from better noise-immunity, or from spectral efficiency at the same noise-immunity. This is important for implementation satellite-based augmentation systems (SBAS) for broadcasting a navigation data that supplements data in satellite ephemeris and thus sufficiently improves navigation accuracy. However, CSK demodulation is much more complicated than that of phase-manipulated symbols and in general needs calculation of a full correlation function within the range of all possible PRN-code offsets for each received symbol. The reception of CSK-signals becomes even more complex (and hence, more expensive), if it is accompanied by range tracking (it is also called delay lock loop, or DLL, in the literature), rather than only data demodulation. In this case, a discriminator outputting a range tracking error needs a more detailed correlation function, for example, in increments of half chip, while for a standard demodulator the correlation function with one chip-increments is sufficient.
There is known a device [2] calculating correlation function for receiving CSK-signals using a set of correlators, each of which calculates a convolution of the received signal and reference signal replica (usually a pseudo-random noise code) shifted by a certain value, different for each correlator. A drawback of such a device is the use of many correlators, wherein the reference replica is stored or generated.
In [5], a method and apparatus are disclosed where additional correlators are applied. The correlators are adjusted to an advance code position, i.e., shifted forward (Early correlator), and to a delayed code position, i.e., shifted backward (Late correlator), in a value less than the chip duration. These correlators are employed for tracking code position (the term “code phase” is also used) via a so-called early-late discriminator. Half chip is an optimal shift. Implementation of this method to demodulate and track CSK-signal delay with P bits per symbol requires calculation of 2*2P+1 correlation points for each symbol being received. For instance, if there are P=8 bits for each symbol, one needs 513 correlators (256 for data demodulation and 257 for DLL early-late discriminator)—this can exceed the total number of correlators for receiving all other satellite navigation signals.
Another device known as a “sliding” or “parallel” correlator [3] calculates correlation function between the received signal and a reference replica using an array of accumulators updated by the same calculation module. Along with the accumulators, such an apparatus includes memory elements to store the input signal and reference replica. These memory blocks have a lot of outputs equal to a number of simultaneously calculated multiplication and adding operations. A typical configuration for these memory blocks is an implementation in the form of a multi-tapped delay line. The availability of the memory block storing the input signal and computation operation at rates exceeding the chip rate of the input signal allows this device to store only one copy of the replica and employ only one computational module for correlation calculations, thereby making this design cheaper than an equivalent set of separate universal correlators.
A correlator can be a circuit where the correlator's output is calculated as a convolution of a single-bit sample of the input signal and single-bit values of code sequence (replica) (see [4]). The advantage of this correlator's configuration is additional lowering of the cost of the equipment due to using an XOR operation, instead of multi-digit multipliers. Together with the above circuit (see [3]), this construction allows implementing the most efficient method of calculating the correlation function. A main drawback of this is some worsening of noise immunity by 2 dB.
There are known devices for reception of navigation CSK-signals in which correlations of the input signal and replica are calculated by small-digit multipliers (ultimately, single-bit), with memory blocks whose size matches the length of the multipliers' input as well as the number of parallel calculated correlation points. To demodulate and track CSK-signals (specifically to track the unmodulated code phase of the signal), 2*2P+1=2*N+1 correlator outputs are calculated in parallel, where N is the range of one CSK symbol data values and P—number of bits per symbol. The doubled number of correlators allows calculation of responses for Early and Late positions. An advantage of such devices is more efficient demodulation and tracking of CSK signals, compared with devices based on universal correlators. A disadvantage is the doubled number of calculators compared to similar devices for demodulating information data only. In addition, there is a 2 dB energy loss if single-digit representation of the input samples in the memory blocks is used. At the same time, single-digit presentation of the input samples is more efficient, since it is implemented on the basis of memory blocks with smallest capacity, which is a further advantage.
In [6], a method and apparatus are disclosed where frequency-phase adjustment of carrier, as well as CSK-signal delay tracking, are implemented by re-calculation of similar tracking parameters of another satellite signal. As such, the satellite signals can be QZSS LEX L6D/L6E CSK-signal and GPS L1 C/A BPSK-signal respectively. This method benefits from demodulation and tracking of CSK-signal with the help of N correlators for the CSK-signal and just one triple correlator for tracking another signal, i.e., N+3 correlators are used instead of 2*N+1 correlators. A drawback of this method is appearance of CSK-signal tracking error due to different conditions of passing CSK-signal and another signal (for example, if the signals are broadcast on different frequencies or from different satellites).
Thus, the known methods for CSK demodulation and code phase tracking with 2P bit-symbol require calculation of the correlation function for 2*2P+1 points which causes serious hardware overhead or—in case of more cost-conscious implementations —they cause an energy loss that can be a critical disadvantage for SBAS-enabled devices.
Accordingly, the present invention is related to a system and method for demodulating and tracking CSK modulated signals that substantially obviates one or more of the disadvantages of the related art.
In one aspect, there is provided a method for demodulating and tracking of CSK-modulated signals comprising the steps of receiving a CSK-modulated signal at a plurality of correlators, the CSK-modulated signal including a transmitted symbol; convolving the CSK-modulated signal with a preset reference replica of the CSK-modulated signal for N shifts (D1 . . . DN) of a PRN code relative to an internal receiver clock corresponding to N possible code shifts in the transmitted symbol, and M other code shifts (T1 . . . TM), where M<N, that satisfies D1−ΔT≤Ti≤DN+ΔT for all Ti from (T1 . . . TM); selecting one of the N shifts (DMAX) where the convolution is a maximum; adjusting a CSK-signal Delay Locked Loop (DLL) based on a ratio of convolution values for the shifts DMAX and Ti for all Ti for which a modulus |DMAX−Ti| is smaller than a predefined threshold; and determining a value of the transmitted symbol based on DMAX.
Additional features and advantages of the invention will be set forth in the description that follows, and in part will be apparent from the description, or may be learned by practice of the invention. The advantages of the invention will be realized and attained by the structure particularly pointed out in the written description and claims hereof as well as the appended drawings.
It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory and are intended to provide further explanation of the invention as claimed.
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the principles of the invention.
In the drawings:
Reference will now be made in detail to the preferred embodiments of the present invention, examples of which are illustrated in the accompanying drawings.
The objective of the present invention is to provide an efficient and interference-immune method of demodulating CSK signals from a satellite and its further delay (range) tracking. The method is based on the fact that DLL for a GNSS signal at typical navigation conditions has better interference immunity, compared to immunity of demodulation of information bits and to the carrier PLL. This allows the DLL to be adjusted less frequently, and intermediate values for loop adaptation to be calculated at a lower quality. The following important fact should be also noted: a DLL for the CSK-signal in the invention is eventually adjusted according to the error signal being generated directly in the CSK demodulator. The method is as follows:
1) A CSK-modulated signal is fed to the correlators' input.
2) A convolution of the CSK-modulated signal and a predetermined replica is calculated for N replica code offsets (D1 . . . DN) relative to the receiver internal time, where the offsets correspond to N possible code shifts in the transmitted symbol, and for M (where M<N) different code shifts (T1 . . . TM), where each Ti from (T1 . . . TM) satisfies: D1−ΔT≤Ti≤DN+ΔT
3) One of N offsets (DMAX) is further selected such that the convolution at this offset is maximal
4) If there are one or several Ti for which the modulus |DMAX−Ti| is smaller than a threshold, then such Ti are selected and the DLL for the CSK signal is adjusted according to the convolution values for the offsets DMAX and the selected Ti
5) A value of the modulated symbol is determined in accordance with DMAX.
The proposed sequence of steps is shown in
In a typical embodiment of the invention, all the correlators are implemented in digital form so that they perform a convolution of a digitized incoming signal and a digitized replica, and form a digital output. The convolution is performed as a sample-by-sample multiplication of the incoming signal by replica elements with accumulation of the multiplication results. In some embodiments, each multiplication result is additionally multiplied by a complex vector corresponding to current carrier phase. Accumulation of the multiplication results starts and stops in accordance with current correlator configuration. In some embodiments, the duration of convolution of all the employed correlators is identical, whereas the start time of convolution depends on the correlator configuration. A delayed start of the convolution of a fixed replica code with the incoming signal is introduced to apply the required Di or Ti code shift. In other embodiments, the required code shift is implemented via a convolution of the incoming signal with a shifted code sequence, whereas the convolution start time is fixed. The code sequence in some embodiments is shifted cyclically. In other embodiments, the original code length is truncated so that a wraparound of the code sequence does not occur.
In some embodiments of the invention, intervals between shifts (T1 . . . TM) can vary over time. In some other embodiments, correlator outputs for shifts (T1 . . . TM) can be calculated less accurate than that of correlator outputs for shifts (D1 . . . DN). For example, this can be done by decreasing the size of multiplier inputs and/or outputs used in calculating convolution. Less accurate calculations of correlator outputs for the intermediate (additional) shifts, as well as DLL adjustments not for each CSK-symbol result in worsening DLL interference immunity, but make the apparatus cheaper. If M is much smaller than N, a decrease in DLL interference immunity can be sufficient.
To increase DLL interference immunity, the feedback loop filter in some embodiments is made more narrowband. At the same time, to keep DLL capability in dynamic, some additional information is transmitted for CSK-signal DLL as follows: a receiver receives CSK-signal and another signal with a modulation different from CSK. This signal is generally broadcast on a frequency different from the carrier frequency of the CSK-modulated signal but is synchronized with the CSK-modulated signal. This differently-modulated signal is used to predict changes in pseudoranges to the satellite transmitting CSK-modulated signal, caused by the receiver movement. Then the DLL of the CSK-modulated signals is additionally adjusted according to the prediction. Some embodiments comprise such predictions based on changing delays for the non-CSK-modulated signal. The delays determine changes in pseudoranges to the satellite transmitting this signal. Other embodiments comprise predictions on the basis of receiver's velocity vector calculated by the non-CSK-modulated signal. In a number of embodiments, the predictions for changes in pseudoranges to the satellite transmitting non-CSK-modulated signal are added to the DLL error of CSK-modulated signal.
In the proposed method, a decision scheme of CSK-symbol demodulation is based on the maximum of scalar correlator outputs. In some embodiments scalar correlator outputs are determined as a sum of squares of convolution quadratures. In other embodiments, scalar correlator outputs are a signed value of one of the quadratures. In other embodiments, scalar correlator outputs can be calculated both as a sum of squares of quadratures and as the signed value of one quadrature. In such embodiments, the selection of one or another scalar output, according to which the information encoded in the symbol is defined, is made by estimation of carrier phase error at carrier PLL output. If this averaged PLL error is small enough (smaller than a threshold), the decision scheme works using maximum value of one of signed correlator quadratures. Otherwise, the decision scheme is switched to the mode based on the sum of quadrature squares. This principle is illustrated in
It should be noted that a block scheme of a CSK-symbol demodulator of
From
An example of embodiment of the invention can be considered for reception QZSS LEX L6D/L6E CSK-signal. Since the L6D signal is multiplexed with the L6E signal, to demodulate two multiplexed 8-bit CSK symbols in parallel as many as 512 correlators configured to (D′1 . . . D′512) code shifts should be employed. The difference between adjacent D′i and D′i+1 should be equal to a time interval between adjacent chips in the multiplexed signal. The correlators perform convolution that forms a single quadrature output, which is sufficient if the whole signal energy is concentrated in one quadrature. Additionally, a small number of correlators with two-quadrature outputs are to be implemented. Demodulation of the L6D and L6E signals are two independent processes. Therefore, they can be considered separately. Demodulation of, for example, the L6D signal, requires 256 single quadrature correlators configured to (D1 . . . D256) code shifts with a shift increment equal to a time interval between consecutive L6D chips. The number of additional correlators with two-quadrature output can be for instance equal to 16. This value can be increased or decreased with no significant effect. Code shifts (T1 . . . T16) of two-quadrature correlators can vary relative to each other, which is used to effectively utilize them during known data reception. Let ΔT be half length of L6D chip. Then, Di+1−Di=ΔT*4 due to multiplexing with L6E chips. For the additional two-quadrature correlators the following two conditions need to be met:
1) min (|Ti−Dj|) should always be equal to ΔT if Ti belongs to (T1 . . . T8) for example, and
2) min (|Ti−Dj|) should always be zero for Ti belonging to (T9 . . . T16).
For time intervals, when known data is broadcast, one of the code shifts Ti in conditions above is set in accordance with known data value. Otherwise, the shifts Ti from above conditions could be arbitrary.
For time intervals, when the first condition above is met for code shift DMAX, which corresponds to the max signed output of 256 single-quadrature correlators, the output value for code shift DMAX and one or two values of correlator outputs with neighboring Ti are fed to early-late DLL discriminator. The early-late discriminator in this particular example is implemented as
where Early value is signed output of correlator with code shift Ti=DMAX−ΔT, Late value is signed output of correlator with code shift Ti=DMAX+ΔT, Prompt is signed output of correlator with code shift DMAX. The limited number of implemented shifts Ti leads to possible absence of Early or Late value or both. The absent values are substituted by zero. The discriminator output after averaging is fed back to all the correlators and tune (D1 . . . D256) and (T1 . . . T16) code shifts.
For time intervals, when the second condition above is met for code shift DMAX, the two-quadrature output for code shift DMAX is fed to the carrier PLL. Information from the carrier PLL is converted to the single-quadrature correlators' settings, so that correlators with code shift corresponding to the received CSK-symbol accumulate output with no energy loss of the CSK signal in the single quadrature, and the value of the quadrature is positive.
Both the DLL and carrier PLL in the above example are also assisted by code and carrier phase forecast obtained from velocity vector calculated by the navigation receiver. The assisted information allows to robustly track carrier and code phases of the CSK signal even at high dynamic movements of the receiver. The residual error is insensitive to high dynamic movements and can be eliminated slowly with help of the CSK-based DLL and PLL. In its turn, slow adaptation of the tracking loops allows to filter out adaptation noise caused by punctured (zeroed from time to time) DLL and PLL inputs. Thus, high quality demodulation and tracking of CSK signal is achieved with reduced hardware resources: with N=256 single quadrature correlators and M=16 two-quadrature correlators.
In another embodiment of the invention, a multiplexed signal such as QZSS LEX L6D/L6E can be demodulated without separation of L6D and L6E components. In such an implementation the single-quadrature correlators with said (D′1 . . . D′512) code shifts and additional two-quadrature correlators with code shifts (T′1 . . . T′32) with a total number of 32 (for example) are connected to a common carrier PLL and a common DLL, same as described in the previous example. Two conditions from the previous example for code shifts Ti need to be met as well. The difference is that two correlators with maximum signed quadrature output are searched. Let DDMAX and DEMAX be the code shifts of the selected correlators. The code shift DDMAX is selected between shifts with odd indices: (D′1, D′3, D′5 . . . D′511), the code shift DEMAX is selected between shifts with even indices: (D′2, D′4, D′6 . . . D′512). Index of the DDMAX code shift is used to determine data value modulated in the L6D signal. Index of the DEMAX code shift is used to determine the data value modulated in the L6E signal.
For time intervals, when the first condition above is met for code shift DDMAX or DEMAX or both, the selected correlator outputs, and outputs from two-quadrature correlators with neighboring T′i, are fed to the early-late DLL discriminator as described in previous example. Similarly, for time intervals, when the second condition above is met for code shift DDMAX or DEMAX or both, the two-quadrature outputs for the code shifts are fed to the carrier PLL. For time intervals, when none of the above conditions is met for DDMAX or DEMAX, the DLL and the carrier PLL are not updated.
In some embodiments, a multiplexed CSK signal, such as considered above, is demodulated in non-coherent mode, i.e., scalar correlator outputs among which the maximum value is selected are calculated as a power of two quadratures. In other embodiments, both coherent and non-coherent demodulation is supported. To implement decision schemes for coherent and non-coherent demodulation, the DDMAX and DEMAX are calculated either according to maximum of one signed quadrature or according to power of two quadrature's correspondingly.
A navigation receiver includes an ADC, a computation device, an array of first type correlators and an array of second type correlators. In the array of correlators of the first type code delay, carrier phase and carrier frequency are simultaneously controlled via a common controlling input. In the array of correlators of the second type, code delay is individually controlled for each correlator or for each sub-group of correlators, whereas carrier phase and carrier frequency of the correlators are controlled via a common controlling input affecting all the correlators at once (similar to the array of first type correlators). The array of first type correlators also includes at least one unit to select a maximum. An ADC output is connected to a signal input of the arrays of first and second type correlators. Correlator outputs in the array of first type correlators are connected to the input of the max selection unit. The output of the max selection unit is connected to the first input of the computation unit. Correlator outputs in the array of second type correlators are connected to the array of second inputs of the computation unit. The first output of the computation unit is connected to the common input controlling code delay, carrier phase and frequency for the array of first type correlators, the array of third outputs of the computation unit is connected to the inputs controlling code delays of correlators in the array of second type correlators. The second output of the computation unit is connected to the common input controlling carrier frequency and phase of the array of second type correlators. The computation unit is typically implemented as a processor.
In a number of embodiments the module of second type correlators includes only code delay individually-adjusted correlators. In other embodiments, the array of second type correlators consists of correlator sub-groups, each of which is individually adjusted so that code delays of all correlators in the sub-groups are reconfigured in a group way. In some other embodiments, in the array of second type correlators there are both individually-adjusted correlators, and individually-adjusted sub-groups of correlators. In this embodiments the whole sub-group is simultaneously controlled in code delay, each correlator in the sub-group being adjusted to its own code delay, which is different from other correlators in the sub-group, according to a law. In some embodiments, code shift separation in one chip is applied, so that the rule of configuring correlators in the sub-group could be as follows. The first correlator in the sub-group is configured to work with a code shift given in a configuration command, the second correlator is configured to work with given code shift increased by one chip, the third correlator is to work with given code shift increased by two chips, and so on. Similarly, each correlator in array of first type correlators is adjusted in group way, according to a law when new information about code delay is fed to the common control input. The law of configuring correlators in the array of first type distributes code shifts so that they correspond to all possible code shifts in the received symbol.
In some embodiments, correlators in the array of first type correlators calculate only one quadrature component. This is enough to demodulate the CSK-symbol when carrier PLL is converged. At the same time, the correlators in the array of second type correlators calculate both quadrature components, enabling carrier PLL to be adapted to the time instants when code shift in the received symbol corresponds to the code shift in one of the correlators of the array of second type correlators.
In embodiments, designated to receive multiplexed signals such as QZSS LEX L6D/L6E, a second unit to select maximum is included in the array of correlators of the first type. In the embodiments one part of correlators in the array of first type correlators is connected by their outputs to the first maximum selection unit, the rest part of correlators is connected to the second maximum selection unit. Outputs of the maximum selection units are connected to the first input of the computation unit.
In some embodiments the correlators are implemented in form of an integrated circuit. In such embodiments correlators in array of first type and in array of second type can be implemented differently. The correlators in the array of first type correlators calculate only one quadrature, whereas other correlators calculate two quadratures. In addition, the correlators from the array of second type correlators in some embodiments contain accumulators with reduced size as well as multipliers and sine/cosine tables with fewer number of digits and table entries. Difference in code phase control in the correlators from array of the first type and array of the second type correlators along with other differences in some embodiments lead to completely different implementations of the correlator arrays. The correlators in array of the first type correlators are implemented as one unit that contains the only module to convolute incoming signal and replica. At the same time the unit contains an array of accumulators to store accumulation results for different code shifts. The unit also contains one or several maximum selection module. All these parts have access to the array of accumulators and are controlled via common interface and internal logic.
In some embodiments the sub-groups of correlators in array of second type correlators are implemented in similar way as array of first type correlators. But unlike the array of first type correlators the unit implementing a sub-group of correlators does not contain maximum selection unit. All correlator outputs in the sub-group are connected to the computation unit.
The computation unit in typical embodiments is a processor, which may be implemented as a standalone processor, a PLD or an ASIC.
In some embodiments the navigation receiver contains correlators (hereafter the correlators of the third type) capable to process a signal with modulation different from CSK. Inputs of these correlators are connected either to the ADC (in some embodiments) or to another ADC (in other embodiments), and the outputs are connected to the computation unit. These correlators are individually controlled by the computation unit (processor) so that code phase, carrier phase and carrier frequency of each correlator are independent. In addition, the third type correlators are capable to be configured for different PRN codes with different code frequencies.
An example of the non-CSK modulated signal to be processed by the correlators of the third type is GPS L2C with carrier frequency 1227.6 MHz, which is close to QZSS LEX carrier frequency 1278.75 MHz. Both these signals can be digitized by the same ADC. Another example of the non-CSK modulated signal is GPS C/A with carrier frequency 1575.42 MHz. To receive this signal, the third type correlators in some embodiments are connected to the other ADC.
In some embodiments the correlators of the first type in the first array feature code shifts (D1 . . . DN) whereas the correlators of the second type from the second array feature (T1 . . . TM) code shifts.
From
The components and blocks shown in
Having thus described a preferred embodiment, it should be apparent to those skilled in the art that certain advantages of the described method and system have been achieved.
It should also be appreciated that various modifications, adaptations, and alternative embodiments thereof may be made within the scope and spirit of the present invention. The invention is further defined by the following claims.
This application is a divisional of U.S. patent application Ser. No. 15/549,209, filed on Aug. 7, 2017, which is a US National Phase of PCT/RU2016/000789 filed on Nov. 16, 2016.
| Number | Name | Date | Kind |
|---|---|---|---|
| 20040042534 | Raphaeli | Mar 2004 | A1 |
| 20120224864 | Walewski | Sep 2012 | A1 |
| 20120308229 | Bahr | Dec 2012 | A1 |
| 20130279541 | Goldberg | Oct 2013 | A1 |
| Number | Date | Country | |
|---|---|---|---|
| 20200072981 A1 | Mar 2020 | US |
| Number | Date | Country | |
|---|---|---|---|
| Parent | 15549209 | US | |
| Child | 16677996 | US |
