RU2409892C2 - Method of navigation signal search - Google Patents

Abstract

FIELD: physics.

SUBSTANCE: proposed method comprises formation and memorisation of complex samples of cross-correlation function for several analysed delay positions, execution of fast Fourier transform for all combinations of data symbols, including those with their shifted limits, non-coherent summation and isolation of one combination that features maximum accumulated signal magnitude. All combinations of signal sequences are shifted in frequency to new frequencies shifted relative to initial medium frequency of indeterminacy interval up and down by quarter interval of indeterminacy. One half of spectral components adjoining the shifted frequencies are taken of all spectral components at output of fast Fourier transform. For every spectral component, maximum signal magnitude is selected for all combinations of data symbols. Selected signals are used to compile the table to determine the most probably combination of frequency, delay and phase of digital data symbol of received signal and to refine frequency and delay magnitudes.

EFFECT: higher efficiency of search, reduced volume of memory for non-coherent accumulation.

3 dwg

Description

The invention relates to electronics, to systems using signals with an extended spectrum and can be used in receivers of navigation signals GPS, GLONASS and Galileo.

There is a method of searching for navigation signals in the time-frequency domain, which includes the calculation of the fast Fourier transform (FFT) from N complex samples, which is the result of coherent integration over the interval τ _{n of the} correlation of the received signal and the generated replica. The coherent integration interval determines the band of a single frequency cell in the form of a bell [1].

This method of searching for navigation signals has a significant limitation in the frequency resolution bandwidth due to a limitation in the duration of the coherent accumulation interval of the input signal, which, in turn, is due to ignorance of the boundaries of the symbols of the received navigation information and their values.

A known method of searching for navigation signals, including the formation of complex samples, each of which is the result of correlation of the received signal and the replica in the time interval corresponding to the region of frequency uncertainty, storing these samples in the interval corresponding to the duration of several symbols of digital information and determining the frequency resolution bandwidth, performing fast Fourier transform for all combinations of information symbols, including when repeating these combinations under shifts within the uncertainty information symbol boundaries, as well as to delay the analyzed position, the incoherent summation and isolation of all combinations of one having the maximum accumulated signal [2].

This method of searching for navigation signals, the closest to the proposed, has the following disadvantages:

- the serial connection of the operation of coherent integration of the received signal in the interval τ _n , where τ _n is determined by the width of the uncertainty zone in frequency ΔF _bit and FFT with a resolution band ΔF _bit , leads to a deterioration in energy characteristics at the edges of the frequency spectrum of the FFT;

- with incoherent summation, the amount of required memory increases due to the fact that the maximum search is carried out for all possible combinations of information symbols;

- when selecting the most probable sequence, the possible signal correlation is not taken into account at single shifts in delay, frequency, and also shifts of the boundaries of information symbols.

The present invention solves the problem of increasing the efficiency of the search for the received signal in frequency and delay, as well as reducing the required memory size with incoherent accumulation.

To achieve this result, a method of searching for navigation signals, including the formation of complex samples, each of which is the result of correlation of the received signal and replica in a time interval corresponding to a region of frequency uncertainty, storing these samples in an interval corresponding to the duration of several symbols of digital information and determining the band frequency resolution, performing fast Fourier transform for all combinations of information symbols, including at the repetition of these combinations during shifts within the uncertainty of the boundaries of information symbols, as well as for the analyzed positions on the delay, incoherent summation and separation from all combinations of the one having the maximum value of the accumulated signal, introduce the transfer in frequency of all combinations of signal sequences to new frequencies shifted relative to the original the middle frequency of the uncertainty interval up and down by a quarter of the uncertainty interval, while for further processing from all spectral the components at the output of the fast Fourier transform take half the spectral components adjacent to the shifted frequencies, for each spectral component, the maximum signal value is selected from all signals corresponding to all possible combinations of information symbols, a table is formed from the selected signals, from which, taking into account the correlation of signals in neighboring cells determine the most probable combination of the frequency and delay of the received signal and the phase of the digital information symbol, and also specify the values frequency and delay.

Signs that distinguish the proposed method of searching for a navigation signal from the one closest to it, known from the publication [2], characterize the frequency transfer operations of all signal combinations to new mid frequencies, the use of n / 2 spectral components adjacent to the new mid frequencies, and selection before incoherent summation the maximum values for all signal combinations and the formation of a table, from which, taking into account the correlation of signals in neighboring cells, the most probable combination in frequency and The received signal and the phase of the digital information symbol, and also specify the frequency and delay values.

Figure 1 shows a block diagram of a device that implements the proposed method of searching for a navigation signal. The device comprises a quadrature correlator (1), a signal replicator (2), a memory for (n + w-1) samples (3), multipliers (4), a first switch (5), a delay line (6), an information symbol combination generator (7), the second switch (8), complex multipliers (9), the third switch (10), the fast Fourier transform (FFT) (11), the maximum isolator (12), the fourth switch (13), the adder with feedback (14 ), table (15).

. Результаты интегрирования, представляющие собой комплексные отсчеты u_l,t, заносят в память (3). Число комплексных отсчетов для каждой фиксированной задержки, l·Δ, составляет (n+m-1), где n определяет длительность интервала когерентного накопления и соответственно ширину полосы разрешения ΔF_paзp на выходе БПФ,

, a «m» - число комплексных отсчетов на длительности информационного символа τ_c,

.The method of searching for navigation signals is implemented as follows. The input signal converted to the middle frequency f _{0 of} the frequency search interval Δf _n in quadrature correlators (1) is multiplied with the signal replica received from the output of the signal replicator (2). The number of different replicas is “k”, the replicas differ by a delay shift. The shift step is Δ (in fractions of the SRP symbol) and the total analyzed delay interval is k · Δ. In the quadrature correlators (1), integration is performed within ε ₁ ÷ τ _u + ε ₁ , where ε ₁ = l · Δ, l∈0 ÷ k-1, the product of the input signal and the signal replica. The integration interval τ _u determine the width of the search range in frequency Δf _n ,

. The integration results, which are complex readings u _{l, t} , are stored in memory (3). The number of complex samples for each fixed delay, l · Δ, is (n + m-1), where n defines the length of the coherent accumulation interval and correspondingly bandwidth ΔF _pazp resolution FFT output,

, a “m” is the number of complex samples on the duration of the information symbol τ _c ,

.

информационных символов. Как правило, значения «n» и «m» выбирают кратными, и отношение

обычно находится в пределах 3÷6.Thus, on the interval of coherent accumulation are located

information symbols. As a rule, the values of "n" and "m" are selected in multiples, and the ratio

usually within 3 ÷ 6.

When searching for navigation signals, the value of the information symbols modulating the navigation signal, as well as the boundaries of these symbols, is a priori unknown. Full-fledged coherent accumulation on the n · τ interval _and with a minimum of energy losses is possible when “guessing” the transmitted information symbols and their boundaries. Such "guessing" is carried out by enumerating all the possible options for the location of information symbols. Taking into account the fact that during further signal processing it does not matter whether the characters are taken in direct or inverse form, the total number of analyzed combinations will be:

комбинаций в формирователе комбинаций информационных символов (7) и размножения указанных комбинаций путем введения задержек

. Последовательность проверки всех комбинаций реализуется первым коммутатором (5), с выхода которого информационная последовательность поступает на перемножители (4), на которых осуществляется снятие модуляции с последовательности комплексных отсчетов u_l,t с «k» выходов памяти (3). Таким образом, общее количество комбинаций анализируемых последовательностей равноThe specified combinatorics is implemented by forming

combinations in the shaper of combinations of information symbols (7) and the propagation of these combinations by introducing delays

. The sequence of checking all combinations is implemented by the first switch (5), from the output of which the information sequence is sent to the multipliers (4), on which the modulation is removed from the sequence of complex samples u _{l, t} from the “k” memory outputs (3). Thus, the total number of combinations of the analyzed sequences is

The second switch (8) provides a sequence of further processing by “k” delays in quadrature correlators (1). For each individual sequence of “n” complex samples, the FFT (11) is converted from the time domain to the frequency domain for further analysis of the spectral composition of the signal.

Since the sequence of complex samples is formed at the output of the correlator (1), which has a frequency response for the power spectral density

будет фиксироваться уменьшение спектральной мощности, что на выходе БПФ приведет к неравнозначности сравнения спектральных компонент. По результатам моделирования это эквивалентно энергетическим потерям порядка 3,9 дБ.shown in figa, when analyzing frequencies that are f _{0 away} by an amount close to

the decrease in spectral power will be recorded, which at the output of the FFT will lead to an unequal comparison of the spectral components. According to the simulation results, this is equivalent to an energy loss of about 3.9 dB.

Losses can be reduced by reducing the integration interval, for example, by half. But this will lead to a doubling of the sequence of complex readings, and also to an increase in the FFT dimension, which is equivalent to a doubling of the memory (3) and more than doubling of the FFT. In many applications this is not acceptable.

и выделения тех участков спектра, которые примыкают к новым срединным частотам

и

, как показано на фиг.2б. То есть при переносе спектра комбинаций последовательностей на срединную частоту

из всех спектральных компонент на выходе БПФ (11) для дальнейшей обработки берут спектральные компоненты, находящиеся в пределах

, и, соответственно для срединной частоты

берут спектральные компоненты, находящиеся в пределах

. Как показывает моделирование, это позволяет улучшить энергетические характеристики при поиске навигационных сигналов на 2,9 дБ. Перенос спектра сигнала с выхода второго коммутатора (8) на частоты

производится в 4 комплексном перемножителе (9), на который поочередно подают частоты смещения

и

.This problem is solved by applying sequential heterodyning of the signal spectrum to

and highlighting those parts of the spectrum that are adjacent to the new mid-frequencies

and

as shown in figb. That is, when transferring the spectrum of sequence combinations to the middle frequency

from all spectral components at the output of the FFT (11), for further processing, take the spectral components that are within

, and, respectively, for the middle frequency

take the spectral components within

. As the simulation shows, this allows you to improve energy performance when searching for navigation signals by 2.9 dB. Transfer of a spectrum of a signal from an output of the second switch (8) to frequencies

produced in 4 complex multiplier (9), to which the bias frequencies are fed alternately

and

.

The complex multiplier multiplies the input complex signal

u _il = A _i (cos2πf ₀ t + jsin2πf ₀ t),

:where A _i is the sequence of samples per signal with a bias frequency

:

The resulting signal has the form:

Replacing time t with a discrete sequence

t = i · τ _u

and given that

,

.we get that the values

,

.

As a result, the structure of the complex multiplier can be reduced to the form:

The block diagram of the complex multiplier is shown in figure 3.

The procedure for processing samples stored in memory (3) consists of the following operations:

1) for a fixed sequence at the output of the FFT (11), n / 2 spectral reports are selected, which are received at the maximum separators (12) corresponding to each spectral count. Without changing the frequency shift setting, the l-th sequence is multiplied with all possible combinations of information symbols. The maximum signal values for each spectral component are selected on the maximum isolators (12) and fed through a fourth switch (13) to a feedback adder (14). After that, the maximum isolators (12) are zeroed.

2) Repeat operation “1” “m” times, varying the possible shifts of the boundaries of information symbols using the delay line (6) and the switch (5).

.3) Repeat operations "1" and "2" with another frequency shift by

.

4) Repeat operations “1”, “2” and “3” “k” times for other delay shifts that are generated in blocks (1) and (2).

As a result of operations “1” - “4”, n · m · k quantities are extracted, which are accumulated in adders with feedback (14).

Feedback adders (14) carry out incoherent signal accumulation during the repetition of operations “1” - “4”, but with other input signal sequences stored in the memory (3). The number of repetitions of NN is determined by the desired signal-to-noise ratio at the output given at the input. The outputs from (14) actually form a table (15) with coordinates for the delay "1", frequency "i" and the shift of the boundaries of the information symbols "j". The determination of the global maximum in (15) immediately determines the position of the signal in the indicated three coordinates. However, a more likely situation is when the global maximum is not clearly expressed, and the desired signal is distributed over neighboring cells. In this case, interpolation is carried out, with the help of which it is possible to clarify the position of the signal by the corresponding coordinate. Moreover, the presence of correlation between cells allows us to reject spurious single outliers, unless, of course, they exceed the sum of the correlated signals. Thus, the formation of the output table allows you to improve signal recognition and improve the accuracy of determining the frequency-time parameters of the signal.

The proposed method of searching for navigation signals increases the efficiency of searching for a received signal by frequency and delay, and also reduces the required memory size with incoherent accumulation.

Literature

1. B. Shayevits, H. Cohen, J. Nir, E.Dochovny. Very Efficient High Sensitivity Fully Coherent AGPS Signal Processing with Almost No Assistance Requirements. ION GPS 2002, 24-27 September 2002.

2. Nesreen I. Ziedan. GNSS Receivers for Weak Signals, chapter 3.2, 3.3. ARTECH HOUSE, 2006.

Claims (1)

A method of searching for navigation signals, including the formation of complex samples, each of which is the result of correlation of received and SRP symbols on a time interval corresponding to the frequency uncertainty region, storing these samples on an interval corresponding to the duration of several symbols of digital information and determining the frequency resolution band, execution fast Fourier transform for all combinations of information symbols, including when repeating these combinations at shifts within the uncertainty of the boundaries of information symbols, as well as for the analyzed positions on the delay, incoherent summation and isolation from all combinations of one having the maximum value of the accumulated signal, characterized in that when incoherent summation is carried out, the frequency transfer of all combinations of signal sequences to new frequencies shifted relative to the original median frequency of the uncertainty interval up and down by a quarter of the uncertainty interval, while for further processing of all spectral components at the output of the fast Fourier transform, take half of the spectral components adjacent to the shifted frequencies, for each spectral component, the maximum signal value is selected from all signals corresponding to all possible combinations of information symbols, a table is formed from the selected signals, from which, taking into account the correlation signals in neighboring cells determine the most likely combination of the frequency and delay of the received signal and the phase of the symbol of digital information, and that specify the value of frequency and delay.

Images