A pseudo-satellite time hopping sequence blind synchronization and multi-frame high-precision joint acquisition method

Abstract

The application discloses a pseudo-satellite time hopping sequence blind synchronization and multi-frame high-precision joint acquisition method, which comprises the following steps: step one, intercepting a single frame of received signals, screening available pseudo-satellites through parallel two-dimensional FFT correlation acquisition, and extracting absolute code phase and carrier Doppler frequency of the pseudo-satellites; step two, extracting the initial phase constraint of each pulse absolute time slot and constructing a one-dimensional mapping sequence, and under a shift shielding cyclic correlation mechanism, converting blind synchronization into a slip correlation search of one-dimensional row numbers of the SIT table, and obtaining a candidate frame index; step three, when multiple candidates are generated due to state collision, verifying the secondary frame coherent energy accumulation on the corresponding time slots of subsequent frames according to the periodic cycle characteristics of the SIT table, and uniquely determining the time hopping starting frame index; and step four, reducing the three-dimensional search space to a two-dimensional subspace according to the time hopping starting frame index, extracting time slot segments from multi-frame signals, and performing non-coherent energy accumulation, so that high-precision joint fine acquisition is completed.

Description

Technical Field

[0001] This invention relates to the field of satellite navigation and pseudo-satellite baseband signal processing technology, and in particular to a method for joint acquisition of pseudo-satellite time-hopping sequence blind synchronization and multi-frame high precision, applicable to signal acquisition of receivers in ground-based pseudo-satellite networking systems using time-hopping / direct sequence spread spectrum (TH / DS-CDMA) signal systems. Background Technology

[0002] Global Navigation Satellite Systems (GNSS) can provide high-precision spatiotemporal reference information in open outdoor environments and are widely used in key fields such as transportation, surveying and mapping, and national defense. However, in scenarios where satellite signals are severely obstructed or completely unavailable, such as indoors, underground, or in urban canyons, the positioning performance of GNSS drops sharply or even fails completely. To compensate for this deficiency, ground-based pseudo-satellite technology has emerged, which provides navigation enhancement or independent positioning services in areas with GNSS signal coverage blind spots by deploying satellite-like signal transmitters on the ground.

[0003] Ground-based pseudosatellite networks typically face a severe near-far effect problem, where strong signals emitted by pseudosatellites closer to the receiver can overwhelm weaker signals emitted from more distant sources, making it difficult to acquire and track the weaker signals. To address this issue, Time-hop / Direct Sequence Spread Spectrum (TH / DS-CDMA) signaling schemes have been introduced into pseudosatellite systems. Under this scheme, each pseudosatellite transmits signals in turn within different time slots according to a predefined time-hop sequence (SIT), ensuring that only one pseudosatellite occupies the channel at any given time, fundamentally avoiding overlapping interference from multiple pseudosatellite signals in the time domain.

[0004] However, while time-hopping mechanisms address the near-far effect, they also introduce new challenges for receiver signal acquisition. Upon power-on, the receiver does not know which starting frame of the time-hopping sequence corresponds to the currently received signal, resulting in blind synchronization of the time-hopping sequence. To achieve correct signal acquisition, the receiver needs to search for the starting frame number of the time-hopping sequence in addition to the traditional two dimensions of Doppler frequency and code phase, expanding the search space from two to three dimensions.

[0005] To address the aforementioned 3D search problem, there are two main typical solutions in the existing technology:

[0006] The first approach is the traditional three-dimensional blind search exhaustive method. This method performs a point-by-point exhaustive search in the three-dimensional solution space composed of Doppler frequency, code phase, and the starting frame number of the time-hopping sequence. This method can utilize long-term incoherent accumulation of multiple frames of signal to obtain excellent noise resistance and high-precision parameter estimation results. However, since the total number of frames in the time-hopping sequence is usually large (up to hundreds of frames), the three-dimensional search space is extremely large, resulting in extremely high computational consumption and acquisition time, making it difficult to meet the requirements of fast first-to-first-flight (TTFF).

[0007] The second approach is the existing rapid joint acquisition method. This method utilizes the spatial orthogonality of pseudosatellites in time slots within a single frame signal to map the multi-satellite acquisition results into an enabling sequence, and estimates the starting frame number of the time hop sequence through cyclic correlation. This method only requires a single frame signal to complete time hop synchronization, making it extremely fast. However, its inherent drawbacks are: the limited information in a single frame makes it highly susceptible to two types of problems: "cross-frame cyclic ambiguity" and "single-frame state overlap" (i.e., different starting frame numbers correspond to the same single-frame mapping pattern). In complex time hop sequences, these two types of problems lead to a very high synchronization error rate (up to 50% or more), severely impacting the reliability of subsequent positioning performance.

[0008] Therefore, there is an urgent need in the existing technology for a method of blind synchronization and signal acquisition of pseudo-satellite time-hopping sequences that can balance acquisition speed and synchronization accuracy. Summary of the Invention

[0009] The purpose of this invention is to provide a method for blind synchronization of pseudo-satellite time-hopping sequences and high-precision joint acquisition of multiple frames, aiming to solve the dual technical problems of excessively long acquisition time of existing traditional three-dimensional blind search exhaustive method and excessively high synchronization error rate of existing fast joint acquisition method, so as to achieve highly reliable signal acquisition that balances fast first positioning time and noise resistance performance.

[0010] To achieve the above objectives, this invention provides a method for blind synchronization of pseudo-satellite time-hopping sequences and high-precision multi-frame joint acquisition, comprising the following steps:

[0011] Step 1: Extract the received signal with a length of one frame, and use parallel two-dimensional FFT correlation acquisition to filter out m usable pseudo-satellites whose correlation peaks exceed the judgment threshold, and extract and record their corresponding absolute code phase and carrier Doppler frequency;

[0012] Step 2: For the m available pseudo-satellites, extract the absolute time slot initial phase constraints of each pulse in a single frame, construct a one-dimensional mapping sequence, introduce the pre-stored global optimal SIT table (i.e., time hopping matrix), and under the shift-masked cyclic correlation mechanism that excludes non-periodic cross-frame splicing false peaks, transform blind synchronization into a one-dimensional sliding correlation search of the one-dimensional row number of the SIT table, and obtain the candidate frame index that maximizes the correlation energy through peak search;

[0013] Step 3: When a state collision occurs within a single frame and multiple candidate frame indices are generated, the absolute row number of the SIT corresponding to the subsequent frame is calculated using the periodic cyclic characteristics of the SIT table. The temporal coherence energy accumulation verification of the next frame is performed on the corresponding time slot segment of the subsequent frame to uniquely determine the true time jump start frame index.

[0014] Step 4: After locking the time hop start frame index, the three-dimensional search space containing carrier frequency offset, code phase and frame start index is reduced to a two-dimensional subspace. The associated SIT table time slot sequence is used to extract time slot segments of multiple frames of signals and accumulate non-coherent energy of multiple frames to complete high-precision joint fine acquisition solution.

[0015] Preferably, in step S2: the shift-masked cyclic correlation mechanism specifically involves: during the one-dimensional row number k sliding search, if the expected time slot position... If the time slot of the next transmitted signal from an adjacent base station overlaps with the time slot at the frame edge, a masking factor is introduced. Forcefully set the relevant weights of this time slot to zero; otherwise... The original values ​​are maintained to eliminate spurious boundary peaks generated during the splicing of aperiodic time-jump sequences.

[0016] Preferably, in step S3: the criteria for state collision determination and next-frame verification include: if the energy ratio of the highest peak to the second highest peak of the single-frame blind synchronization output is lower than a set threshold. Then, the received signal segment of the third frame duration is further extracted, and temporal coherent energy accumulation verification is performed again for candidate frame numbers with similar energy, until the unique correct coarse acquisition hop index is selected.

[0017] Preferably, in step S4: the spatial dimensionality reduction also includes dimensionality reduction search of phase frequency parameters: using the carrier frequency and absolute code phase obtained in step S1 as the prior search center, a two-dimensional fine search is performed only within its preset narrowed frequency adjacent domain.

[0018] Preferably, in step S2, the one-dimensional slip-correlated energy distribution of the i-th pseudo-satellite is calculated using the following formula:

[0019]

[0020] In the formula, This represents the total number of pulses within a single frame. To find the expected time slot position of the i-th pseudo-satellite in the n-th pulse within the k-th frame, which is obtained by looking up the table, This is the time slot energy mapping function.

[0021] Preferably, in step S3, the absolute line number of the SIT corresponding to the subsequent frame is calculated using the following formula:

[0022]

[0023] Where L is the total number of frames in the SIT table.

[0024] This invention also provides a pseudo-satellite time-hop sequence blind synchronization and multi-frame high-precision joint acquisition system for implementing the above method, including a coarse acquisition module, a shift-masked blind synchronization module, a second-frame verification anti-collision module, and a dimension-reduction fine acquisition module. The coarse acquisition module is used to perform single-frame truncation and parallel two-dimensional FFT correlation on the received signal, filter out m usable pseudo-satellites, and extract the absolute code phase and carrier Doppler frequency. The shift-masked blind synchronization module is used to extract the initial phase constraint of the time slot and perform a one-dimensional row number search of the SIT table under the shift-masked cyclic correlation mechanism to obtain the candidate frame index. The second-frame verification anti-collision module is used to calculate the absolute row number of the SIT corresponding to the second frame and perform temporal coherent energy accumulation of the second frame when encountering state collision, uniquely locking the time-hop start frame index. The dimension-reduction fine acquisition module is used to reduce the three-dimensional search space to a two-dimensional subspace and perform multi-frame incoherent accumulation on the extracted time slot segments based on the locked frame start index, outputting high-precision pseudorange and Doppler measurement values.

[0025] Compared with the prior art, the present invention has the following beneficial effects:

[0026] (1) True blind synchronization of time-hopping signals is achieved without any prior frame synchronization information. Existing time-hopping signal acquisition methods (such as the THSI method and methods based on dynamic Bayesian networks) usually assume that the receiver knows the starting frame boundary of the time-hopping sequence, or require continuous observation of multiple pulses to complete the time-hopping parameter estimation. This invention creatively introduces an absolute time slot initial phase constraint ( The system employs a coherent energy verification mechanism with the next frame, enabling the time-hopping start frame number to be locked to a very small candidate set (typically 2-5) based solely on a single frame signal, without any external time reference. This mechanism provides the receiver with full cold-start blind synchronization capability, greatly enhancing the system's autonomy and robustness in complex environments.

[0027] (2) Eliminating single-frame collision states to achieve highly reliable time-hopping synchronization. When different frame numbers in the Time-hopping Sequence (SIT) table have the same single-frame time slot allocation pattern (i.e., "single-frame state collision"), existing joint acquisition algorithms will make incorrect decisions due to ambiguity. This invention, through the sub-frame coherent energy verification mechanism in step S3, only needs to extract a signal of one additional frame duration. Using the carrier frequency and code phase obtained in step S1, targeted coherent integration is performed on the expected time slot of the next frame for the candidate frame number. With an extremely low additional overhead of 0.026 seconds, the synchronization accuracy is significantly improved to 97.5%. If a third frame verification is further performed on candidate frame numbers with similar verification energy, the accuracy can approach 100%.

[0028] (3) Achieving both acquisition speed and accuracy through a coarse-fine hybrid architecture. Although the traditional three-dimensional exhaustive method (traversing Doppler frequency × code phase × time-hopping frame number) can achieve high accuracy and high precision estimation, its computational complexity is F×C×L, and the acquisition time is tens of seconds, which is difficult to meet the real-time requirements. In the coarse acquisition stage of steps S1 to S3, the time-hopping frame number is uniquely determined using a single frame to two frames of signal (taking only about 0.1 seconds); then, through spatial dimensionality reduction in step S4, the search space for fine acquisition is reduced from three dimensions to two dimensions, reducing the complexity to F×C, and the processing gain is the same as that of the traditional method by using multi-frame incoherent accumulation.

[0029] (4) Low computational resource consumption. The single-pulse acquisition in step S1 and the fine acquisition in step S4 are both implemented using FFT parallel correlation, which makes full use of the frequency domain acceleration capability of modern digital signal processors. The mapping sequence construction and cyclic correlation in step S2 only involve vector inner product operations of very low dimension, and the sub-frame verification in step S3 only involves coherent integration of a finite number of time slots in the time domain. The entire acquisition process does not involve complex iterative or recursive calculations, has a small memory footprint, high computational efficiency, and is suitable for FPGA, DSP or embedded ARM platforms, with good engineering practicality. Attached Figure Description

[0030] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention.

[0031] Figure 1 The above is an overall flowchart of a pseudo-satellite time-hopping sequence blind synchronization and multi-frame high-precision joint acquisition method provided in an embodiment of the present invention.

[0032] Figure 2 This is a comparison chart of the accuracy of the embodiments of the present invention and existing technologies (traditional three-dimensional exhaustive search method and original multi-satellite joint acquisition algorithm) in 600 Monte Carlo tests. The horizontal axis represents different methods, and the vertical axis represents the percentage of accuracy.

[0033] Figure 3 This is a comparison chart of the capture time distribution between the method of this invention and the prior art. In 600 Monte Carlo simulations:

[0034] Figure 3 (a) shows the time distribution of the multi-satellite joint capture method;

[0035] Figure 3 (b) is the time distribution of the coarse and fine two-stage mixed capture method proposed in this invention;

[0036] Figure 3 (c) represents the time distribution of the traditional three-dimensional blind search exhaustive method. Detailed Implementation

[0037] The present invention will be further described in detail below with reference to specific embodiments, which is intended to explain the present invention and should not be construed as limiting the present invention.

[0038] This embodiment provides a method for blind synchronization of pseudo-satellite time-skip sequences and joint acquisition of high-precision multi-frame acquisition. This method abandons the traditional basis stitching mode limited by the prime number dimension of the Galois field and directly performs optimization globally. Figure 1 As shown, the specific implementation steps are as follows:

[0039] S1: Single pulse signal capture;

[0040] S2: Enabled mapping and correlation with absolute phase constraints;

[0041] S3: Sub-frame verification based on temporal coherent energy;

[0042] S4: Search space dimensionality reduction and multi-frame precision capture.

[0043] The following provides further explanation in conjunction with the specific implementation steps:

[0044] Step S1: Single Pulse Signal Acquisition

[0045] Assume the pseudo-satellite networking system contains N pseudo-satellites (e.g., ), using the TH / DS-CDMA signaling system, the total number of frames in the time hop sequence (SIT) is L frames (e.g. Each frame contains N time slots, and each time slot corresponds to the launch window of a pseudosatellite. Specific parameter examples are shown in Table 1 below:

[0046] Table 1

[0047] symbol meaning Example value N Number of pseudo-satellites 5 L SIT total frames 100 Number of chips per time slot 2046 Sampling rate 62MHz Number of sampling points per time slot 12400

[0048] After the receiver starts up, it first extracts a segment of the baseband received signal with a length of one frame (i.e., N time slots), denoted as . ,in This is the index for the sampling points.

[0049] Parallel two-dimensional acquisition of single-pulse signals is performed on this single-frame signal segment. Specifically, for the i-th pseudo-satellite ( F frequency hypothesis grid points are set in the Doppler frequency dimension. (The frequency search step size is set to, for example, 500Hz), and C code phase hypothesis grid points are set in the code phase dimension (usually). In this embodiment, the duration of a single time slot is 0.2 ms, and the pseudo-code period is 0.1 ms. Therefore, each active time slot contains two complete pseudo-code periods. Single-pulse acquisition can coherently integrate these two code periods separately and incoherently accumulate (or take the maximum value) the correlation energy of both to overcome the influence of navigation message bit flipping on the coherent integration gain, thereby improving the detection probability. For each set of frequency-phase assumptions ( ), calculate the relevant energy:

[0050]

[0051] In engineering implementation, to reduce the computational complexity of time-domain sliding, it is preferable to use FFT parallel acceleration to calculate the aforementioned related energies:

[0052]

[0053] in, This is the intermediate frequency (IF) of the receiver. Let i be the local pseudocode sequence corresponding to the i-th pseudosatellite. ( () represents the system sampling period. This represents the computation of the complex conjugate of the Fast Fourier Transform. This represents the inverse fast Fourier transform.

[0054] Let the capture decision threshold be... . It can be dynamically determined by multiplying the noise floor statistics of the received signal by a constant factor (e.g., taking the mean of the noise floor). (Multiple times), to ensure a constant false alarm probability (CFAR). If the relevant energy value of a pseudosatellite at any grid point is... Exceed If the pseudo-satellite is found to be active and usable within the current signal frame, the absolute code phase of its corresponding correlation peak is recorded. (In units of sampling points, i.e., the absolute position index of the correlation peak within the entire frame signal) and carrier frequency estimate .

[0055] This step requires identifying at least two usable pseudo-satellites to facilitate the construction of the time slot mapping sequence in subsequent steps. If the number of usable pseudo-satellites exceeds the threshold by less than two, it indicates that the signal quality of the current frame is insufficient to support subsequent time hop synchronization. In this case, you can choose to wait for the extraction of the next frame signal or switch the acquisition strategy (such as extending the incoherent integration time).

[0056] Since this step only captures a single frame of signal and uses FFT frequency domain multiplication to replace time domain convolution, its computational load is comparable to that of traditional single-star two-dimensional blind search. The overall coarse acquisition process can be completed quickly in milliseconds.

[0057] Step S2: Enable mapping and correlation with absolute phase constraints

[0058] The absolute code phase set of at least two usable pseudo-satellites obtained in step S1 ( , From the sample points (in units), select the smallest value. The absolute code phase of the earliest arriving pseudosatellite. Divide by the number of sampling points contained in a single time slot Then, round down to obtain the initial phase constraint of the absolute time slot. :

[0059]

[0060] in, This indicates the floor function. For the range of values ​​within The integer represents the absolute timeslot number (indexed from 0) of the earliest available pseudosatellite signal within the current received frame. The physical meaning of this constraint is that regardless of which frame the time hop sequence begins to cycle from, the alignment of the timeslot boundaries on the absolute time axis is fixed. The absolute time anchor point of the first active time slot in the received frame was uniquely determined.

[0061] Next, we construct the receiver mapping sequence R and the local mapping sequence P.

[0062] The method for constructing the receiving end mapping sequence R is as follows:

[0063] Let m be the number of available pseudo-satellites participating in the mapping, and select the number of bits for the identifier code. ( (This represents rounding up), meaning each time slot state is represented by x elements. A unique x-dimensional orthogonal identifier vector is assigned to each available pseudosatellite. Each component A time slot not occupied by any available pseudo-satellites is identified by its corresponding x-dimensional all-zero vector. For example, when At that time, it can be taken The identifier vectors of the four pseudo-satellites can be set as follows: , , , .

[0064] The total length of the receiving end mapping sequence R is For the s-th time slot within the received frame ( If the time slot is occupied by the kth available pseudosatellite, then the identifier vector will be... Fill in the corresponding position; if the time slot is not occupied, fill in the all-zero vector. Among them, the absolute time slot positions occupied by the available pseudo-satellites. The absolute code phase obtained in step S1 is converted to obtain:

[0065]

[0066] Rounding is used here to tolerate minor jitter errors caused by physical propagation delay, ensuring robustness of time slot assignment.

[0067] The local mapping sequence P is constructed as follows:

[0068] The total length of the local mapping sequence P is , where L is the total number of frames in the time-skip sequence (SIT). For the f-th frame ( The s-th time slot in ) According to the prior definition of the SIT table, if the time slot is assigned to pseudo-satellite i and pseudo-satellite i belongs to the set of available pseudo-satellites selected in step S1, then fill in the corresponding position. If allocated to an unavailable pseudo-satellite or an idle time slot, fill with a vector of all zeros. .

[0069] Error shift masking and loop-related operations:

[0070] Perform a cyclic shift correlation operation between the received mapping sequence R and the local mapping sequence P. For the time slot-level shift... Correlation value Defined as R and P with The inner product of equal-length subsequences starting at:

[0071]

[0072] Introducing an error shift masking mechanism: For each shift amount q, calculate its corresponding time slot offset. ,like This indicates that the shift amount causes the slot boundary of the local sequence to be inconsistent with the initial phase of the absolute slot observed in the received frame, which is a false assumption resulting from cross-frame splicing. Therefore, an absolute phase constraint piecewise function is introduced to force the correlation value of the false assumption to be set to negative infinity (or zero):

[0073]

[0074] After the above filtering and masking, the global maximum value is found among the remaining valid relevant values. Extracting the desired result ( For example, a very small positive number All displacements q (used to tolerate system floating-point operation errors) are converted into corresponding candidate jump start frame indices:

[0075]

[0076] Here, +1 is used to convert the frame offset starting from 0 to an absolute frame number starting from 1. After deduplicating all candidate frame numbers that meet the conditions, a candidate set is obtained. .

[0077] If candidate set If there is no state overlap, then the unique candidate frame number is directly used as the final coarse capture hop index. And skip the verification step and proceed directly to step S4. If This indicates a single-frame state collision (i.e., there are multiple different starting frame numbers in the SIT table, which produce the exact same time slot allocation pattern in different frames). In this case, it is necessary to proceed to step S3 to perform sub-frame coherent energy verification to eliminate ambiguity.

[0078] Step S3: Sub-frame verification based on temporal coherence energy

[0079] When step S2 generates a candidate set and This indicates that there are at least two different starting frame number assumptions under a single frame information, which produce the exact same time slot allocation pattern within the current received frame (i.e., a "single frame state collision" phenomenon has occurred). In this case, the correct time hop starting frame number cannot be uniquely determined based solely on the spatial mapping information of a single frame.

[0080] This invention breaks the spatial symmetry by introducing the time-domain signal features of the next frame. The specific verification method is as follows:

[0081] (a) Extracting a segment of the next frame signal

[0082] In the time domain, a segment of the baseband received signal, the duration of the next frame immediately following the first frame, is continuously extracted and denoted as... ,in The index of the sampling point within the next frame (N is the total number of time slots in a single frame, (This refers to the number of sampling points in a single time slot). This signal segment corresponds to the second frame continuously received by the receiver.

[0083] (b) Determine the expected time slot for the next frame corresponding to each candidate frame number.

[0084] For each candidate start frame number in the candidate set First, calculate its next frame line number in the time-skip sequence (SIT). Considering the periodic and cyclical nature of the SIT table, Calculate using the following formula:

[0085]

[0086] Where L is the total number of frames in the SIT table, mod is the modulo operation, and +1 is used to convert the result into an absolute frame number starting from 1. Based on the prior definition of the SIT table, the frame number is found... The activation slot number of each available pseudosatellite in the frame. For the i-th available pseudosatellite (i belongs to the set of available pseudosatellites selected in step S1), its expected activation slot number in the next frame is denoted as . ,and .

[0087] (c) Perform signal extraction and coherent integration on the expected time slot of the next frame.

[0088] For the i-th available pseudosatellite, in the next frame signal Extract the first The signal segment corresponding to each time slot. This time slot is in The starting and ending range of the sampling points in the sample is:

[0089]

[0090]

[0091] in, The number of sampling points corresponding to a single time slot is used to extract the signal segment within that time slot. Local index within time slots .

[0092] Using the absolute code phase of the i-th pseudo-satellite obtained in step S1 and carrier frequency estimates The time slot segment is then subjected to time-domain decarrier and despreading processing. Let the number of sampling points corresponding to a single pseudocode period be... (In this embodiment, the pseudocode period is 0.1ms and the sampling rate is 62MHz, therefore...) Since both the frame length and the slot length are integer multiples of the pseudocode period, the absolute code phase... The relative code phase within the time slot is obtained by taking the modulus of the pseudocode period. :

[0093]

[0094] Since there are two complete pseudocode cycles within an active time slot, to prevent the cancellation of integrated energy caused by navigation message bit flipping, the coherent integration within the time slot is divided into segments based on a single pseudocode cycle, and the maximum value of the integrated energy of each segment is taken (or incoherent accumulation is performed). The coherent integrated energy... The calculation is as follows:

[0095]

[0096] in, For the pseudocode periodic segment index within the time slot, Let i be the local pseudocode sequence of the i-th pseudosatellite. This is the system sampling period.

[0097] (d) Accumulate the verification energy of each star

[0098] Candidate frame number Under the assumption that the slotted coherent integral energy of all m available pseudosatellites is accumulated, the total verification energy of the candidate frame number can be obtained. :

[0099]

[0100] Where m is the number of available pseudo-satellites selected in step S1 ( ).

[0101] (e) Select the candidate frame number with the highest energy.

[0102] Compare all candidate frame numbers The total verification energy is selected to make The candidate frame number that reaches the global maximum value is used as the unique correct coarse capture time hop index. :

[0103]

[0104] The principle behind this step is: for the correct time-skip start frame number The expected activation time slot position of the next frame perfectly matches the actual transmission time slot positions of each pseudo-satellite in the actual air interface signal. After decarrier and pseudo-code removal using the known phase frequency parameters, the coherent integral will accumulate an extremely high energy peak within the time slot window; however, for erroneous candidate frame numbers, the expected activation time slot position is misaligned with the actual transmission time slot, and the extracted signal segment is actually pure noise or cross-correlation interference from other pseudo-satellites, with the coherent integral result approaching the noise floor level. Therefore, this verification mechanism has extremely high confidence.

[0105] Furthermore, since this step only needs to process a limited number of candidate frame numbers (usually...) It performs targeted pure time-domain inner product operations within one frame, completely avoiding the massive FFT frequency domain search. Therefore, its additional computing power consumption is extremely low, perfectly maintaining the fast characteristics of the coarse acquisition stage of this invention.

[0106] As a further optimization, if the verification energy difference of multiple candidate frame numbers in step S3 is less than a preset threshold (e.g., the ratio of the second largest value to the largest value exceeds 0.8), it indicates that the time slot allocation pattern of the next frame still has collisions. In this case, a received signal segment of the third frame duration can be extracted, and temporal coherent energy verification can be performed again for candidate frame numbers with similar energy. Since the probability of the time slot allocation patterns of three consecutive frames being exactly the same in actual time hop sequence design is extremely low, the verification of the third frame can almost certainly completely eliminate the time hop index error rate, further improving the accuracy of coarse acquisition time hop index to close to 100%.

[0107] Step S4: Search Space Dimensionality Reduction and Multi-Frame Precision Capture

[0108] After verification in step S3 (or direct output of step S2 in the case of a single candidate), the system has determined the unique and correct coarse capture time-lapse start frame index. (This index starts from 1,) ).

[0109] The search space of the traditional three-dimensional blind exhaustive search method can be represented as a huge three-dimensional hypothesis space, whose three dimensions are: the Doppler frequency hypothesis (let's say it's a hypothesis that the Doppler frequency hypothesis is ... (number of grid points), code phase assumption (set to) (number of grid points), assumption of the start frame number of the time jump sequence (total number of grid points), (Each frame number). Traditional methods require traversing... Each set of hypothetical grid points... Calculating the relevant energies separately results in an extremely large computational load, with a complexity of O(n log n). .

[0110] This invention is based on a determined unique time-hopping index. Dimensionality reduction based on prior constraints is performed on the three-dimensional search space: the dimension of the starting frame number of the time-skip sequence is forcibly locked to 1. Only the two-dimensional search subspace consisting of Doppler frequency and code phase is retained. The search space size ranges from... Reduced to The computational load is directly reduced compared to traditional methods. times.

[0111] Furthermore, to obtain accurate parameter estimates with high signal-to-noise ratio gain, the receiver performs incoherent accumulation using multiple consecutively received frames of signal. Let the number of accumulated frames be... (For example ), continuously received The frame signal is denoted as Each frame is [length]. One sampling point.

[0112] For each hypothetical lattice point in the reduced two-dimensional search space ,in Assuming Doppler frequency shift, Assuming a code phase delay, its value in the first... frame( The relevant energy in ) The calculation process is as follows:

[0113] First, based on the known coarse capture time-hop index Based on the periodic and cyclical properties of the jump time sequence (SIT), determine the first... The absolute line number of the SIT corresponding to the frame :

[0114]

[0115] in, For modulo operation. According to the SIT table, ... The definition of a line is used to find the currently processed line number. pseudo-satellite ( The set of available pseudo-satellites selected in step S1 consists of a total of (piece) in the first Activation slot number in the frame ,and .

[0116] In signal frame Extract the first The signal segment corresponding to each time slot has a global sampling point start and end range of:

[0117]

[0118]

[0119] Extract the time slot segment and introduce a local discrete index within the time slot. , recorded as Using the same FFT parallel correlation method as in step S1, the pseudo-satellites within this time slot are calculated. In the assumption The relevant energy below:

[0120]

[0121] in, The intermediate frequency, For the first The local pseudocode sequence of a pseudosatellite. The sampling period is defined as follows. If a single time slot contains multiple pseudocode periods, the maximum value of the relevant energy of each period is taken, or incoherent accumulation is performed.

[0122] The first All within the frame The time-slot correlation energy of each pseudosatellite can be accumulated to obtain the frame under the assumption that... Total relevant energy below:

[0123]

[0124] right The correlation energy of each frame is incoherently accumulated frame by frame to obtain the hypothetical grid. Multi-frame cumulative energy:

[0125]

[0126] Multi-frame incoherent accumulation can significantly smooth random noise and improve the signal-to-noise ratio, enabling the present invention to still achieve the great processing gain of traditional long-time accumulation methods under the premise of fast time-hopping synchronization.

[0127] Cumulative energy at all two-dimensional hypothetical lattice points Search for the global maximum value and its corresponding Doppler frequency shift. Sum code phase delay This is a high-precision estimate:

[0128]

[0129] Ultimately, the high-precision carrier frequency output by the system is High-precision code phase is .

[0130] As a further optimization, the two-dimensional spatial search in step S4 does not require a blind search across the entire domain. The system can directly use the coarsely estimated carrier frequency obtained quickly in step S1. and rough estimation of absolute code phase As a priori search center, search only within its nearby neighborhood (e.g., frequency range). Code phase range This method performs a fine-grained search after removing the time-skip frame number dimension. Based on this, the phase frequency dimension was further reduced. The grid size enables secondary acceleration of the search space.

[0131] In summary, step S4 reduces the 3D search to a 2D search through spatial dimensionality reduction based on prior constraints, thus improving capture time performance. The high precision and robustness of the fine capture output are ensured by multi-frame incoherent accumulation. Based on this, by combining the coarse estimation results of step S1, the overall capture calculation time is reduced by more than two orders of magnitude compared with the traditional three-dimensional exhaustive method, thus achieving a significant improvement in capture performance.

[0132] This embodiment also provides a pseudo-satellite time-hopping sequence blind synchronization and multi-frame high-precision joint acquisition system that implements the above method. The system corresponds one-to-one with steps S1 to S4 of the above method and includes a coarse acquisition module, a blind synchronization module with shift shielding, a second-frame verification anti-collision module, and a dimension reduction fine acquisition module. Each module can be implemented by digital signal processing platforms such as FPGA, DSP, or embedded ARM.

[0133] The coarse acquisition module corresponds to step S1. It is used to intercept a single frame of received signal and perform parallel two-dimensional FFT correlation acquisition, filter out m usable pseudo-satellites whose correlation peak exceeds the judgment threshold, extract and record the absolute code phase and carrier Doppler frequency of each usable pseudo-satellite, and output them to the blind synchronization module with shift shielding.

[0134] The shift-masked blind synchronization module corresponds to step S2. It is used to extract the initial phase constraint of the absolute time slot based on the absolute code phase of each available pseudo-satellite, construct the receiver mapping sequence and the local mapping sequence, and perform a sliding correlation search on the one-dimensional row number of the SIT table under the shift-masked cyclic correlation mechanism to output one or more candidate frame indices.

[0135] The subframe verification anti-collision module corresponds to step S3. When the shift-masked blind synchronization module outputs multiple candidate frame indices and a single frame state collision occurs, it calculates the next frame absolute row number of each candidate frame according to the periodic cyclic characteristics of the SIT table, extracts the subframe signal, performs temporal coherent energy accumulation verification on the corresponding expected time slot, and uniquely locks the true time jump start frame index after comparing the verification energy of each candidate frame.

[0136] The dimension reduction and fine capture module corresponds to step S4. After the time hop start frame index is locked, it reduces the three-dimensional search space composed of carrier frequency offset, code phase and frame start index to a two-dimensional subspace composed of carrier frequency offset and code phase. Based on the SIT table time slot sequence, it extracts time slot segments and accumulates multi-frame incoherent energy for consecutive frames of signals, and outputs high-precision carrier frequency, code phase and pseudorange and Doppler measurement values ​​calculated therefrom.

[0137] To verify the effectiveness of this invention, 600 Monte Carlo comparison tests were conducted between the method of this invention and the prior art. The results are as follows: Figure 2 and Figure 3 As shown.

[0138] Regarding capture accuracy, such as Figure 2As shown, the blind synchronization accuracy of the method of the present invention reaches 97.5%, which is more than 4 times higher than the original multi-satellite joint acquisition algorithm of 22.0%, and close to the 100% of the traditional three-dimensional blind search exhaustive method.

[0139] Regarding capture time, such as Figure 3 As shown, Figure 3 The multi-satellite joint capture method shown in (a) has an average time of 0.086 seconds. Figure 3 (b) The coarse-fine hybrid acquisition method of the present invention, after introducing absolute phase constraints and sub-frame verification, has an average acquisition time of only 0.112 seconds. Figure 3 (c) shows that the traditional three-dimensional blind search exhaustive method takes an average of 36.153 seconds and the distribution range is wide. It can be seen that the present invention can significantly improve the synchronization accuracy with only an additional overhead of about 0.026 seconds, and takes into account both the acquisition speed and the synchronization reliability.

Claims

1. A method for blind synchronization of pseudo-satellite time-skip sequences and joint high-precision acquisition of multiple frames, characterized in that, Includes the following steps: Step S1: Extract the received signal of length one frame, and use parallel two-dimensional fast Fourier transform (FFT) for correlation acquisition to filter out m usable pseudo-satellites whose correlation peaks exceed the decision threshold, and extract and record their corresponding absolute code phases. and carrier Doppler frequency ; Step S2: For the m available pseudo-satellites, extract the absolute time slot initial phase constraints of each pulse within a single frame, construct a one-dimensional mapping sequence, introduce a pre-stored globally optimal SIT table, and under the shift-masked cyclic correlation mechanism that excludes non-periodic cross-frame splicing spurious peaks, transform blind synchronization into a one-dimensional glide correlation search of the one-dimensional row number k of the SIT table. Obtain the candidate frame index that maximizes the correlation energy through peak search. ; Step S3: When multiple network nodes overlap in time slots within a single frame, causing state collisions and generating multiple candidate frame indices that meet the threshold, the absolute row number of the SIT corresponding to the w-th frame is calculated using the periodic cyclic characteristic of the SIT table. In the subsequent M... acc The temporal coherence energy accumulation verification of the next frame is performed on the corresponding time slot segment of the frame, that is, only if M consecutive M acc When the coherent energy after frame accumulation continues to increase and uniquely exceeds the decision threshold, the current blind synchronization is determined to be successful, and the true time-hop start frame index is uniquely determined. ; Step S4: Successfully lock the time-hop start frame index during blind synchronization. Then, it will contain carrier frequency offset, code phase, and frame start index. Dimensional reduction of the three-dimensional search space Two-dimensional subspace; within the two-dimensional subspace, utilizing the The associated SIT table time slot sequence is used to extract time slot segments of signals within multiple frames and to accumulate non-coherent energy across multiple frames, thereby achieving high-precision joint fine acquisition and resolution.

2. The method for blind synchronization of pseudo-satellite time-skip sequences and high-precision multi-frame joint acquisition as described in claim 1, characterized in that, In step S2: The shift-masked cyclic correlation mechanism is specifically as follows: during the one-dimensional row number k sliding search, if the expected time slot position... If the time slot of the next transmitted signal from an adjacent base station overlaps with the time slot at the frame edge, a masking factor is introduced. Forcefully set the relevant weights of this time slot to zero, otherwise The original values ​​are maintained to eliminate spurious boundary peaks generated during the splicing of aperiodic time-jump sequences.

3. The method for blind synchronization of pseudo-satellite time-skip sequences and high-precision multi-frame joint acquisition as described in claim 1, characterized in that, In step S3: The criteria for state collision determination and next-frame verification include: If the energy ratio of the highest to the second highest peak of the single-frame blind synchronization output is lower than a set threshold... Then, the received signal segment of the third frame duration is further extracted, and temporal coherent energy accumulation verification is performed again for candidate frame numbers with similar energy, until the unique correct coarse acquisition hop index is selected.

4. The method for blind synchronization of pseudo-satellite time-skip sequences and high-precision multi-frame joint acquisition as described in claim 1, characterized in that, In step S4: The spatial dimensionality reduction also includes dimensionality reduction search of phase frequency parameters: using the carrier frequency and absolute code phase obtained in step S1 as the prior search center, a two-dimensional fine search is performed only within its preset narrowed frequency adjacent domain.

5. The method for blind synchronization of pseudo-satellite time-skip sequences and high-precision multi-frame joint acquisition as described in claim 1, characterized in that, In step S2, the one-dimensional slip-correlation energy distribution of the i-th pseudo-satellite is calculated using the following formula: In the formula, This represents the total number of pulses within a single frame. To find the expected time slot position of the i-th pseudo-satellite in the n-th pulse within the k-th frame, which is obtained by looking up the table, This is the time slot energy mapping function.

6. The method for blind synchronization of pseudo-satellite time-skip sequences and high-precision multi-frame joint acquisition as described in claim 1, characterized in that, In step S3, the absolute line number of the SIT corresponding to the subsequent frame is calculated using the following formula: Where L is the total number of frames in the SIT table.

7. A pseudo-satellite time-hopping sequence blind synchronization and multi-frame high-precision joint acquisition system for implementing the method of any one of claims 1-6, characterized in that, include: Coarse acquisition module: Used to perform single-frame truncation and parallel two-dimensional FFT correlation on the received signal, filter out m usable pseudo-satellites and extract absolute code phase. and carrier Doppler frequency ; Blind synchronization module with shift mask: used to extract the initial phase constraint of the time slot and perform a one-dimensional search for the row number k under the shift masked cyclic correlation mechanism to obtain the candidate frame index. ; Next-frame verification anti-collision module: used to detect state collisions based on... Calculate the absolute row number of SIT corresponding to the next frame, and accumulate the temporal correlation energy of the next frame to uniquely lock the starting frame index of the time jump; Dimensionality Reduction and Precision Capture Module: This module reduces the dimensionality of the 3D search space to a 2D subspace and performs multi-frame incoherent accumulation of the extracted time slot segments based on the locked frame start index, outputting high-precision pseudorange and Doppler measurements.

Images