A symbol timing synchronization method, system, device, medium and program product for a receiving end in a MIMO-OFDM architecture

Abstract

The application belongs to the field of digital baseband signal processing, and discloses a symbol timing synchronization method, system, device, medium and program product of a receiving end in a MIMO-OFDM architecture, which comprises the following steps: a sending end generates a MIMO orthogonal frequency division multiplexing data frame, introduces cyclic shift diversity, and transmits the data frame to a receiving end; the receiving end performs radio frequency front-end processing and coarse frequency offset compensation to obtain a discrete receiving signal, performs cross-time splicing on the discrete receiving signal, and then performs maximum ratio combination after cross-correlation matching with a local long training sequence to output a single joint cross-correlation metric; based on the joint cross-correlation metric, two sliding windows are constructed along a time axis, the ratio of the integral energies of the two sliding windows is calculated, and finally the tail of the long training sequence is locked as a globally unique symbol timing synchronization point after peak searching of the ratio; the system, the device and the medium are used for realizing the method, and the program product comprises a computer program used for realizing the method; and the application has a unique main peak under any channel condition.

Description

Technical Field

[0001] This invention belongs to the field of digital baseband signal processing technology, specifically relating to a symbol timing synchronization method, system, device, medium, and program product for the receiver in a MIMO-OFDM architecture. Background Technology

[0002] When performing long-range air-to-ground communication missions, unmanned aerial vehicles (UAVs) face extremely challenging channel conditions, including very low signal-to-noise ratios (SNR) and strong multipath delay spread. To meet the high reliability and high throughput communication requirements in complex scenarios, UAV broadband data links typically employ a multiple-input multiple-output orthogonal frequency division multiplexing (MIMO-OFDM) architecture (such as a multi-antenna transceiver architecture based on the IEEE 802.11n protocol standard). In the physical layer baseband processing of the MIMO-OFDM architecture, the receiver must rely on the long training sequence (L-LTF) in the preamble for precise symbol timing synchronization to determine the accurate starting truncation position for the subsequent Fast Fourier Transform (FFT).

[0003] Existing symbol timing synchronization techniques have made numerous attempts to address multipath and multi-antenna environments. For example, invention patent CN119520211A proposes a MIMO-OFDM time-frequency synchronization method suitable for strong multipath channels of UAVs. This method achieves synchronization by designing an orthogonal preamble sequence independent of the number of antennas at the transmitting end and jointly performing the strongest path search at the receiving end. However, such methods require changes to the preamble structure of standard protocols and are incompatible with the cyclic shift diversity (CSD) mechanism forcibly introduced in existing standards such as IEEE 802.11n.

[0004] When dealing with the complex air-to-ground link scenarios described above, traditional symbol timing synchronization algorithms reveal significant limitations in adaptability: (1) Traditional continuous sliding cross-correlation mechanisms exhibit multiple peaks under standard protocol (such as IEEE 802.11n) frame formats: Traditional synchronization schemes rely entirely on the local sequence and continuously received data for single-dimensional sliding matching. This complete reliance on conventional continuous sliding mechanisms leads to ambiguity in the matching between the cyclic prefix (CP) and the periodic characteristics of the training sequence, resulting in multiple correlation peaks, such as... Figure 1 As shown. This results in an extremely high risk of alignment errors at the receiving end, and time-domain synchronization is completely out of control.

[0005] (2) Traditional cross-correlation results cannot be adequately assessed by using a fixed threshold, as this leads to CSD and multipath-induced cross-correlation peak attenuation in standard protocols (such as the two-antenna mode of the IEEE 802.11n protocol). If both antennas simultaneously transmit identical baseband signals (standard L-LTFs with the same frequency and phase), these coherent radio frequency electromagnetic waves will coherently superimpose in space. Therefore, the IEEE 802.11n protocol introduces CSD in the two-antenna architecture (the first antenna remains unchanged, and the first 8 points (-200ns) of each L-LTF symbol of the second antenna are moved to the end of the L-LTF symbol, while the rest are moved forward by 8 points). Due to the CSD of the second antenna causing some signals to advance, an additional peak will appear in the first 8 points of the main peak in the cross-correlation results, such as... Figure 2 As shown. Meanwhile, under multipath channel conditions, the cross-correlation peak further splits, and the peak value further decreases, as... Figure 3 As shown. If a fixed decision threshold is used, on the one hand, if the threshold is too high, synchronization will be missed under CSD and severe multipath conditions; on the other hand, if the threshold is too low, the low peak caused by the cyclic prefix (CP) will be incorrectly identified when the channel conditions are good. Figure 1 The first low point in the signal is the moment the signal arrives. Summary of the Invention

[0006] To overcome the shortcomings of the prior art, the present invention aims to provide a symbol timing synchronization method, system, device, medium, and program product for the receiver in a MIMO-OFDM architecture. By introducing a time-reconstruction mechanism in the receiver to construct a time-reconstruction input vector, and deeply integrating multi-antenna joint cross-correlation metric and asymmetric dual sliding window integral energy ratio decision, a symbol timing synchronization framework that does not require dynamic adaptive adjustment of the absolute threshold, is resistant to multipath fading, and can converge accurately is formed. It has the characteristics of a unique main peak and an extremely sharp main peak under any channel conditions (easy to set the threshold value).

[0007] To achieve the above objectives, the technical solution adopted by the present invention is as follows: A symbol timing synchronization method for the receiver in a MIMO-OFDM architecture includes the following steps: Step 1: The transmitting baseband generates multiple-input multiple-output orthogonal frequency division multiplexing (OFDM) data frames. The transmitting physical layer preamble contains a standardized long training sequence for synchronization. The transmitting OFDM data frames are mapped to different spatial streams and cyclic shift diversity is forcibly introduced. Then, the OFDM data frames with cyclic shift diversity are passed to multiple transmit antennas. Subsequently, each transmit antenna transmits the radio frequency signal with cyclic shift diversity characteristics to the receiving end through a long-distance air-to-ground channel containing strong multipath delay spread. Step 2: After receiving the radio frequency signal with cyclic shift diversity characteristics, the receiver performs radio frequency front-end processing and coarse frequency offset compensation. The multiple receiving antennas of the receiver obtain discrete received signals. The receiver splices the discrete received signals across time to obtain the time-reconstructed input vector. Step 3: Perform cross-correlation matching operation between the time-reconstructed input vector and the locally pre-stored long training sequence, and merge the maximum ratio of the cross-correlation results of all receiving antennas to output a single joint cross-correlation metric. Step 4: The receiver constructs two continuous and asymmetric sliding windows along the time axis based on the joint cross-correlation metric. The sliding windows include an energy capture window and a noise reference window. Step 5: The receiver calculates the ratio of the integrated energy of the energy capture window to the noise reference window, and then performs a peak search on the ratio of the integrated energy based on a preset fixed empirical threshold to lock the end of the long training sequence, and uses the end of the long training sequence as a globally unique symbol timing synchronization point.

[0008] In step 1, the specific steps for forcibly introducing cyclic shift diversity are as follows: Let the first The time-domain transmitted signal of the root transmitting antenna is The introduced cyclic shift diversity sample delay is Then the specific operation of the sending end on the long training sequence is equivalent to: in, For the baseband time-domain standard discrete sequence at the transmitting end, , For the set of complex numbers, For the number of points in a long training sequence, For discrete sampling times.

[0009] In step 2, the first of the plurality of receiving antennas The discrete received signal of the antenna is At the current discrete sampling time As the reference anchor point for the sliding window, an independent anchor point of length is constructed for each receiving antenna. Time-separated reconstruction of input vector Before reconstructing the input vector The points are taken from the lag. Historical data of points, later Each point is taken from current real-time data, and its mathematical mapping relationship is defined as follows: in, Represents the time-reconstructed input vector Internal discrete sampling point index.

[0010] In step 3, the specific steps of the cross-correlation matching operation are as follows: Time-series reconstruction of the input vector Compared with the locally pre-stored standard discrete sequence of the transmitter baseband time domain Perform cross-correlation calculations and define the joint cross-correlation metric as follows: The calculation formula is as follows: Where * denotes the conjugate operation. Represents the time-reconstructed input vector Internal discrete sampling point index.

[0011] In step 4, the length of the energy capture window Covering the maximum cyclic shift diversity (CSD) delay at the transmitter, locked to the current moment, the joint cross-correlation metric accumulated within the energy capture window. Represented as: in, For discrete sampling times, for The index value of the inner joint cross-correlation metric of the time-matter energy capture window. For joint cross-correlation measurement; The joint cross-correlation metric of the noise reference window accumulated energy capture window over a period of time prior to the noise reference window, and the length of the noise reference window. The joint cross-correlation metric accumulated within the noisy reference window is at least twice the maximum cyclic shift diversity (CSD) delay at the transmitter and less than the period of the long training sequence (L-LTF). Represented as: .

[0012] In step 5, the receiver calculates the ratio of the transient energy capture window to the noise reference window as the final symbol synchronization decision metric. : in, The smallest positive real number preset by the system. This refers to the accumulated cross-correlation metric within the energy capture window. This is the accumulated cross-correlation metric within the noise reference window; Based on a preset fixed threshold Symbolic synchronization decision measurement Perform peak search to extract symbol timing synchronization points: in, For symbol timing synchronization point, For all discrete sampling times In the search for symbolic synchronization decision metrics The discrete sampling time corresponding to the attainment of the maximum value .

[0013] A synchronization system based on the symbol timing synchronization method at the receiver in a MIMO-OFDM architecture includes: The radio frequency signal generation module generates multiple-input multiple-output orthogonal frequency division multiplexing (OFDM) data frames at the transmitting end baseband. The transmitting end physical layer preamble contains a standardized long training sequence for synchronization. The transmitting end OFDM data frames are mapped to different spatial streams and cyclic shift diversity is forcibly introduced. Then, the OFDM data frames with cyclic shift diversity are transmitted to multiple transmit antennas. Subsequently, each transmit antenna transmits the radio frequency signal with cyclic shift diversity characteristics to the receiving end through a long-distance air-to-ground channel containing strong multipath delay spread. The radio frequency signal time-separation module performs radio frequency front-end processing and coarse frequency offset compensation after receiving a radio frequency signal with cyclic shift diversity characteristics. Multiple receiving antennas at the receiving end obtain discrete received signals, and the receiving end performs time-separation splicing on the discrete received signals to obtain a time-separation reconstruction input vector. The joint cross-correlation metric extraction module performs cross-correlation matching operations on the time-series reconstructed input vector and the locally pre-stored long training sequence, merges the maximum ratio of the cross-correlation results of all receiving antennas, and outputs a single joint cross-correlation metric. The dual sliding window construction module constructs two continuous and asymmetric sliding windows along the time axis based on the joint cross-correlation metric at the receiver. The sliding windows include an energy capture window and a noise reference window. The symbol timing synchronization point acquisition module calculates the ratio of the integrated energy of the energy capture window to the noise reference window at the receiving end, and then performs peak search on the ratio of the integrated energy based on a preset fixed empirical threshold to lock the end of the long training sequence, and uses the end of the long training sequence as the globally unique symbol timing synchronization point.

[0014] A symbol timing synchronization device for the receiver in a MIMO-OFDM architecture includes: Memory: Used to store computer programs that implement symbol timing synchronization methods at the receiver in a MIMO-OFDM architecture; Processor: Used to implement the symbol timing synchronization method at the receiver in the MIMO-OFDM architecture when executing the computer program.

[0015] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of a symbol timing synchronization method for a receiver in a MIMO-OFDM architecture.

[0016] A computer program product includes a computer program that, when executed by a processor, implements a symbol timing synchronization method for the receiver in a MIMO-OFDM architecture.

[0017] Compared with the prior art, the beneficial effects of the present invention are as follows: First, this invention overcomes the physical limitations of traditional continuous sliding correlation, fundamentally eliminating the multi-peak false alarm interference unique to multiple-input multiple-output (MIMO) mechanisms. Addressing the peak splitting problem caused by cyclic shift diversity (CSD) and cyclic prefix (CP) in MIMO-OFDM architectures, this invention proposes constructing a time-reconstructed input vector based on a long training sequence (L-LTF). The time-reconstructed input vector is cross-correlated with the local training sequence, ensuring that the received signal only produces a unique and significant main peak when the input vector is perfectly aligned with the local sequence. This completely eliminates the spurious peaks generated by traditional continuous sliding matching, achieving high-precision, unambiguous alignment of temporal resources in complex air-to-ground links.

[0018] Secondly, this invention, relying on an asymmetric dual-sliding-window energy integration architecture, achieves ultimate re-aggregation of multipath-spreading energy and maximizes diversity gain. Addressing the widespread temporal energy dispersion caused by multipath delay spread in long-range UAV communication, this invention completely overcomes the shortcomings of traditional single-window architectures, which are prone to energy leakage. By constructing an energy capture window and a noise reference window, and combining it with maximum ratio combining (MRC) of multiple antennas, it successfully re-aggregates severely dispersed temporal energy, greatly improving the effective signal detection probability and anti-fading capability under extremely low signal-to-noise ratio (SNR) conditions.

[0019] Third, this invention, based on a ratio decision model that senses abrupt changes in relative signal-to-noise ratio (SNR), possesses extremely high threshold robustness against disturbances and engineering practicality. Traditional synchronization algorithms heavily rely on absolute power thresholds, which fluctuate dramatically in the highly dynamic and non-stationary channels of UAVs, easily leading to threshold failure and false alarms. This invention utilizes the ratio of the integrated energy of the energy capture window to the noise reference window as a global synchronization decision metric, effectively offsetting the dramatic fluctuations in the absolute power of the received signal. This mechanism transforms the traditional "absolute energy detection" into "relative SNR abrupt change detection," allowing for accurate convergence using only a fixed threshold without complex dynamic evaluation of the noise floor or adaptive adjustment of the absolute threshold. This fundamentally eliminates the problem of subsequent Fast Fourier Transform (FFT) window failure caused by temporal misalignment.

[0020] In summary, this invention achieves high-precision, unambiguous alignment of time-domain resources in complex air-to-ground links, while greatly improving the effective signal detection probability and anti-fading capability under extreme conditions of extremely low signal-to-noise ratio (SNR), and possesses extremely high threshold disturbance robustness and engineering practicality. Attached Figure Description

[0021] Figure 1 This is a traditional continuous cross-correlation peak diagram under ideal channel conditions.

[0022] Figure 2 To introduce CSD under ideal channel conditions, a conventional continuous cross-correlation peak plot is used.

[0023] Figure 3 To introduce CSD in multipath channel conditions, a conventional continuous cross-correlation peak plot is used.

[0024] Figure 4 This is a flowchart of the method described in this invention.

[0025] Figure 5 This is a peak cross-correlation diagram of the method described in this invention under ideal channel conditions.

[0026] Figure 6 To introduce CSD in multipath channel conditions, the present invention describes a method for intertemporal cross-correlation peak plot.

[0027] Figure 7 A ratio diagram of the dual sliding window integral energy in the method described in this invention is provided to introduce CSD in multipath channel conditions.

[0028] Figure 8 This is a flowchart of constructing the time-series reconstruction input vector in the method described in this invention.

[0029] Figure 9 This is a comparison chart of the symbol timing missing synchronization probability between the method described in this invention and existing traditional algorithms under different root mean square delay spreads. Detailed Implementation

[0030] The present invention will now be described in detail with reference to the accompanying drawings.

[0031] The symbol timing synchronization method for the receiver in a MIMO-OFDM architecture described in this invention is based on the IEEE 802.11n protocol standard and is applied to long-range, low signal-to-noise ratio UAV air-to-ground broadband data links. This embodiment adopts a two-transmit, two-receive (2T2R) architecture. The transmitter is configured with a dual-antenna XC7Z035 ZYNQ board, and the receiver is configured with a dual-antenna UAV. The relationship between the transmitter's antenna configuration and the CSD is shown in Table 1. Table 2 shows the real and imaginary parts of the long training sequence (L-LTF). Under the IEEE 802.11n protocol standard, the local baseband time-domain standard discrete sequence of the transmitter's long training sequence (L-LTF) is denoted as... In this embodiment, considering the integrity of multi-antenna joint processing, the sequence length... 128 sampling points were selected.

[0032] Table 1 Antenna serial number Loop delay duration / ns 1 0 2 -200 Table 2 Real and Imaginary Parts of L-LTF Sequences Sampling sequence number Real part virtual part Sampling sequence number Real part virtual part 1 693 693 65 -693 -693 2 -701 -169 66 195 727 3 -511 557 67 -488 379 4 529 -806 68 927 -32 5 670 -317 69 877 -63 6 -532 -439 70 -471 493 7 -62 -797 71 -103 919 8 -260 598 72 54 -863 9 -30 217 73 226 -251 10 -494 1 74 -274 -455 11 -655 124 75 -5 -662 12 230 -387 76 -411 209 13 -266 -756 77 -656 -474 14 222 910 78 653 777 15 -301 641 79 614 467 16 681 -112 80 -712 141 17 548 318 81 -825 514 18 -225 -433 82 706 -364 19 -219 -255 83 347 154 20 489 -99 84 -18 -604 21 515 -506 85 93 -628 22 -64 804 86 -139 224 23 397 817 87 -371 -246 24 -265 -530 88 717 346 25 368 -151 89 775 -45 26 -741 -88 90 -325 603 27 -327 -194 91 293 872 28 -486 355 92 -486 -718 29 -813 613 93 -79 -428 30 288 -733 94 -469 307 31 -547 -511 95 -600 273 32 894 125 96 284 5 33 555 0 97 0 555 34 -5 -284 98 -125 -894 35 273 -600 99 -511 -547 36 -307 469 100 733 -288 37 -428 -79 101 613 -813 38 718 486 102 -355 486 39 872 293 103 -194 -327 40 -603 325 104 88 741 41 -45 775 105 -151 368 42 -346 -717 106 530 265 43 -246 -371 107 817 397 44 -224 139 108 -804 64 45 -628 93 109 -506 515 46 604 18 110 99 -489 47 154 347 111 -255 -219 48 364 -706 112 433 225 49 514 -825 113 318 548 50 -141 712 114 112 -681 51 467 614 115 641 -301 52 -777 -653 116 -910 -222 53 -474 -656 117 -756 -266 54 -209 411 118 387 -230 55 -662 -5 119 124 -655 56 455 274 120 -1 494 57 -251 226 121 217 -30 58 863 -54 122 -598 260 59 919 -103 123 -797 -62 60 -493 471 124 439 532 61 -63 877 125 -317 670 62 32 -927 126 806 -529 63 379 -488 127 557 -511 64 -727 -195 128 169 701 like Figure 4 As shown, the method of the present invention includes the following steps: Step 1: The transmitting baseband generates multiple-input multiple-output orthogonal frequency division multiplexing (MIMO-OFDM) data frames. The transmitting physical layer preamble contains a long training sequence (L-LTF) for synchronization. To avoid accidental beam cancellation in space when multiple antennas transmit the same signal, the transmitting orthogonal frequency division multiplexing data frames are mapped to different spatial streams and a cyclic shift diversity delay offset is forcibly introduced. Then, the orthogonal frequency division multiplexing data frames with cyclic shift diversity are passed to two transmit antennas. Subsequently, each transmit antenna transmits the radio frequency signal with cyclic shift diversity characteristics to the receiving end through a long-distance air-to-ground channel containing strong multipath delay spread. Step 2: After receiving the radio frequency signal with cyclic shift diversity characteristics, the receiver performs radio frequency front-end processing and coarse frequency offset compensation. The multiple receiving antennas of the receiver obtain discrete received signals. In order to avoid the multiple correlation peaks caused by the cyclic shift diversity (CSD) offset of the transmitter, the periodic repetition characteristics of the long training sequence, and the cyclic prefix (CP) in the discrete received signals, the receiver performs time-crossing splicing on the discrete received signals to obtain the time-crossing reconstructed input vector. This time-crossing splicing mechanism breaks the original continuity of the received signal from the perspective of physical alignment logic. Step 3: Perform cross-correlation matching operation between the time-series reconstructed input vector and the locally pre-stored long training sequence (L-LTF). Since the signals from the two receiving antennas have experienced different channels, to maximize the utilization of the information from both receiving antennas, the cross-correlation results of all receiving antennas are combined using maximum ratio combining (MRC), outputting a single joint cross-correlation metric. Due to the strong constraints of the time-series splicing mechanism, this joint cross-correlation metric successfully eliminates multiple main peaks caused by the repetitive characteristics of L-LTF in the standard protocol (802.11n), resulting in only a single main peak, such as... Figure 5 As shown; Step 4: In response to the CSD introduced in the MIMO architecture of the standard protocol (802.11n) and the problem of peak attenuation caused by severe dispersion of effective signal energy in the strong multipath environment of UAVs, the receiver constructs two continuous and asymmetric sliding windows along the time axis based on the joint cross-correlation metric. The sliding windows include an energy capture window and a noise reference window. Step 5: The receiver calculates the ratio of the integrated energy of the energy capture window to the integrated energy of the noise reference window. This ratio of integrated energy makes the correlation peaks sharper and is independent of signal power (e.g., ...). Figure 6 To jointly measure cross-correlation, Figure 7 (The result is the ratio of the double sliding window). The receiving end then performs a peak search on the ratio of the integral energy based on a preset fixed empirical threshold to lock the end of the long training sequence, and uses the end of the long training sequence as the globally unique symbol timing synchronization point.

[0033] In step 1, the specific steps for forcibly introducing cyclic shift diversity are as follows: To prevent accidental beam cancellation (i.e., deep fading nulls) from multiple transmit antennas transmitting the same preamble sequence in the spatial channel, the baseband processing unit forces cyclic shift diversity (CSD) onto the physical layer preambles mapped to different spatial streams and transmit antennas.

[0034] Let the first The time-domain transmitted signal of the root transmitting antenna is In this embodiment The introduced cyclic shift diversity (CSD) sample delay is Then, the specific operation of the sending end on the long training sequence (L-LTF) is equivalent to: in, For the baseband time-domain standard discrete sequence at the transmitting end, , For the set of complex numbers, For the number of points in a long training sequence (L-LTF), For discrete sampling times.

[0035] In this embodiment, under the 40MHz bandwidth of the 802.11n standard, This manifests as a specific nanosecond-level bias, with a value of -8, precisely because The introduction of this technology, along with the fact that each transmit antenna generates multipath superposition of radio frequency signals with cyclic shift diversity characteristics through a long-distance air-to-ground channel containing strong multipath delay spread, leads to uncontrollable multi-peak energy splitting when conventional receivers perform continuous sliding correlation.

[0036] In step 2, the first of the plurality of receiving antennas The discrete received signal of the antenna is In this embodiment ,like Figure 8 As shown, the reconstruction process of the time-spanning input vector is illustrated. To avoid the multi-peak artifacts caused by the CP and the multi-cycle repetition characteristics of the training sequence, the receiver introduces a time-spanning splicing mechanism in the time domain, using the current discrete sampling time... As the reference anchor point for the sliding window, an independent anchor point of length is constructed for each receiving antenna. Time-separated reconstruction of input vector In this embodiment Before reconstructing the input vector The points are taken from the lag. Historical data of points, later Each point is taken from current real-time data, and its mathematical mapping relationship is defined as follows: in, Represents the time-reconstructed input vector Internal discrete sampling point index.

[0037] This time-reconstruction mechanism physically disrupts the original continuity of the received signal. Only when... When accurately aligned to the end of the second actual received L-LTF symbol, the input vector is reconstructed over time. The first half of the sequence will exactly match the first half of the first L-LTF of the transmitted sequence, and the second half will exactly match the second half of the second L-LTF. This non-continuous spatiotemporal matching condition fundamentally prevents the possibility of multi-peak matching caused by the cyclical nature of the sequence.

[0038] In step 3, the specific steps of the cross-correlation matching operation are as follows: Time-series reconstruction of the input vector Compared with the locally pre-stored standard discrete sequence of the transmitter baseband time domain Perform cross-correlation calculations and define the joint cross-correlation metric as follows: The calculation formula is as follows: Where * denotes the conjugate operation. Represents the time-reconstructed input vector Internal discrete sampling point index.

[0039] The joint metric output has only one significant main peak in the time domain. At other times, due to the physical misalignment caused by time-series splicing, the cross-correlation energy is greatly suppressed, thus filtering out spurious peak ambiguities caused by the repetition of CP and L-LTF.

[0040] In step 4, the energy capture window length is determined in order to collect peak dispersion caused by maximum cyclic shift diversity (CSD) and multipath. The maximum cyclic shift diversity (CSD) delay at the transmitting end is locked to the current time and used to accumulate the effective signal energy at the current time due to time-domain tailing caused by CSD and multipath fading. The joint cross-correlation metric accumulated within the energy capture window is also included. Represented as: in, For discrete sampling times, for The index value of the inner joint cross-correlation metric of the time-matter energy capture window. For joint cross-correlation measurement; The joint cross-correlation metric of the noise reference window accumulated energy capture window over a period of time prior to the noise reference window, and the length of the noise reference window. It is at least twice the maximum cyclic shift diversity (CSD) delay length at the transmitter and less than the long training sequence (L-LTF) period, used for real-time extraction of background interference and noise floor energy of the local channel, and the joint cross-correlation metric accumulated within the noise reference window. Represented as: .

[0041] In step 5, to eliminate the severe non-stationary fluctuations in received absolute power caused by the high-dynamic flight of the UAV, the receiver calculates the ratio of the transient energy capture window to the noise reference window as the final symbol synchronization decision metric. : in, A very small positive real number is preset for the system to prevent the denominator from approaching zero abnormally during extremely high signal-to-noise ratios or pure silent periods. This refers to the accumulated cross-correlation metric within the energy capture window. The cross-correlation metric accumulated within the noise reference window is used to offset the non-stationary fluctuations in the absolute power of the received signal through ratio calculation, resulting in a steep relative signal-to-noise ratio peak when the signal slides into the acquisition window.

[0042] Based on a preset fixed threshold Symbolic synchronization decision measurement Perform peak search to extract symbol timing synchronization points: In this embodiment, a fixed threshold is used. The default value is 5; in, For symbol timing synchronization point, For all discrete sampling times In the search for symbolic synchronization decision metrics The discrete sampling time corresponding to the attainment of the maximum value .

[0043] A synchronization system based on the symbol timing synchronization method at the receiver in a MIMO-OFDM architecture includes: The radio frequency signal generation module generates multiple-input multiple-output orthogonal frequency division multiplexing (OFDM) data frames at the transmitting end baseband. The transmitting end physical layer preamble contains a standardized long training sequence for synchronization. The transmitting end OFDM data frames are mapped to different spatial streams and cyclic shift diversity is forcibly introduced. Then, the OFDM data frames with cyclic shift diversity are transmitted to multiple transmit antennas. Subsequently, each transmit antenna transmits radio frequency signals with cyclic shift diversity characteristics to the receiving end through a long-distance air-to-ground channel containing strong multipath delay spread, thereby implementing step 1 of the method described in this invention. The radio frequency signal time-separation module, after the receiver receives the radio frequency signal with cyclic shift diversity characteristics, performs radio frequency front-end processing and coarse frequency offset compensation. The multiple receiving antennas of the receiver obtain discrete received signals. The receiver performs time-separation splicing on the discrete received signals to obtain a time-separation reconstruction input vector, which is used to implement step 2 of the method described in this invention. The joint cross-correlation metric extraction module performs cross-correlation matching operations on the time-series reconstructed input vector and the locally pre-stored long training sequence, merges the maximum ratio of the cross-correlation results of all receiving antennas, and outputs a single joint cross-correlation metric to implement step 3 of the method described in this invention. The dual sliding window construction module constructs two continuous and asymmetric sliding windows along the time axis based on the joint cross-correlation metric at the receiving end. The sliding windows include an energy capture window and a noise reference window, which are used to implement step 4 of the method described in this invention. The symbol timing synchronization point acquisition module calculates the ratio of the integrated energy of the energy capture window to the noise reference window at the receiving end, and then performs peak search on the ratio of the integrated energy based on a preset fixed empirical threshold to lock the end of the long training sequence. The end of the long training sequence is used as a globally unique symbol timing synchronization point to implement step 5 of the method described in this invention.

[0044] A symbol timing synchronization device for the receiver in a MIMO-OFDM architecture includes: Memory: Used to store computer programs that implement symbol timing synchronization methods at the receiver in a MIMO-OFDM architecture; Processor: Used to implement the symbol timing synchronization method at the receiver in the MIMO-OFDM architecture when executing the computer program.

[0045] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of a symbol timing synchronization method for a receiver in a MIMO-OFDM architecture.

[0046] A computer program product includes a computer program that, when executed by a processor, implements a symbol timing synchronization method for the receiver in a MIMO-OFDM architecture.

[0047] Experimental Analysis The root mean square delay spread (RMS delay spread) is used to measure the strength of multipath channels; a larger value indicates a more severe multipath channel. Under high signal-to-noise ratio (SNR) (20 dB) and with an RMS delay spread of 20–300 ns, simulations were performed to compare the missed synchronization probabilities of the traditional algorithm and our proposed algorithm. The results are as follows: Figure 9 As shown in the simulation results, the comparison curves clearly show that the traditional algorithm has a missed synchronization probability of about 4% when the root mean square delay spread is small (20ns); as the root mean square delay spread gradually increases, the missed synchronization probability exceeds 30% at 300ns. In contrast, the present invention not only has an extremely low initial missed synchronization probability (about 1.5%), but also maintains a missed synchronization probability of around 5% even with a root mean square delay spread of 300ns, greatly suppressing the overall trend of worsening missed detection rate. This simulation demonstrates that, compared to traditional synchronization mechanisms, the present invention possesses extremely significant anti-interference capabilities and symbol timing synchronization stability in the face of highly dynamic and multipath delay dispersion environments of UAVs.

Claims

1. A symbol timing synchronization method at the receiver in a MIMO-OFDM architecture, characterized in that, Includes the following steps: Step 1: The transmitting baseband generates multiple-input multiple-output orthogonal frequency division multiplexing (OFDM) data frames. The transmitting physical layer preamble contains a standardized long training sequence for synchronization. The transmitting OFDM data frames are mapped to different spatial streams and cyclic shift diversity is forcibly introduced. Then, the OFDM data frames with cyclic shift diversity are passed to multiple transmit antennas. Subsequently, each transmit antenna transmits the radio frequency signal with cyclic shift diversity characteristics to the receiving end through a long-distance air-to-ground channel containing strong multipath delay spread. Step 2: After receiving the radio frequency signal with cyclic shift diversity characteristics, the receiver performs radio frequency front-end processing and coarse frequency offset compensation. The multiple receiving antennas of the receiver obtain discrete received signals. The receiver splices the discrete received signals across time to obtain the time-reconstructed input vector. Step 3: Perform cross-correlation matching operation between the time-reconstructed input vector and the locally pre-stored long training sequence, and merge the maximum ratio of the cross-correlation results of all receiving antennas to output a single joint cross-correlation metric. Step 4: The receiver constructs two continuous and asymmetric sliding windows along the time axis based on the joint cross-correlation metric. The sliding windows include an energy capture window and a noise reference window. Step 5: The receiver calculates the ratio of the integrated energy of the energy capture window to the noise reference window, and then performs a peak search on the ratio of the integrated energy based on a preset fixed empirical threshold to lock the end of the long training sequence, and uses the end of the long training sequence as a globally unique symbol timing synchronization point.

2. The method according to claim 1, characterized in that, In step 1, the specific steps for forcibly introducing cyclic shift diversity are as follows: Let the first The time-domain transmitted signal of the root transmitting antenna is The introduced cyclic shift diversity sample delay is Then the specific operation of the sending end on the long training sequence is equivalent to: in, For the baseband time-domain standard discrete sequence at the transmitting end, , For the set of complex numbers, For the number of points in a long training sequence, For discrete sampling times.

3. The method according to claim 1, characterized in that, In step 2, the first of the plurality of receiving antennas The discrete received signal of the antenna is At the current discrete sampling time As the reference anchor point for the sliding window, an independent anchor point of length is constructed for each receiving antenna. Time-separated reconstruction of input vector Before reconstructing the input vector The points are taken from the lag. Historical data of points, later Each point is taken from current real-time data, and its mathematical mapping relationship is defined as follows: in, Represents the time-reconstructed input vector Internal discrete sampling point index.

4. The method according to claim 1, characterized in that, In step 3, the specific steps of the cross-correlation matching operation are as follows: Time-series reconstruction of the input vector Compared with the locally pre-stored standard discrete sequence of the transmitter baseband time domain Perform cross-correlation calculations and define the joint cross-correlation metric as follows: The calculation formula is as follows: Where * denotes the conjugate operation. Represents the time-reconstructed input vector Internal discrete sampling point index.

5. The method according to claim 1, characterized in that, In step 4, the length of the energy capture window Covering the maximum cyclic shift diversity (CSD) delay at the transmitter, locked to the current moment, the joint cross-correlation metric accumulated within the energy capture window. Represented as: in, For discrete sampling times, for The index value of the inner joint cross-correlation metric of the time-matter energy capture window. For joint cross-correlation measurement; The joint cross-correlation metric of the noise reference window accumulated energy capture window over a period of time prior to the noise reference window, and the length of the noise reference window. The joint cross-correlation metric accumulated within the noisy reference window is at least twice the maximum cyclic shift diversity (CSD) delay at the transmitter and less than the period of the long training sequence (L-LTF). Represented as: 。 6. The method according to claim 1, characterized in that, In step 5, the receiver calculates the ratio of the transient energy capture window to the noise reference window as the final symbol synchronization decision metric. : in, The smallest positive real number preset by the system. This refers to the accumulated cross-correlation metric within the energy capture window. This is the accumulated cross-correlation metric within the noise reference window; Based on a preset fixed threshold Symbolic synchronization decision measurement Perform peak search to extract symbol timing synchronization points: in, For symbol timing synchronization point, For all discrete sampling times In the search for symbolic synchronization decision metrics The discrete sampling time corresponding to the attainment of the maximum value .

7. A symbol timing synchronization system for the receiver in a MIMO-OFDM architecture, characterized in that, The method described in any one of claims 1 to 6 includes: The radio frequency signal generation module generates multiple-input multiple-output orthogonal frequency division multiplexing (OFDM) data frames at the transmitting end baseband. The transmitting end physical layer preamble contains a standardized long training sequence for synchronization. The transmitting end OFDM data frames are mapped to different spatial streams and cyclic shift diversity is forcibly introduced. Then, the OFDM data frames with cyclic shift diversity are transmitted to multiple transmit antennas. Subsequently, each transmit antenna transmits the radio frequency signal with cyclic shift diversity characteristics to the receiving end through a long-distance air-to-ground channel containing strong multipath delay spread. The radio frequency signal time-separation module performs radio frequency front-end processing and coarse frequency offset compensation after receiving a radio frequency signal with cyclic shift diversity characteristics. Multiple receiving antennas at the receiving end obtain discrete received signals, and the receiving end performs time-separation splicing on the discrete received signals to obtain a time-separation reconstruction input vector. The joint cross-correlation metric extraction module performs cross-correlation matching operations on the time-series reconstructed input vector and the locally pre-stored long training sequence, merges the maximum ratio of the cross-correlation results of all receiving antennas, and outputs a single joint cross-correlation metric. The dual sliding window construction module constructs two continuous and asymmetric sliding windows along the time axis based on the joint cross-correlation metric at the receiver. The sliding windows include an energy capture window and a noise reference window. The symbol timing synchronization point acquisition module calculates the ratio of the integrated energy of the energy capture window to the noise reference window at the receiving end, and then performs peak search on the ratio of the integrated energy based on a preset fixed empirical threshold to lock the end of the long training sequence, and uses the end of the long training sequence as the globally unique symbol timing synchronization point.

8. A symbol timing synchronization device for the receiver in a MIMO-OFDM architecture, characterized in that, include: Memory: for storing a computer program that implements the symbol timing synchronization method of the receiver in the MIMO-OFDM architecture as described in any one of claims 1 to 6; Processor: Used to implement the symbol timing synchronization method of the receiver in the MIMO-OFDM architecture as described in any one of claims 1 to 6 when executing the computer program.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the symbol timing synchronization method for the receiver in the MIMO-OFDM architecture as described in any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the symbol timing synchronization method for the receiver in the MIMO-OFDM architecture as described in any one of claims 1 to 6.

Images