A high-speed rail scenario automatic frequency control method and system
By adaptively adjusting the frequency control method and system in the high-speed rail scenario, and by using the set classification of frequency offset estimates and channel type identification, the problems of frequency synchronization accuracy and response speed in the high-speed rail scenario are solved, and the frequency tracking accuracy and response speed in the high-speed rail scenario are improved.
Patent Information
- Application Number
- CN202611095032.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-07-22
- Publication Date
- 2026-11-10
AI Technical Summary
Existing automatic frequency control schemes cannot meet the high-precision frequency synchronization requirements of high-speed rail scenarios, cannot autonomously identify single-frequency network scenarios, and cannot separate crystal oscillator temperature drift and Doppler frequency offset, resulting in the accumulation of compensation errors and deterioration of receiving performance.
By implementing the set classification, temperature drift estimation, and multipath identification of frequency offset estimates in the terminal equipment, the compensation strategy of the automatic frequency control loop is adaptively adjusted. The channel type is identified by utilizing the directional cancellation characteristics of Doppler components and power delay distribution data, and the AFC loop compensation strategy is adaptively configured.
It achieves effective separation of crystal oscillator temperature drift and Doppler frequency offset in high-speed rail scenarios, improves frequency offset tracking accuracy and response speed, adapts to optimal compensation in different channel environments, and improves receiving performance.
Smart Images

Figure CN123075386A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of automatic frequency control technology, and in particular to an automatic frequency control method and system for high-speed rail scenarios. Background Technology
[0002] Frequency synchronization is a prerequisite for the normal operation of the receiver in a mobile communication system. Because there is a frequency deviation between the local crystal oscillator of the terminal device and the transmission frequency of the base station, and the Doppler effect generated during the movement of the terminal device further introduces frequency offset, the superposition of these two types of deviations severely affects the terminal's signal reception performance. Currently, terminal devices typically estimate and compensate for frequency deviations through an Automatic Frequency Control (AFC) loop. The basic principle of the AFC loop is as follows: The frequency offset estimate of the current received signal is obtained through a Frequency Offset Estimation (FOE) module. After smoothing by a loop filter, the operating frequencies of the analog domain voltage-controlled oscillator (VCO) and the digital domain numerically controlled oscillator (NCO) are adjusted respectively. Through a closed-loop control mechanism, the frequency deviation of the received signal is compensated and corrected, thus achieving system frequency synchronization.
[0003] With the development of mobile communication technology, the 3rd Generation Partnership Project (3GPP) New Radio (NR) Release 15 (R15) standard began to support high-speed rail scenarios, and the 3GPP Release 16 (R16) and Release 17 (R17) standards further strengthened the support for high-speed rail scenarios. According to the specification definition of 3GPP protocol TS 38.101-4, terminal equipment must support three typical high-speed rail channel scenarios: High Speed Train (HST) channel (up to 500 km / h), High Speed Train-Single Frequency Network (HST-SFN) channel, and High Speed Train-Dynamic Point Selection (HST-DPS) channel.
[0004] Compared to conventional mobile communication scenarios, the high-speed movement in high-speed rail scenarios significantly amplifies the frequency offset problem of terminals, increasing the difficulty of frequency synchronization. In high-speed rail scenarios, when the communication subcarrier spacing is 15kHz, the system Doppler frequency offset can reach up to 870Hz, and when the subcarrier spacing is 30kHz, the Doppler frequency offset can reach up to 1677Hz. The actual frequency offset values far exceed the approximately 300Hz Doppler assumption supported by 3GPP in the initial design of NR. At the same time, the transmission characteristics of different high-speed rail channel scenarios are different. Among them, the HST channel and HST-DPS channel are single-path channels, while the HST-SFN channel is a typical multipath channel. In multipath channels, multiple signal direct paths are transmitted from different remote radio units to the terminal device. Its Doppler transmission characteristics are significantly different from the traditional Jakes fading model, increasing the complexity of frequency offset estimation.
[0005] Existing frequency synchronization compensation schemes cannot meet the high-precision frequency synchronization requirements of high-speed rail communication. For example, traditional automatic frequency control schemes use fixed-bandwidth loop filters and a single frequency offset estimation module. Assuming the Doppler frequency offset is within 300Hz, they cannot distinguish between temperature drift and Doppler components, nor can they differentiate between single-path and multipath scenarios. In high-speed rail scenarios, fixed bandwidth cannot simultaneously meet the requirements of fast tracking and low noise. When Doppler frequency offset changes drastically, the tracking lag of the automatic frequency control loop is significant, and traditional automatic frequency control loops can no longer meet the frequency offset tracking and compensation requirements of high-speed rail scenarios. Another example is the high-speed rail mode switching scheme based on high-level signaling, which relies solely on the high-speed rail scenario indication signaling configured on the network side to switch automatic frequency control parameters. It cannot autonomously identify high-speed rail scenarios and does not support bidirectional identification and temperature drift separation on a single-frequency network. When network signaling is unreliable or not updated in a timely manner, the terminal cannot adaptively adjust. Yet another example is the direct Doppler frequency offset compensation scheme, which directly uses the frequency offset estimate for VCO or NCO compensation, which cannot distinguish between temperature drift and Doppler components. In multipath scenarios, directly compensating for Doppler frequency offset may lead to overcompensation, which degrades reception performance.
[0006] In summary, traditional automatic frequency control loops, due to their fixed bandwidth, cannot adapt to the rapid changes in Doppler frequency offset in high-speed rail scenarios, resulting in insufficient frequency offset tracking accuracy and response speed. Existing solutions cannot autonomously identify single-frequency network scenarios, leading to ambiguous frequency offset estimation in multi-directional incoming wave conditions and an inability to separate Doppler components from different directions. Furthermore, existing solutions fail to effectively separate crystal oscillator temperature drift from Doppler frequency offset, preventing separate compensation and resulting in accumulated compensation errors. Finally, existing solutions cannot distinguish between single-path and multi-path scenarios, and employing a single-path compensation strategy in multi-path channels leads to performance degradation. Summary of the Invention
[0007] To address the above technical problems, this invention provides an automatic frequency control method for high-speed rail scenarios; on the other hand, it also provides an automatic frequency control system for high-speed rail scenarios.
[0008] The technical problem solved by this invention can be achieved by the following technical solutions: An automatic frequency control method for high-speed rail scenarios includes: Step S101: When the target terminal enters the high-speed rail scenario mode, determine whether to enter the single-frequency network mode based on the frequency offset estimation value reported by the target terminal, and classify the frequency offset estimation value into the first set and the second set according to the direction of arrival. Step S102: Average the frequency offset estimates in the first set and the second set to obtain the temperature drift estimate, wherein the Doppler components cancel each other out because they are in opposite directions; Step S103: Perform multipath identification based on the power delay distribution data reported by the target terminal to determine the channel type; Step S104: Adaptively adjust the compensation strategy of the automatic frequency control loop according to the channel type and the temperature drift estimate.
[0009] Preferably, the method further includes the following steps before step S101: Step S100: Determine whether the target terminal has entered the high-speed rail scenario mode; When it is determined that the target terminal has entered the high-speed rail scene mode, proceed to step S101; When it is determined that the target terminal has exited the high-speed rail scene mode, return to step S100.
[0010] Preferably, determining whether the target terminal has entered the high-speed rail scene mode includes: Step S1001A: Receive the first signaling, which is used to indicate whether the target terminal is in high-speed rail scenario mode; Step S1002A: Based on the first signaling, determine whether the target terminal is in high-speed rail scenario mode: When the first signaling instructs the target terminal to enter the high-speed rail scenario mode, proceed to step S101; When the first signaling instructs the target terminal to exit the high-speed rail scenario mode, proceed to step S100.
[0011] Preferably, determining whether the target terminal has entered the high-speed rail scene mode includes: Step S1001B: Receive the frequency offset estimate reported by the target terminal; Step S1002B: Calculate the standard deviation of the frequency offset estimate; Step S1003B: Determine whether the standard deviation of the frequency offset estimate meets the preset entry condition, wherein the preset entry condition is that the standard deviation of the frequency offset estimate is greater than the preset first threshold and the consecutive cumulative number is not less than the preset first cumulative number threshold. If the standard deviation of the frequency offset estimate meets the preset entry condition, then the target terminal is determined to enter the high-speed rail scenario mode and proceeds to step S101. If the standard deviation of the frequency offset estimate does not meet the preset entry condition, then return to step S1001B; After confirming that the target terminal has entered the high-speed rail scenario mode, the following steps are also included: Step S1004B: Determine whether the standard deviation of the frequency offset estimate meets the preset exit condition, wherein the preset exit condition is that the standard deviation of the frequency offset estimate is not greater than the preset first threshold and the cumulative number is not less than the preset second cumulative number threshold in the high-speed rail scenario mode. If the standard deviation of the frequency offset estimate meets the preset exit condition, then the target terminal is determined to exit the high-speed rail scenario mode and proceeds to step S100. If the standard deviation of the frequency offset estimate does not meet the preset exit condition, then return to step S1001B.
[0012] Preferably, step S101 includes: Step S1011: Based on the frequency offset estimation value reported by the target terminal, sort the received first frequency offset estimation value, second frequency offset estimation value, and third frequency offset estimation value from smallest to largest, calculate the difference between two adjacent frequency offset estimation values, and obtain the first gap and the second gap. The first gap is the difference between the second largest value and the smallest value among the first frequency offset estimation value, second frequency offset estimation value, and third frequency offset estimation value, and the second gap is the difference between the largest value and the second largest value among the first frequency offset estimation value, second frequency offset estimation value, and third frequency offset estimation value. Step S1012: Compare the first gap and the second gap, and compare the first gap and the second gap with a preset second threshold; If both the first gap and the second gap are less than or equal to the second threshold, then all frequency offset estimates are classified into the first set, the second set is empty, and it is determined that the target terminal has not entered the single-frequency network mode. If the maximum gap between the first gap and the second gap is greater than the second threshold, then the frequency offset estimates on both sides of the maximum gap are classified into the first set and the second set, respectively.
[0013] Preferably, after step S1012, the method further includes: In the single-frequency network mode, when a new frequency offset estimate is received, the absolute value of the difference between the new frequency offset estimate and the previous frequency offset estimate is compared with the second threshold and the preset third threshold. If the absolute value of the difference between the new frequency offset estimate and the previous frequency offset estimate is greater than the third threshold, then the new frequency offset estimate is classified into a different set than the set in which the previous frequency offset estimate is located. If the absolute value of the difference between the new frequency offset estimate and the previous frequency offset estimate is less than or equal to the second threshold, then the new frequency offset estimate is classified into the same set as the previous frequency offset estimate. If the absolute value of the difference between the new frequency offset estimate and the previous frequency offset estimate is greater than the second threshold and less than or equal to the third threshold, then the new frequency offset estimate is discarded.
[0014] Preferably, it further includes: When the target terminal enters the single-frequency network mode, monitor the last update time of the first set and the second set; If the time difference between the last update time of either the first set or the second set and the current time exceeds the preset maximum update interval threshold, the target terminal is determined to exit the single-frequency network mode, and the first set and the second set are cleared.
[0015] Preferably, the first set and the second set respectively use a first-in-first-out buffer to store the classified frequency offset estimates. The first-in-first-out buffer is configured to, upon receiving a newly classified frequency offset estimate, remove the oldest frequency offset estimate and keep the storage count unchanged if the current storage amount exceeds a preset storage limit, and increment the storage count if the current storage amount does not exceed the preset storage limit. The first set and the second set are respectively configured with sum accumulators. Each sum accumulator is connected to the first-in-first-out buffer of the corresponding set. The sum accumulator is configured to add the newly classified frequency offset estimate to the accumulation result and subtract the oldest frequency offset estimate removed by the first-in-first-out buffer.
[0016] Preferably, step S103 includes: Step S1031: Receive power delay distribution data reported by the target terminal, wherein the power delay distribution data includes power information and delay information of several paths; Step S1032: Remove paths whose power information is lower than the preset noise threshold; Step S1033: Calculate the delay difference between the delay information of each remaining effective path after elimination, and classify the effective paths with delay differences less than the preset intra-cluster threshold into the same cluster. Step S1034: Clusters whose minimum inter-cluster delay difference is greater than a preset inter-cluster threshold and whose power information of the representative path of each cluster is higher than a preset power threshold are determined as distinguishable clusters; wherein, the representative path is the path with the largest power information within the cluster. Step S1035: Determine the channel type based on the number of resolvable clusters; If there are two or more distinguishable clusters, the channel type is determined to be a multipath channel; If one or more of the resolvable clusters described above exist, the channel type is determined to be a single-path channel.
[0017] Preferably, in step S104, if the channel type is determined to be a single-path channel, the compensation amount in the compensation strategy includes a temperature drift component and a Doppler component; wherein, the temperature drift component is the temperature drift estimate in step S102, and the Doppler component is the difference between the frequency offset estimate and the temperature drift estimate. If the channel type is determined to be a multipath channel, then the compensation amount in the compensation strategy only includes the temperature drift component.
[0018] On the other hand, an automatic frequency control system for high-speed rail scenarios is provided, for implementing the automatic frequency control method for high-speed rail scenarios as described above, including: The single-frequency network bidirectional judgment module is used to determine whether to enter the single-frequency network mode based on the frequency offset estimation value reported by the target terminal when the target terminal enters the high-speed rail scenario mode, and classify the frequency offset estimation value into the first set and the second set according to the direction of arrival. The temperature drift estimation module is used to average the frequency offset estimates in the first set and the second set to obtain the temperature drift estimate, wherein the Doppler components cancel each other out because they are in opposite directions. The multipath identification module is used to perform multipath identification based on the power delay distribution data reported by the target terminal and determine the channel type. The compensation module is used to adaptively adjust the compensation strategy of the automatic frequency control loop according to the channel type and the temperature drift estimate.
[0019] The advantages or beneficial effects of the technical solution of this invention are as follows: This invention enables the terminal to identify the single-frequency network mode based on the frequency offset estimation value reported by the terminal when operating in a high-speed rail scenario. It collects and averages the frequency offset estimation values from different incoming wave directions to obtain the crystal oscillator temperature drift estimation value. It utilizes the characteristic of mutual cancellation of Doppler components in the positive and negative directions to achieve effective separation of crystal oscillator temperature drift and Doppler frequency offset. At the same time, it combines the power delay distribution data reported by the terminal to complete the identification of single-path / multipath channels. Finally, it adaptively configures the AFC loop compensation strategy based on the identified channel type and the decoupled temperature drift estimation value. It can adaptively select the optimal compensation strategy under different channel environments, thereby improving the frequency offset tracking accuracy and response speed in high-speed rail scenarios. Attached Figure Description
[0020] Figure 1This is a flowchart illustrating an automatic frequency control method in a high-speed rail scenario, as a preferred embodiment of the present invention. Figure 2 This is a flowchart illustrating an automatic frequency control method in a high-speed rail scenario, as shown in another preferred embodiment of the present invention. Figure 3 This is a flowchart illustrating an automatic frequency control method in a high-speed rail scenario, as shown in another preferred embodiment of the present invention. Figure 4 This is a flowchart illustrating an automatic frequency control method in a high-speed rail scenario, as shown in another preferred embodiment of the present invention. Figure 5 This is a schematic diagram of the initial clustering process in step S101, which is a preferred embodiment of the present invention. Figure 6 This is a schematic diagram of the non-first-time clustering process in step S101, as described in a preferred embodiment of the present invention. Figure 7 This is a schematic diagram of the process of exiting single-frequency network mode in a preferred embodiment of the present invention; Figure 8 This is a flowchart illustrating step S103 in a preferred embodiment of the present invention. Figure 9 This is a structural block diagram of an automatic frequency control system in a high-speed rail scenario, as shown in a preferred embodiment of the present invention. Detailed Implementation
[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0022] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other.
[0023] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, but this is not intended to limit the scope of the invention.
[0024] In a preferred embodiment of the present invention, based on the above-mentioned problems existing in the prior art, an automatic frequency control method for high-speed rail scenarios is provided, such as... Figure 1 As shown, it includes: Step S101: When the target terminal enters the high-speed rail scenario mode, determine whether to enter the single-frequency network mode based on the frequency offset estimation value reported by the target terminal, and classify the frequency offset estimation value into the first set and the second set according to the direction of arrival. Step S102: Average the frequency offset estimates in the first set and the second set to obtain the temperature drift estimate, where the Doppler components cancel each other out due to their opposite directions; Step S103: Perform multipath identification based on the power delay distribution data reported by the target terminal to determine the channel type; Step S104: Adaptively adjust the compensation strategy of the automatic frequency control loop according to the channel type and temperature drift estimate.
[0025] Specifically, the target terminal is a user terminal device that supports the NR high-speed rail communication protocol. The high-speed rail scenario mode is the high-speed mobile communication working mode pre-configured locally on the terminal.
[0026] In this embodiment, when the terminal is operating in the high-speed rail scenario mode, it can autonomously determine the single-frequency network mode based on the frequency offset estimation value reported by the terminal. It can collect and average the frequency offset estimation values from different incoming wave directions to obtain the crystal oscillator temperature drift estimation value. It can effectively separate the crystal oscillator temperature drift and Doppler frequency offset by utilizing the characteristic that the forward and reverse Doppler components cancel each other out. At the same time, it can identify single-path / multipath channels by combining the power delay distribution data reported by the terminal. Finally, it can adaptively configure the AFC loop compensation strategy based on the identified channel type and the decoupled temperature drift estimation value. It can adaptively select the optimal compensation strategy under different channel environments, thereby improving the frequency offset tracking accuracy and response speed in the high-speed rail scenario.
[0027] The working principle of this invention is as follows: Based on the frequency offset estimate, signal-to-noise ratio (SNR), and power delay distribution data periodically reported by the terminal's internal hardware, the terminal software sequentially performs high-speed rail scenario pattern recognition, single-frequency network mode discrimination and frequency offset estimate set classification, temperature drift estimation, and channel multipath identification. Finally, it adaptively matches the automatic frequency control (AFC) loop compensation strategy, enabling terminal devices that only support non-high-speed rail scenarios to support high-speed rail scenarios through software upgrades, thus solving the problems of frequency offset tracking lag and insufficient compensation in high-speed rail scenarios under traditional automatic frequency control strategies.
[0028] In this embodiment, high-speed rail scene recognition is included before step S101, such as... Figure 2 As shown, high-speed rail scene recognition includes the following steps: Step S100: Determine whether the target terminal has entered the high-speed rail scenario mode; When it is determined that the target terminal has entered the high-speed rail scene mode, proceed to step S101; When it is determined that the target terminal has exited the high-speed rail scene mode, return to step S100.
[0029] Specifically, in this embodiment, the terminal continuously executes the high-speed rail scene mode recognition process, determines the current working scene status in real time, and only starts the automatic frequency control method of the present invention after confirming that it has entered the high-speed rail scene mode; if it is in a non-high-speed rail scene mode, the traditional automatic frequency control scheme is used to achieve conventional frequency synchronization, and the high-speed rail scene mode recognition steps are continuously executed in a loop.
[0030] The terminal can activate the high-speed rail scenario mode by either high-level signaling or frequency offset fluctuation measurement. After activation, the terminal switches to high-speed mobile communication working mode, that is, the terminal enters the high-speed rail scenario mode.
[0031] In a preferred embodiment of the present invention, such as Figure 3 As shown, determining whether the target terminal has entered the high-speed rail scenario mode includes: Step S1001A: Receive the first signaling, which is used to indicate whether the target terminal is in high-speed rail scenario mode; In this step, the physical layer or media access control layer (MAC layer) of the target terminal receives Radio Resource Control (RRC) signaling or the first signaling from the MAC control layer control unit (MAC CE) on the network side. This first signaling is the high-speed rail scenario indication signaling issued by the network side, which carries configuration parameters for explicitly indicating whether the current cell or the current geographical area is a high-speed rail dedicated network (HST Network).
[0032] Furthermore, the first signaling may also include subtype information of the current HST scenario, such as whether it is used for HST-SFN scenario or HST-DPS scenario, so that the terminal can pre-configure the corresponding processing flow.
[0033] Step S1002A: Based on the first signaling, determine whether the target terminal is in high-speed rail scenario mode: When the first signaling instructs the target terminal to enter the high-speed rail scenario mode, proceed to step S101; When the first signaling instructs the target terminal to exit the high-speed rail scenario mode, proceed to step S100.
[0034] In this step, the target terminal responds to the received first signaling by parsing it and extracting the scene indication bit. If the parsing result indicates that the first signaling indicates the target terminal to enter the high-speed rail scene mode (e.g., the indication bit is set to "1"), the target terminal enters the high-speed rail scene mode. If the parsing result indicates that the first signaling indicates the target terminal to exit the high-speed rail scene mode (e.g., the indication bit is set to "0"), the target terminal neither enters nor exits the high-speed rail scene mode.
[0035] In another preferred embodiment of the invention, such as Figure 4 As shown, determining whether the target terminal has entered the high-speed rail scenario mode includes: Step S1001B: Receive the frequency offset estimate reported by the target terminal; Step S1002B: Calculate the standard deviation of the frequency offset estimate; Step S1003B: Determine whether the standard deviation of the frequency offset estimate meets the preset entry conditions, wherein the preset entry conditions are that the standard deviation of the frequency offset estimate is greater than the preset first threshold and the consecutive cumulative number is not less than the preset first cumulative number threshold. If the standard deviation of the frequency offset estimate meets the preset entry conditions, then the target terminal is determined to enter the high-speed rail scenario mode and proceeds to step S101. If the standard deviation of the frequency offset estimate does not meet the preset entry condition, return to step S1001B; After confirming that the target terminal has entered the high-speed rail scenario mode, the following steps are also included: Step S1004B: Determine whether the standard deviation of the frequency offset estimate meets the preset exit condition. The preset exit condition is that the standard deviation of the frequency offset estimate in the high-speed rail scenario mode is not greater than the preset first threshold and the cumulative number is not less than the preset second cumulative number threshold. If the standard deviation of the frequency offset estimate meets the preset exit condition, then the target terminal is determined to exit the high-speed rail scenario mode and proceeds to step S100. If the standard deviation of the frequency offset estimate does not meet the preset exit condition, return to step S1001B.
[0036] Specifically, in the high-speed rail scenario, the Doppler frequency offset changes drastically, and the fluctuation range of the frequency offset estimate is significantly greater than in the non-high-speed rail scenario. This invention uses the Exponentially Weighted Moving Standard Deviation (EWMSTD) as a measure of the degree of frequency offset fluctuation.
[0037] The methods for calculating the exponentially moving weighted standard deviation include: First, based on the frequency offset estimate at time k and the exponential moving average at time k-1, the residual deviation of the current frequency offset sample relative to the historical smoothed mean is calculated. This residual deviation is the difference between the current frequency offset estimate and the exponential moving average at the previous time, and its calculation formula is as follows:
[0038] in, This represents the frequency offset residual at time k, which characterizes the deviation of the current frequency offset sample from the historical smoothed mean. This represents the frequency offset estimate at time k. This represents the exponential moving average at time k-1, i.e., the historical smoothed mean.
[0039] Then, based on the exponential moving average at time k-1 and the frequency offset residual at time k, combined with a preset smoothing coefficient, iterative calculations are performed to update the current exponential moving average, the calculation formula of which is:
[0040] in, This represents the exponential moving average at time k. This represents the preset smoothing coefficient, used to assign a weighted proportion between the statistics of new samples and historical samples. New samples are given a higher weight, while the weight of historical samples decays exponentially, enabling the statistics to respond quickly to channel changes while maintaining constant computational complexity.
[0041] The relationship between the smoothing coefficient and the equivalent sliding window length of the EWMSTD algorithm is as follows:
[0042] in, This indicates the equivalent sliding window length.
[0043] Compared with the traditional sliding window method, this invention adopts the recursive method of Infinite Impulse Response (IIR) filter, which does not require storing historical sample windows. Each frequency offset estimate update only requires constant multiplication and addition operations, making it suitable for terminal devices with limited hardware resources.
[0044] Next, the exponential moving variance, smoothing coefficient, and the squared residual term from the previous time step are weighted and iterated to update the current exponential moving variance, which is calculated using the following formula:
[0045] in, This represents the exponentially moving standard deviation at time k. This represents the exponentially moving standard deviation at time k-1. This represents the exponential moving variance at time k.
[0046] Next, the arithmetic square root of the current exponential moving variance is calculated using the following formula:
[0047] The final output is the exponentially moving weighted standard deviation, which is the standard deviation of the frequency deviation estimate, and serves as a statistical measure of the fluctuation range of the frequency deviation estimate.
[0048] Based on this, the following two criteria apply to terminals entering high-speed rail scenarios; meeting either one is sufficient: (1) The high-level signaling indicates that the scenario is a high-speed rail scenario; (2) The standard deviation of the frequency offset estimate is greater than the preset first threshold TH1 and the cumulative occurrence is set.
[0049] In this embodiment, the second cumulative number threshold used to determine the exit condition of the high-speed rail scene is greater than the first cumulative number threshold used to determine the entry condition of the high-speed rail scene.
[0050] Specifically, this invention employs a hysteresis mechanism, comprising two independent counters: a first counter and a second counter. The first counter is a hysteresis counter for entering the high-speed rail scenario, configured to: increment its own count when the standard deviation of the frequency offset estimate is greater than a preset first threshold TH1, and reset its count when the standard deviation of the frequency offset estimate is less than or equal to the preset first threshold TH1, to prevent accidental triggering of the high-speed rail scenario mode due to instantaneous fluctuations. The second counter is a hysteresis counter for exiting the high-speed rail scenario, which starts counting only when the terminal has confirmed that it is in the high-speed rail scenario mode, configured to: increment its own count when it has entered the high-speed rail scenario mode and the standard deviation of the frequency offset estimate is less than or equal to the preset first threshold TH1.
[0051] In this invention, the first cumulative count threshold corresponding to the first counter is the high-speed rail scene entry hysteresis threshold count, representing the minimum number of consecutive samplings required for the terminal to formally switch to the high-speed rail scene mode and meet the threshold condition. The second cumulative count threshold corresponding to the second counter is the high-speed rail scene exit hysteresis threshold count, representing the minimum number of consecutive samplings required for the terminal to exit the high-speed rail scene mode and meet the threshold condition. This invention configures independent hysteresis determination counts for scene entry and scene exit, and the second cumulative count threshold corresponding to exit hysteresis is higher than the first cumulative count threshold corresponding to entry hysteresis. This forms a numerical difference hysteresis interval within the first threshold TH1 critical fluctuation range, suppressing frequent ping-pong switching between the high-speed rail scene mode and non-high-speed rail scene mode, and improving the stability of the scene recognition process.
[0052] Each time the terminal receives a newly reported frequency offset estimate, it first performs the aforementioned iterative calculation of the exponentially moving weighted standard deviation: sequentially solving for the current period's frequency offset residual, iteratively updating the exponentially moving average, and iteratively updating the exponentially moving variance, finally calculating and outputting the exponentially moving weighted standard deviation corresponding to the frequency offset estimate for this round; after completing the standard deviation calculation, the statistic is compared with the preset first threshold TH1, and counting and scene switching operations are performed according to the hierarchical judgment logic. The complete judgment process is as follows: If the exponentially moving weighted standard deviation of the frequency offset estimate in this round is greater than the first threshold TH1: the count value of the first counter is incremented by 1; then the current count value of the first counter is compared with the first cumulative count threshold. If the count value of the first counter is greater than or equal to the first cumulative count threshold, the terminal is determined to meet the high-speed rail scene entry condition and is switched to the high-speed rail scene mode; if the count value of the first counter is less than the first cumulative count threshold, the current scene state is maintained, and the determination process is repeated after the frequency offset estimate of the next frame is input.
[0053] If the exponentially moving weighted standard deviation of the frequency offset estimate in this round is less than or equal to the first threshold TH1: immediately reset the first counter value to zero; then check the current scene working state of the terminal; if the terminal is not in the high-speed rail scene mode, maintain the non-high-speed rail scene working state, wait for the next frame frequency offset estimate value input, and repeat the judgment process; if the terminal is already in the high-speed rail scene mode, control the second counter value to increment by 1; then compare the current counter value with the second cumulative count threshold. If the second counter value is greater than or equal to the second cumulative count threshold, determine that the terminal meets the high-speed rail scene exit condition and exits the high-speed rail scene mode; if the second counter value is less than the second cumulative count threshold, maintain the high-speed rail scene working state, wait for the next frame frequency offset estimate value input, and repeat the judgment process.
[0054] In this embodiment, as Figure 5 As shown, step S101 includes: Step S1011: Based on the frequency offset estimation value reported by the target terminal, sort the received first frequency offset estimation value, second frequency offset estimation value, and third frequency offset estimation value from smallest to largest, calculate the difference between two adjacent frequency offset estimation values, and obtain the first gap and the second gap. The first gap is the difference between the second largest value and the smallest value among the first, second, and third frequency offset estimation values, and the second gap is the difference between the largest value and the second largest value among the first, second, and third frequency offset estimation values. Step S1012: Compare the first gap and the second gap, and compare the first gap and the second gap with a preset second threshold; If the largest gap between the first gap and the second gap is less than or equal to the preset second threshold, that is, if both the first gap and the second gap are less than or equal to the second threshold, then all frequency offset estimates are classified into the first set, the second set is empty, and it is determined that the target terminal has not entered the single-frequency network mode. If the maximum gap between the first and second gaps is greater than the second threshold, then the frequency offset estimates on both sides of the maximum gap are classified into the first set and the second set, respectively. Furthermore, using this maximum gap as the segmentation boundary, all frequency offset samples to the left of the maximum gap are collected into the first set, and all frequency offset samples to the right of the maximum gap are collected into the second set.
[0055] Specifically, this embodiment determines the Single Frequency Network (SFN) mode based on the frequency offset estimation value reported by the terminal in real time and whether the difference between the frequency offset estimation values at adjacent time points is greater than a preset threshold. If the frequency offset value is greater than the first threshold and accumulates to a preset number of times, it is considered to be in SFN mode. In SFN mode, the Doppler frequency offset comes from multipath from different directions. This invention stores the frequency offset estimation values corresponding to different directions separately in a first set and a second set.
[0056] In this step, the initial clustering employs the maximum gap method: a fixed number of consecutive frequency offset estimates are sorted by size, the gap between adjacent values is calculated, and the data is segmented into a first set and a second set at the maximum gap. This method is based on the fact that downlink signals from base stations in different directions in a single-frequency network scenario exhibit significantly different Doppler frequency offsets, and the frequency offset estimates display a bimodal distribution. The maximum gap method is an unsupervised nonparametric clustering method that does not require prior parameters such as the number of clusters or cluster centers; it identifies the segmentation point solely through the gap between adjacent values after sorting the frequency offset samples.
[0057] Preferably, a fixed number of frames is preset to 3. From the frequency offset estimates periodically reported by the target terminal hardware, three consecutive frames of frequency offset estimation samples are selected and defined as the first frequency offset estimate f0, the second frequency offset estimate f1, and the third frequency offset estimate f2, respectively. The three sets of frequency offset estimation samples are sorted in ascending order according to their numerical values to obtain an ordered sample sequence s0≤s1≤s2. The numerical differences between adjacent frequency offset samples after sorting are calculated sequentially to generate two numerical gaps: the difference between the minimum value s0 and the second largest value s1 in the ordered sequence is defined as the first gap gap1, and the difference between the second largest value s1 and the maximum value s2 in the ordered sequence is defined as the second gap gap2. Then, the numerical values of the first gap and the second gap are compared, and the maximum gap in the two gaps is selected as max_gap=max(gap1,gap2). Next, the maximum gap max_gap is compared with a preset threshold, and the single-frequency network pattern recognition and frequency offset sample set collection are completed by combining the threshold determination result of the maximum gap.
[0058] Preferably, the preset thresholds include two, namely a second threshold TH2 and a third threshold TH3. The second threshold TH2 is smaller than the third threshold TH3. As an example and not a limitation, the second threshold TH2 is set to 200Hz and the third threshold TH3 is set to 1000Hz.
[0059] The second threshold TH2 is a critical threshold for set splitting / merging, serving as the core decision threshold for distinguishing single-source signals and dual-source signals. When the maximum gap max_gap is greater than the second threshold TH2, two sets are split with the maximum gap as the boundary, and the single frequency network mode is entered; when the maximum gap max_gap is less than or equal to the second threshold TH2, the three frequency offset estimation values are merged into a single set, and the single frequency network mode is not entered.
[0060] The third threshold TH3 is a high-confidence judgment threshold, used to characterize the confidence level of the clustering result. When the maximum gap max_gap is greater than the third threshold TH3, it is a high-confidence single frequency network mode, and the two sides of the maximum gap clearly come from different signal sources.
[0061] The present invention adopts two thresholds for coordinated judgment, which can completely cover all numerical distribution scenarios of the three-frame ordered frequency offset samples, and the specific scenarios are summarized as follows.
[0062] Taking the ascending ordered samples s0≤s1≤s2 as an example, corresponding to the first gap gap1=s1-s0 and the second gap gap2=s2-s1, 9 types of complete sample distribution scenarios are divided according to the magnitude relationship between the first gap gap1 and the second gap gap2, and the numerical interval relationship between the maximum gap and the two-level thresholds. The clustering rules and single frequency network mode judgment results corresponding to each type of scenario are as follows: In scenario 1, the second gap gap2 is greater than the first gap gap1, and the second gap gap2 is greater than the third threshold TH3. At this time, the maximum gap is the second gap gap2, that is, gap2>gap1 and gap2>TH3, max_gap=gap2, which is judged as a high-confidence single frequency network mode. With the second gap gap2 as the splitting boundary, the first set={s0,s1}, and the second set={s2}.
[0063] In scenario 2, the second gap gap2 is greater than the first gap gap1, and the second gap gap2 is greater than the second threshold TH2 and less than or equal to the third threshold TH3. At this time, the maximum gap is the second gap gap2, that is, gap2>gap1 and TH2<gap2≤TH3, max_gap=gap2, which is judged as a gray area single frequency network scenario. With the second gap gap2 as the splitting boundary, the first set={s0,s1}, and the second set={s2}.
[0064] In scenario 3, the second gap gap2 is greater than the first gap gap1, and the second gap gap2 is less than or equal to the second threshold TH2, that is, gap2>gap1 and gap2≤TH2, max_gap=gap2≤TH2, which is judged as a homologous single signal source scenario. All three groups of samples are uniformly classified into the first set, the first set={s0,s1,s2}, the second set is empty, and the terminal does not enter the single frequency network mode.
[0065] Case 4, the first gap gap1 is greater than the second gap gap2, and the first gap is greater than the third threshold TH3. In this case, the maximum gap is the first gap gap1, that is, gap1>gap2 and gap1>TH3, max_gap=gap1, which is determined as a high-confidence single frequency network scenario. Taking the first gap gap1 as the division boundary, the first set={s0}, and the second set={s1,s2}.
[0066] Case 5, the first gap gap1 is greater than the second gap gap2, and the first gap gap1 is greater than the second threshold TH2 and less than or equal to the third threshold TH3. In this case, the maximum gap is the first gap gap1, that is, gap1>gap2 and TH2<gap1≤TH3, max_gap=gap1, which is determined as a gray area single frequency network scenario. Taking the first gap gap1 as the division boundary, the first set={s0}, and the second set={s1,s2}.
[0067] Case 6, the first gap gap1 is greater than the second gap gap2, and the first gap gap1 is less than or equal to the second threshold TH2, that is, gap1>gap2 and gap1≤TH2, max_gap=gap1≤TH2, which is determined as a homologous single signal source scenario. All three groups of samples are uniformly classified into the first set, the first set={s0,s1,s2}, the second set is empty, and the terminal does not enter the single frequency network mode.
[0068] Case 7, the values of the first gap gap1 and the second gap gap2 are equal, and the gap value is greater than the third threshold TH3, that is, max_gap=gap1=gap2>TH3, which is determined as a high-confidence single frequency network scenario. Taking the first gap gap1 as the division boundary, the first set={s0}, and the second set={s1,s2}.
[0069] Case 8, the values of the first gap gap1 and the second gap gap2 are equal, and the gap value is greater than the second threshold TH2 and less than or equal to the third threshold TH3, that is, TH2<max_gap=gap1=gap2≤TH3, which is determined as a gray area single frequency network scenario; the first gap is fixedly taken as the division boundary, the first set={s0}, and the second set={s1,s2}.
[0070] Case 9, the values of the first gap gap1 and the second gap gap2 are equal, and the gap value is less than or equal to the second threshold TH2, that is, max_gap=gap1=gap2≤TH2, which is determined as a homologous single signal source scenario. All three groups of samples are uniformly classified into the first set, the first set={s0,s1,s2}, the second set is empty, and the terminal does not enter the single frequency network mode.
[0071] Based on the above nine sub-scenarios, the initial clustering can be simplified to a single rule: if the maximum gap (max_gap) > TH2, then the frequency-offset samples are separated at the maximum gap (max_gap) and enter single-frequency network mode; if the maximum gap (max_gap) ≤ TH2, then all frequency-offset samples are merged into a single set and the non-single-frequency network operation state is maintained. The third threshold (TH3) is only used to label the confidence level of the single-frequency network judgment results, distinguishing between high-confidence single-frequency network scenarios and gray-area single-frequency network scenarios, and does not affect the sample segmentation or merging decision.
[0072] In this embodiment, as Figure 6 As shown, after step S1012, the following steps are also included: S1013, in single-frequency network mode, when a new frequency offset estimate is received, the absolute value of the difference between the new frequency offset estimate and the previous frequency offset estimate is calculated. S1014, compare the absolute value of the difference between the new frequency offset estimate and the previous frequency offset estimate with the second threshold and the preset third threshold; If the absolute value of the difference between the new frequency offset estimate and the previous frequency offset estimate is greater than the third threshold, then the new frequency offset estimate is classified into a different set than the set in which the previous frequency offset estimate is located. If the absolute value of the difference between the new frequency offset estimate and the previous frequency offset estimate is less than or equal to the second threshold, then the new frequency offset estimate is classified into the same set as the previous frequency offset estimate. If the absolute value of the difference between the new frequency offset estimate and the previous frequency offset estimate is greater than the second threshold and less than or equal to the third threshold, then the new frequency offset estimate is discarded.
[0073] Specifically, considering that frequency offset estimates from the same signal source are continuous in time and adjacent values differ little, but when the relative motion state between the terminal and different signal sources changes abruptly, such as when a train passes through the coverage area of different Remote Radio Heads (RRHs), the frequency offset estimates will show significant jumps. Based on this, after the initial clustering, the temporal continuity of the Doppler frequency offset is used for value-by-value tracking and classification in subsequent classifications.
[0074] The new frequency offset estimate is compared with the previous frequency offset estimate. If the difference between the new frequency offset estimate and the previous frequency offset estimate exceeds the third threshold and the direction conforms to the cross-set jump characteristic, then it is assigned to another set opposite to the set containing the previous frequency offset estimate. For example, if the previous frequency offset estimate belongs to the first set, then the new frequency offset estimate is assigned to the second set; if the previous frequency offset estimate belongs to the second set, then the new frequency offset estimate is assigned to the first set.
[0075] If the difference between the new frequency offset estimate and the previous frequency offset estimate is less than the third threshold, it indicates that they both originate from the same signal source, and the new frequency offset estimate is assigned to the same set as the previous frequency offset estimate. For example, if the previous frequency offset estimate belongs to the first set, the new frequency offset estimate is also assigned to the first set; if the previous frequency offset estimate belongs to the second set, the new frequency offset estimate is also assigned to the second set.
[0076] For all other cases besides the two mentioned above, the new frequency offset estimate is considered an outlier and discarded.
[0077] In this embodiment, as Figure 7 As shown, it also includes: S1015, When the target terminal enters single-frequency network mode, monitor the last update time of the first set and the second set; S1016, if the time difference between the last update time of either the first set or the second set and the current time exceeds the preset maximum update interval threshold, determine that the target terminal exits the single-frequency network mode and clear the first set and the second set.
[0078] Specifically, in this embodiment, when the last update time of the first set or the second set exceeds the preset maximum update interval threshold, it indicates that a signal from a certain direction has not been detected for a long time and no longer meets the entry conditions for the single-frequency network mode. At this time, the single-frequency network mode is exited, and the temperature drift estimation related state and frequency offset buffer are reset. The exit mechanism of the single-frequency network mode scenario does not depend on the result of a single frequency offset estimation, but is a timeout exit mechanism based on the activity of the set, avoiding premature exit due to instantaneous signal obstruction.
[0079] In this embodiment, the first set and the second set respectively use a first-in-first-out buffer to store the frequency offset estimates after classification. The first-in-first-out buffer is configured such that when a new classified frequency offset estimate is received, if the current storage amount exceeds the preset storage limit, the oldest frequency offset estimate is removed and the storage amount count remains unchanged; if the current storage amount does not exceed the preset storage limit, the storage amount count is incremented. The first set and the second set are each configured with a sum accumulator. Each sum accumulator is connected to the first-in-first-out buffer of the corresponding set. The sum accumulator is configured to add the newly classified frequency offset estimate to the accumulation result and subtract the oldest frequency offset estimate removed from the first-in-first-out buffer.
[0080] Specifically, crystal oscillator temperature drift is an inherent deviation of the local frequency reference at both ends of the terminal. Its value changes slowly, unlike the Doppler frequency offset, which is generated by the high-speed relative motion between the terminal and the base station. It has the characteristics of drastic changes and polarity reversal with the direction of the propagating wave. The two are coupled and superimposed in the original frequency offset estimate. This invention achieves decoupling and separation of the two components through diversity buffering.
[0081] In a single-frequency network (SNNF) scenario, the first and second sets capture frequency offset estimates from incoming waves from two different directions. The mean frequency offsets of the two sets each contain the same temperature drift component and different Doppler components. Averaging the means of the two sets results in the Doppler components canceling each other out due to their opposite directions, yielding the temperature drift estimate. In a non-SNNF mode, only a single set exists: the first set stores valid frequency offset estimates, and the second set is empty. The incremental moving average of the first set is directly used as the temperature drift estimate.
[0082] To efficiently calculate the moving average in an embedded environment, the first and second sets each maintain a FIFO circular buffer and a sum accumulator. First, the newly classified frequency offset estimate is pushed into the corresponding set's FIFO circular buffer. Then, it is determined whether the number of samples currently stored in the buffer has reached the preset storage capacity limit. If the buffer is full, the oldest frequency offset estimate stored at the front of the buffer is removed, and the effective sample count in the buffer remains constant. If the buffer has not reached the preset storage capacity limit and there are no historical old sample removal operations, the equivalent removal value is set to zero, and the effective sample count in the buffer is incremented by 1. After the old sample removal process is completed, the new frequency offset estimate is written to the tail storage unit of the buffer, and the buffer's head and tail access pointers are updated synchronously, completing the data refresh of the circular buffer. The sum accumulator performs iterative calculations on the stored cumulative sum, with the calculation rule: Sum accumulator cumulative sum = original sum accumulator sum + new classification frequency offset estimate - value removed this time. By using the incremental moving average mechanism, the sum of all frequency offset samples in the set can be refreshed in real time with only two addition and subtraction operations. There is no need to traverse and read all historical frequency offset samples in the buffer to re-sum them, so that the moving average calculation maintains constant time complexity and is suitable for embedded terminal operating environments with limited hardware resources.
[0083] In this embodiment, as Figure 8 As shown, step S103 includes: Step S1031: Receive power delay distribution data reported by the target terminal. The power delay distribution data includes power information and delay information of several paths. Step S1032: Remove paths whose power information is lower than the preset noise threshold; Step S1033: Calculate the delay difference between the delay information of each remaining effective path after elimination, and classify the effective paths with delay differences less than the preset intra-cluster threshold into the same cluster. Step S1034: Clusters whose minimum inter-cluster delay difference is greater than a preset inter-cluster threshold and whose power information of the representative path of each cluster is higher than a preset power threshold are identified as resolvable clusters; wherein, the representative path is the path with the largest power information within the cluster. Step S1035: Determine the channel type based on the number of resolvable clusters; If there are two or more distinguishable clusters, the channel type is determined to be a multipath channel; If there is one or fewer resolvable clusters, the channel type is determined to be a single-path channel.
[0084] Specifically, in this embodiment, the hardware-reported Power Delay Profile (PDP) data includes power and delay information for several paths, used to describe the characteristic parameters of the power-delay relationship for each path in a multipath channel. In single-path channels, such as HST channels and HST-DPS channels, the PDP data exhibits a single, concentrated energy peak; in multipath channels, such as HST-SFN channels and traditional fading channels, the PDP data exhibits multiple, dispersed energy peaks. This invention performs cluster identification based on the joint use of two-dimensional information of delay and power from the PDP data, overcoming the limitation of relying solely on a single threshold for hard determination of delay spread.
[0085] The core basis for clustering is that the reflection or scattering components of the same physical propagation path have similar time delays, meaning the time delay difference is less than a preset intra-cluster threshold; while the time delay difference between different physical propagation paths is significant, meaning the time delay difference is greater than a preset inter-cluster threshold. Through hierarchical clustering along the time delay dimension, supplemented by the power dimension, it is ensured that each cluster has sufficient signal energy to form an effective propagation path, thus comprehensively determining the multipath characteristics of the channel.
[0086] Multipath identification includes the following clustering process: acquiring power delay distribution data reported by the hardware; filtering out weak multipath components with power information below a preset noise threshold, eliminating invalid paths dominated by background noise, and retaining the set of effective paths with acceptable signal-to-noise ratio and effective propagation energy; calculating the difference in delay information between each pair of all retained effective paths, grouping effective paths with delay differences less than a preset intra-cluster delay threshold into the same cluster, completing the initial clustering of scattering / reflection components along the same propagation path. Selecting the path with the largest power value from each divided cluster as the representative path of that cluster; traversing all clusters to calculate the minimum inter-cluster delay difference, marking clusters with a minimum inter-cluster delay difference greater than a preset inter-cluster delay threshold and whose corresponding representative path power is higher than a preset power threshold as resolvable clusters. Preliminary channel judgment is made based on the statistical number of resolvable clusters: if two or more resolvable clusters are identified, the channel is preliminarily determined to be a multipath channel; if the number of resolvable clusters is ≤1, the channel is preliminarily determined to be a single-path channel.
[0087] To avoid frequent ping-pong switching between scenarios, a hysteresis counting mechanism is used for both entering and exiting multipath channel mode. If the number of resolvable clusters is ≤1, or the minimum inter-cluster delay difference is less than or equal to a preset inter-cluster delay threshold, or the power of the representative path of each cluster is less than or equal to a preset power threshold, the third counter is cleared and single-path channel mode is maintained. If the number of resolvable clusters is ≥2, the minimum inter-cluster delay difference is greater than a preset inter-cluster delay threshold, and the power of the corresponding representative path of the cluster is higher than a preset power threshold, the third counter is incremented, and it is further determined whether the count value of the third counter is greater than or equal to a preset entry hysteresis count threshold. If the count value of the third counter is less than the preset entry hysteresis count threshold, the multipath identification process is repeated after waiting for the next power delay distribution input. If the count value of the third counter is greater than or equal to the preset entry hysteresis count threshold, the multipath channel mode is set.
[0088] In multipath channel mode, it is determined whether the number of clusters is less than 2 or whether the clusters are indistinguishable. If so, the fourth counter is incremented, and it is further determined whether the count value of the fourth counter is greater than or equal to the preset exit hysteresis number threshold. If the count value of the fourth counter is less than the preset exit hysteresis number threshold, the multipath identification process is repeated after waiting for the next power delay distribution input. If the count value of the fourth counter is greater than or equal to the preset exit hysteresis number threshold, it is set to single-path channel mode.
[0089] Table 1 Compensation strategy for automatic frequency control loop
[0090] Referring to Table 1, in step S104, if the channel type is determined to be a single-path channel, the compensation amount in the compensation strategy includes temperature drift component and Doppler component; wherein, the temperature drift component is the temperature drift estimate in step S102, and the Doppler component is the difference between the frequency offset estimate and the temperature drift estimate; if the channel type is determined to be a multipath channel, the compensation amount in the compensation strategy only includes temperature drift component.
[0091] The automatic frequency control loop adopts a hierarchical compensation architecture in the analog and digital domains. The analog domain voltage-controlled oscillator (VCO) provides a wide range and coarse-grained frequency adjustment, while the digital domain numerically controlled oscillator (NCO) provides a narrow range and fine-grained frequency adjustment. The two levels of compensation work together to meet the requirements of both wide dynamic range and high compensation accuracy in high-speed rail scenarios.
[0092] When the loop is working, the corresponding control registers are updated according to the currently determined analog domain compensation and digital domain compensation values respectively. The frequency deviation estimate at the input of the automatic frequency control loop remains unchanged and does not affect the subsequent frequency deviation tracking calculation.
[0093] Table 2 Frequency Adjustment Strategy for Transmitter Numerical Control Oscillator:
[0094] In non-high-speed rail scenarios, traditional automatic frequency control loops typically employ fixed-bandwidth loop filters. The frequency offset estimate is directly distributed to the voltage-controlled oscillator and digitally controlled oscillator after a single filter, without distinguishing between temperature drift and Doppler components. However, this approach has fundamental problems in high-speed rail scenarios: the large dynamic range of Doppler frequency offset exceeds the tracking capability of traditional loop filters; and the mixing of temperature drift and Doppler aliasing leads to incorrect selection of compensation strategies.
[0095] Based on this, in the high-speed rail scenario mode, referring to Tables 1 and 2, if it is a multipath channel, directly compensating for Doppler frequency offset may lead to overcompensation and deterioration of reception performance. This invention only compensates for temperature drift, and the corresponding transmit CNC oscillator frequency is set to zero. At this time, no Doppler compensation is performed, and the receiving frequency only corrects the temperature drift part. If it is a single-path channel, both temperature drift and Doppler are compensated. The transmit CNC oscillator frequency is jointly determined by the current analog domain compensation amount (VCO), digital domain compensation amount (NCO), and temperature drift estimate. The compensation direction of the Doppler part is opposite to the receiving Doppler direction to ensure that the uplink transmitted signal reaches the base station with the correct frequency after Doppler frequency shift.
[0096] This invention provides an automatic frequency control system for high-speed rail scenarios, used to implement the automatic frequency control method for high-speed rail scenarios described above. Figure 9 As shown, it includes: The single-frequency network bidirectional judgment module 1 is used to determine whether to enter the single-frequency network mode based on the frequency offset estimation value reported by the target terminal when the target terminal enters the high-speed rail scenario mode, and classify the frequency offset estimation value into the first set and the second set according to the direction of arrival. Temperature drift estimation module 2 is used to average the frequency offset estimates in the first set and the second set to obtain the temperature drift estimate, wherein the Doppler components cancel each other out because they are in opposite directions; Multipath identification module 3 is used to identify multipath based on the power delay distribution data reported by the target terminal and determine the channel type; Compensation module 4 is used to adaptively adjust the compensation strategy of the automatic frequency control loop according to the channel type and temperature drift estimate.
[0097] Furthermore, the system also includes a high-speed rail scene recognition module. This module measures the fluctuation of the frequency offset estimate based on the exponentially moving weighted standard deviation (EWMSTD), and combines a dual hysteresis mechanism of entry hysteresis and exit hysteresis to achieve robust recognition and exit from high-speed rail scenes. Compared with traditional AFC schemes, this invention monitors the frequency offset fluctuation in real time through exponentially moving weighted standard deviation, enabling it to autonomously identify high-speed rail scenes and switch AFC strategies, rather than relying on fixed-bandwidth loop filters, thus improving the frequency offset tracking accuracy and response speed in high-speed rail scenes.
[0098] The single-frequency network bidirectional judgment module 1 uses the maximum gap method for initial clustering and utilizes the time continuity of Doppler frequency offset for value-by-value tracking and classification. Combined with the timeout exit mechanism, it achieves effective separation of Doppler frequency offsets from different directions in the SFN scenario.
[0099] In single-frequency network mode, temperature drift estimation module 2 cancels the Doppler component and extracts the temperature drift component by averaging the frequency offset mean of the wave sets from two directions. In non-single-frequency network mode, it directly uses the mean of a single set as the temperature drift estimate and adopts incremental moving average to achieve efficient calculation with constant time complexity.
[0100] The multipath identification module 3 performs cluster identification based on the two-dimensional information of power delay distribution, time delay and power, and combines the entry / exit hysteresis mechanism to distinguish between single-path and multipath channels.
[0101] The compensation module 4 adaptively selects different AFC compensation strategies based on the multipath identification results. In single-path scenarios, it compensates for temperature drift and Doppler frequency shift, while in multipath scenarios, it only compensates for temperature drift. It also adopts a hierarchical compensation architecture that combines analog domain voltage-controlled oscillators and digital domain numerically controlled oscillators to balance large dynamic range and compensation accuracy.
[0102] Compared to traditional schemes based on high-level signaling handover or direct compensation, this invention adds bidirectional judgment of single-frequency network, temperature drift separation, multipath identification, and adaptive compensation strategy selection, forming a complete automatic frequency control processing link for high-speed rail scenarios. It can adaptively select the optimal compensation strategy under different channel environments, solving the problem that a single scheme cannot simultaneously take into account various high-speed rail channels.
[0103] The present invention provides a terminal device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the automatic frequency control method in the high-speed rail scenario described above.
[0104] Overall, this invention achieves a complete automatic frequency control scheme for high-speed rail scenarios through pure software processing, without the need for hardware modifications. This allows devices that only support non-high-speed rail scenarios to support high-speed rail scenarios through software upgrades, resulting in significant cost advantages and market competitiveness.
[0105] The above are merely preferred embodiments of the present invention and are not intended to limit the implementation methods and protection scope of the present invention. Those skilled in the art should recognize that any equivalent substitutions and obvious changes made using the content of this specification and illustrations should be included within the protection scope of the present invention.
Claims
1. An automatic frequency control method for high-speed rail scenarios, characterized in that, include: Step S101: When the target terminal enters the high-speed rail scenario mode, determine whether to enter the single-frequency network mode based on the frequency offset estimation value reported by the target terminal, and classify the frequency offset estimation value into the first set and the second set according to the direction of arrival. Step S102: Average the frequency offset estimates in the first set and the second set to obtain the temperature drift estimate, wherein the Doppler components cancel each other out because they are in opposite directions; Step S103: Perform multipath identification based on the power delay distribution data reported by the target terminal to determine the channel type; Step S104: Adaptively adjust the compensation strategy of the automatic frequency control loop according to the channel type and the temperature drift estimate.
2. The automatic frequency control method for high-speed rail scenarios according to claim 1, characterized in that, The procedure preceding step S101 also includes: Step S100: Determine whether the target terminal has entered the high-speed rail scenario mode; When it is determined that the target terminal has entered the high-speed rail scene mode, proceed to step S101; When it is determined that the target terminal has exited the high-speed rail scene mode, return to step S100.
3. The automatic frequency control method for high-speed rail scenarios according to claim 2, characterized in that, Determining whether the target terminal has entered the high-speed rail scenario mode includes: Step S1001A: Receive the first signaling, which is used to indicate whether the target terminal is in high-speed rail scenario mode; Step S1002A: Based on the first signaling, determine whether the target terminal is in high-speed rail scenario mode: When the first signaling instructs the target terminal to enter the high-speed rail scenario mode, proceed to step S101; When the first signaling instructs the target terminal to exit the high-speed rail scenario mode, proceed to step S100.
4. The automatic frequency control method for high-speed rail scenarios according to claim 2, characterized in that, Determining whether the target terminal has entered the high-speed rail scenario mode includes: Step S1001B: Receive the frequency offset estimate reported by the target terminal; Step S1002B: Calculate the standard deviation of the frequency offset estimate; Step S1003B: Determine whether the standard deviation of the frequency offset estimate meets the preset entry condition, wherein the preset entry condition is that the standard deviation of the frequency offset estimate is greater than the preset first threshold and the consecutive cumulative number is not less than the preset first cumulative number threshold. If the standard deviation of the frequency offset estimate meets the preset entry condition, then the target terminal is determined to enter the high-speed rail scenario mode and proceeds to step S101. If the standard deviation of the frequency offset estimate does not meet the preset entry condition, then return to step S1001B; After confirming that the target terminal has entered the high-speed rail scenario mode, the following steps are also included: Step S1004B: Determine whether the standard deviation of the frequency offset estimate meets the preset exit condition, wherein the preset exit condition is that the standard deviation of the frequency offset estimate is not greater than the preset first threshold and the cumulative number is not less than the preset second cumulative number threshold in the high-speed rail scenario mode. If the standard deviation of the frequency offset estimate meets the preset exit condition, then the target terminal is determined to exit the high-speed rail scenario mode and proceeds to step S100. If the standard deviation of the frequency offset estimate does not meet the preset exit condition, then return to step S1001B.
5. The automatic frequency control method for high-speed rail scenarios according to claim 1, characterized in that, Step S101 includes: Step S1011: Based on the frequency offset estimation value reported by the target terminal, sort the received first frequency offset estimation value, second frequency offset estimation value, and third frequency offset estimation value from smallest to largest, calculate the difference between two adjacent frequency offset estimation values, and obtain the first gap and the second gap. The first gap is the difference between the second largest value and the smallest value among the first frequency offset estimation value, second frequency offset estimation value, and third frequency offset estimation value, and the second gap is the difference between the largest value and the second largest value among the first frequency offset estimation value, second frequency offset estimation value, and third frequency offset estimation value. Step S1012: Compare the first gap and the second gap, and compare the first gap and the second gap with a preset second threshold; If both the first gap and the second gap are less than or equal to the second threshold, then all frequency offset estimates are classified into the first set, the second set is empty, and it is determined that the target terminal has not entered the single-frequency network mode. If the maximum gap between the first gap and the second gap is greater than the second threshold, then the frequency offset estimates on both sides of the maximum gap are classified into the first set and the second set, respectively.
6. The automatic frequency control method for high-speed rail scenarios according to claim 5, characterized in that, Following step S1012, the following is also included: In the single-frequency network mode, when a new frequency offset estimate is received, the absolute value of the difference between the new frequency offset estimate and the previous frequency offset estimate is compared with the second threshold and the preset third threshold. If the absolute value of the difference between the new frequency offset estimate and the previous frequency offset estimate is greater than the third threshold, then the new frequency offset estimate is classified into a different set than the set in which the previous frequency offset estimate is located. If the absolute value of the difference between the new frequency offset estimate and the previous frequency offset estimate is less than or equal to the second threshold, then the new frequency offset estimate is classified into the same set as the previous frequency offset estimate. If the absolute value of the difference between the new frequency offset estimate and the previous frequency offset estimate is greater than the second threshold and less than or equal to the third threshold, then the new frequency offset estimate is discarded.
7. The automatic frequency control method for high-speed rail scenarios according to claim 5, characterized in that, Also includes: When the target terminal enters the single-frequency network mode, monitor the last update time of the first set and the second set; If the time difference between the last update time of either the first set or the second set and the current time exceeds the preset maximum update interval threshold, the target terminal is determined to exit the single-frequency network mode, and the first set and the second set are cleared.
8. The automatic frequency control method for high-speed rail scenarios according to claim 1, characterized in that, The first set and the second set respectively use a first-in-first-out buffer to store the frequency offset estimates after classification. The first-in-first-out buffer is configured to remove the oldest frequency offset estimate and keep the storage count unchanged if the current storage amount exceeds the preset storage limit when a new frequency offset estimate is received, and if the current storage amount does not exceed the preset storage limit, the storage count is incremented. The first set and the second set are respectively configured with sum accumulators. Each sum accumulator is connected to the first-in-first-out buffer of the corresponding set. The sum accumulator is configured to add the newly classified frequency offset estimate to the accumulation result and subtract the oldest frequency offset estimate removed by the first-in-first-out buffer.
9. The automatic frequency control method for high-speed rail scenarios according to claim 1, characterized in that, Step S103 includes: Step S1031: Receive power delay distribution data reported by the target terminal, wherein the power delay distribution data includes power information and delay information of several paths; Step S1032: Remove paths whose power information is lower than the preset noise threshold; Step S1033: Calculate the delay difference between the delay information of each remaining effective path after elimination, and classify the effective paths with delay differences less than the preset intra-cluster threshold into the same cluster. Step S1034: Clusters whose minimum inter-cluster delay difference is greater than a preset inter-cluster threshold and whose power information of the representative path of each cluster is higher than a preset power threshold are determined as distinguishable clusters; wherein, the representative path is the path with the largest power information within the cluster. Step S1035: Determine the channel type based on the number of resolvable clusters; If there are two or more distinguishable clusters, the channel type is determined to be a multipath channel; If one or more of the resolvable clusters described above exist, the channel type is determined to be a single-path channel.
10. The automatic frequency control method for high-speed rail scenarios according to claim 1, characterized in that, In step S104, if the channel type is determined to be a single-path channel, the compensation amount in the compensation strategy includes a temperature drift component and a Doppler component; wherein, the temperature drift component is the temperature drift estimate in step S102, and the Doppler component is the difference between the frequency offset estimate and the temperature drift estimate. If the channel type is determined to be a multipath channel, then the compensation amount in the compensation strategy only includes the temperature drift component.
11. An automatic frequency control system for high-speed rail scenarios, characterized in that, The method for implementing automatic frequency control in a high-speed rail scenario as described in any one of claims 1 to 10 includes: The single-frequency network bidirectional judgment module is used to determine whether to enter the single-frequency network mode based on the frequency offset estimation value reported by the target terminal when the target terminal enters the high-speed rail scenario mode, and classify the frequency offset estimation value into the first set and the second set according to the direction of arrival. The temperature drift estimation module is used to average the frequency offset estimates in the first set and the second set to obtain the temperature drift estimate, wherein the Doppler components cancel each other out because they are in opposite directions. The multipath identification module is used to perform multipath identification based on the power delay distribution data reported by the target terminal and determine the channel type. The compensation module is used to adaptively adjust the compensation strategy of the automatic frequency control loop according to the channel type and the temperature drift estimate.
