A vehicle position tracking method, system and electronic device for a highway

Abstract

The application provides a highway vehicle position tracking method, system and electronic equipment, and belongs to the field of highway vehicle detection. The method comprises the following steps: preprocessing road surface vibration signals obtained from multiple measurement areas to obtain a characteristic sequence reflecting local vibration intensity; obtaining multiple candidate target intervals according to the characteristic sequence; the candidate target interval is an interval with a continuous amplitude exceeding a dynamic threshold; determining sample points with the largest amplitude in each candidate target interval as initial positioning points, respectively; associating the initial positioning points belonging to the same target vehicle as candidate trajectory segments according to the spatial distribution relationship of the multiple measurement areas; calculating Poisson distribution parameters based on the candidate trajectory segments, and screening effective trajectory segments according to the Poisson distribution parameters; and connecting the effective trajectory segments in chronological order to form a driving trajectory of the target vehicle within the coverage range of the multiple measurement areas. The application improves the accuracy of highway vehicle position detection.

Description

Technical Field

[0001] This invention relates to the field of highway vehicle detection, and specifically to a method, system, and electronic device for tracking the location of vehicles on highways. Background Technology

[0002] With the advancement of smart highway construction, traffic flow monitoring technology based on DAS (Distributed Acoustic Sensing) is receiving increasing attention. This technology uses distributed vibration sensors composed of UWFBG (Ultra-Weak Fiber Bragg Grating) arrays, embedded under the highway surface, to detect signals generated by passing vehicles, thereby achieving vehicle detection and location. However, the actual highway operating environment is complex and variable. Road surface vibration signals are often affected by environmental noise (wind noise, vibration from surrounding construction, etc.) and the temperature drift and baseline drift of the fiber optic sensing system itself, resulting in significant low-frequency drift and DC components. Furthermore, the raw sampled data typically has a high sampling rate and large data volume, accompanied by anomalies such as pulse spikes. Without effective processing, this can easily lead to false alarms, missed alarms, and excessive real-time computational burden.

[0003] To suppress the aforementioned interference and achieve vehicle position detection, it is generally necessary to preprocess the original road vibration signal (using methods such as bandpass filtering and matched filtering to suppress spike noise and slow drift, and retain signal components within the effective frequency band of vehicle vibration), extract features (extracting feature quantities that can characterize vehicle events from the preprocessed signal, such as energy features and root mean square), detect (obtaining candidate positioning points corresponding to vehicle events based on fixed thresholds), and associate (combining vehicle motion continuity and spatiotemporal constraints to associate positioning points to form trajectory output).

[0004] However, the above-mentioned technologies still have the following shortcomings: First, matched filtering relies on a single template, which is difficult to cover all target signal patterns. The vibration waveforms of vehicles under different vehicle models and speeds are significantly different, and matched filtering is prone to matching deviation, missed detection, or false detection. Second, the decision method based on fixed thresholds has poor adaptability to environmental changes. There are differences in background noise levels and temperature drift between different test areas. The candidate localization points obtained by fixed thresholds and simple peak detection often include many false localization points caused by noise or background disturbances. Third, in the multi-lane scenario of highways, the vibrations caused by heavy-load vehicles in adjacent lanes can be coupled to the current lane through the road structure, forming accompanying interference that is similar to the vibration waveform of real vehicles and even exceeds the intensity of light vehicle signals. Summary of the Invention

[0005] In view of this, it is necessary to provide a vehicle position tracking method, system and electronic device for highways to solve the technical problem of low vehicle position accuracy caused by the accompanying trajectory interference generated by the vibration coupling of adjacent lanes in the prior art.

[0006] To address the aforementioned technical problems, in a first aspect, the present invention provides a vehicle location tracking method for highways, comprising: Preprocessing of road vibration signals obtained from multiple measurement areas yields a characteristic sequence reflecting local vibration intensity; Multiple candidate target intervals are obtained based on the feature sequence; the candidate target intervals are intervals where the continuous amplitude exceeds the dynamic threshold. The sample point with the largest amplitude in each of the candidate target intervals is determined as the initial positioning point; Based on the spatial distribution relationship of the multiple measurement areas, the initial positioning points belonging to the same target vehicle are associated as candidate trajectory segments; Calculate the Poisson distribution parameters based on the candidate trajectory segments, and then filter the valid trajectory segments according to the Poisson distribution parameters; The effective trajectory segments are connected in chronological order to form the driving trajectory of the target vehicle within the coverage area of ​​multiple measurement zones.

[0007] In one possible implementation, the preprocessing of road vibration signals acquired from multiple measurement areas to obtain a feature sequence reflecting local vibration intensity includes: Acquire the road vibration signal collected by the distributed fiber optic vibration sensing system; The road vibration signal is subjected to median filtering to obtain a first intermediate signal sequence; The signal sequence obtained after median filtering is downsampled to obtain the second intermediate signal sequence; The local vibration intensity of the second intermediate signal sequence is extracted to obtain the feature sequence.

[0008] In one possible implementation, the step of calculating the moving variance of the second intermediate signal sequence and extracting the local vibration intensity features of the second intermediate signal sequence to obtain a feature sequence includes: A sliding window is constructed based on the second intermediate signal sequence. The sliding window includes the currently processed sample point in the second intermediate signal sequence and each of its N / 2 neighboring sample points on both sides, where N is a preset positive integer. Calculate the arithmetic mean of all the neighboring sample points within the sliding window; Based on the amplitude of each neighboring sample point within the sliding window and the arithmetic mean, the moving variance feature value corresponding to the current processing sample point is calculated. The moving variance feature value is used to characterize the local vibration intensity characteristics at the current processing sample point. Traverse all currently processed sample points in the second intermediate signal sequence, and arrange the moving variance feature values ​​corresponding to each currently processed sample point in order to obtain the feature sequence.

[0009] In one possible implementation, the step of obtaining multiple candidate target intervals based on the feature sequence; the candidate target intervals are intervals where the continuous amplitude exceeds a dynamic threshold, including: Taking the current detection sample point in the feature sequence as the center, Nc / 2 reference sample points are selected on both sides of the current detection sample point in the feature sequence to form a reference window, where Nc is a preset positive integer; Sort all the reference sample points within the reference window according to their amplitude. The background noise power estimate is obtained based on the magnitude of the kth reference sample point after sorting; k is a preset positive integer and k≤Nc / 2; Based on the background noise power estimate and the threshold factor, the dynamic threshold corresponding to the current detection sample point is obtained; the threshold factor is determined according to the preset system false alarm rate. The amplitude of the current detection sample point is compared with the dynamic threshold. If the amplitude of the current detection sample point is greater than the dynamic threshold, the current detection sample point is determined as a candidate point. The sample segment consisting of multiple consecutive candidate points in the feature sequence is taken as the candidate target interval.

[0010] In one possible implementation, associating the initial positioning points belonging to the same target vehicle as candidate trajectory segments based on the spatial distribution relationship of the multiple measurement areas includes: Based on the spatial distribution relationship of the multiple measurement areas, the initial positioning points are associated and matched; The initial positioning points that meet the preset spatiotemporal constraints are assigned to the same candidate trajectory segment; the preset spatiotemporal constraints are that the time difference between the target vehicle and adjacent test areas is within the theoretical passage time interval.

[0011] In one possible implementation, the calculation of the Poisson distribution parameters based on the candidate trajectory segment includes: The number of initial positioning points within a preset statistical time window for the candidate trajectory segment is counted. The Poisson distribution parameters of the candidate trajectory segments are obtained by using the maximum likelihood estimation algorithm based on the number of parameters.

[0012] In one possible implementation, the step of filtering valid trajectory segments based on the Poisson distribution parameters includes: The Poisson distribution parameters of each candidate trajectory segment are compared with a preset threshold. If the Poisson distribution parameter of the candidate trajectory segment is greater than or equal to the preset threshold, the candidate trajectory segment is determined to be the valid trajectory segment.

[0013] One possible implementation also includes: If the Poisson distribution parameter of the candidate trajectory segment is less than the preset threshold, the candidate trajectory segment is determined to be an interference trajectory segment and the interference trajectory segment is deleted.

[0014] Secondly, the present invention also provides a vehicle location tracking system for highways, comprising: The processing module is used to preprocess the road vibration signals obtained from multiple measurement areas to obtain a feature sequence reflecting the local vibration intensity. The detection module is used to obtain multiple candidate target intervals based on the feature sequence; the candidate target intervals are intervals where the continuous amplitude exceeds a dynamic threshold. The extraction module is used to determine the sample point with the largest amplitude in each of the candidate target intervals as the initial positioning point; The association module is used to associate the initial positioning points belonging to the same target vehicle as candidate trajectory segments based on the spatial distribution relationship of the multiple measurement areas; The filtering module is used to calculate the Poisson distribution parameters based on the candidate trajectory segments, and to filter valid trajectory segments according to the Poisson distribution parameters. The generation module is used to connect the effective trajectory segments in chronological order to form the driving trajectory of the target vehicle within the coverage area of ​​multiple measurement zones.

[0015] Thirdly, the present invention also provides an electronic device, including a memory and a processor, wherein, The memory is used to store programs; The processor, coupled to the memory, is used to execute the program stored in the memory to implement the steps in the vehicle location tracking method for highways described in any of the above implementations.

[0016] The beneficial effects of this invention are as follows: First, the invention preprocesses the road vibration signal, uses a dynamic threshold to compare the amplitude of the feature sequence and divide it into candidate target intervals, which can more flexibly adapt to the differences in signal amplitude between different test areas and different time periods, improving the stability and reliability of the detection results. Furthermore, within each candidate target interval, the sample point with the largest amplitude is selected as the initial positioning point, improving the temporal accuracy of the positioning point and reducing positioning deviations caused by signal waveform fluctuations. Furthermore, by utilizing the spatial distribution relationship between multiple test areas and the vehicle motion continuity constraint, the initial positioning points of different test areas are associated and matched, effectively clustering positioning points belonging to the same target vehicle into candidate trajectory segments, enhancing the coherence and consistency of trajectory segments, and reducing the probability of incorrect positioning point association. Furthermore, by leveraging the spatial distribution relationships among multiple measurement areas to jointly match dispersed initial positioning points, initial positioning points belonging to the same vehicle event can be effectively clustered to obtain candidate trajectory segments, reducing the probability of incorrect positioning point association. Statistical characteristic analysis and calculation of corresponding distribution parameters on the associated candidate trajectories can effectively identify and eliminate false segments caused by noise or adjacent lane interference from a large number of candidate segments, significantly improving the purity and reliability of the output trajectory. Finally, the effective trajectory segments are sequentially connected to ensure the integrity and continuity of the output trajectory, thereby improving the accuracy and reliability of vehicle driving positioning and tracking under complex conditions such as multi-lane and multi-vehicle types. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 A schematic flowchart of an embodiment of the vehicle location tracking method for highways provided by the present invention; Figure 2 For the present invention Figure 1 A schematic diagram of an embodiment of S101; Figure 3 A basic structural block diagram of the distributed acoustic wave sensing system based on UWFBG array used in the vehicle position tracking method for highways provided by the present invention. Figure 4 A schematic diagram of the deployment scheme of the distributed acoustic wave sensing system used in the vehicle location tracking method for highways provided by the present invention; Figure 5 A schematic diagram of the vibration intensity distribution of highway pavement within a certain time period in the vehicle location tracking method for highways provided by the present invention; Figure 6 A schematic diagram showing the relationship between the vibration intensity of the highway pavement and time in a certain measurement area of ​​a two-channel highway within a certain time period in the vehicle location tracking method for highways provided by the present invention. Figure 7 For the present invention Figure 2 A schematic diagram of an embodiment of S204; Figure 8 For the present invention Figure 1 A schematic diagram of an embodiment of S102; Figure 9 The OS-CFAR target detection flowchart in the vehicle position tracking method for highways provided by this invention; Figure 10 A schematic diagram of the movement variance and positioning results of a certain measurement area in a two-channel system during a certain time period in the vehicle location tracking method for highways provided by the present invention. Figure 11 For the present invention Figure 1 A schematic diagram of an embodiment of S104; Figure 12 This is a schematic diagram of the initial positioning results of vehicles on the highway within a certain time period in the vehicle location tracking method for highways provided by the present invention. Figure 13 This is a schematic diagram of an embodiment of the present invention for calculating Poisson distribution parameters based on candidate trajectory segments; Figure 14 A schematic diagram of the spatiotemporal correlation results of the initial positioning points of vehicles on the highway surface within a certain time period in the vehicle location tracking method for highways provided by the present invention; Figure 15 This is a schematic diagram of an embodiment of the present invention for selecting effective trajectory segments based on Poisson distribution parameters; Figure 16 A schematic diagram illustrating the Poisson distribution statistical distribution of the accompanying trajectory and the actual trajectory, and the determination of the optimal threshold in the vehicle position tracking method for highways provided by this invention. Figure 17 A schematic diagram of the vehicle position detection results on the highway surface within a certain time period in the vehicle position tracking method for highways provided by the present invention; Figure 18 A schematic diagram of an embodiment of the vehicle location tracking system for highways provided by the present invention; Figure 19 A schematic diagram of an embodiment of the electronic device provided by the present invention. Detailed Implementation

[0019] 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 a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0020] In the description of the embodiments of the present invention, unless otherwise stated, "multiple" means two or more. "And / or" describes the relationship between related objects, indicating that there can be three relationships. For example, A and / or B can represent three situations: A exists alone, A and B exist simultaneously, and B exists alone.

[0021] The terms "first," "second," etc., used in the embodiments of this invention are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, a technical feature defined with "first" or "second" may explicitly or implicitly include at least one of that feature.

[0022] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0023] This invention provides a method, system, and electronic device for tracking vehicle location on highways, which will be described below.

[0024] Figure 1 A schematic flowchart of an embodiment of the vehicle location tracking method for highways provided by the present invention is shown below. Figure 1 As shown, the vehicle location tracking method for highways includes: S101. Preprocess the road vibration signals obtained from multiple measurement areas to obtain a characteristic sequence reflecting the local vibration intensity.

[0025] It should be noted that the test area refers to the sensing region buried under the highway surface that can detect road vibration signals. The raw road vibration signals collected from each test area are preprocessed to transform the raw amplitude information into a feature sequence that can sensitively reflect the intensity of vibrations caused by vehicles passing over the road surface at different times. The feature sequence is a set of values ​​that can quantitatively characterize the intensity of local vibrations on the road surface at each moment.

[0026] S102. A plurality of candidate target intervals are obtained based on the feature sequence; the candidate target intervals are intervals in which the continuous amplitude exceeds the dynamic threshold.

[0027] It should be noted that: using the feature sequence as the processing object, the amplitude of each sample point in the feature sequence is compared with the corresponding dynamic threshold one by one, and adjacent sample points with amplitudes continuously higher than the dynamic threshold are grouped into candidate target intervals. The candidate target interval is a continuous time period in the feature sequence where vehicle travel events may occur.

[0028] S103. The sample point with the largest amplitude in each of the candidate target intervals is determined as the initial positioning point.

[0029] It should be noted that since the candidate target interval only indicates the time period during which a vehicle may pass through, it cannot pinpoint the exact time a vehicle passes through the test area. Therefore, an extreme value search operation is performed within each candidate target interval to select the peak point with the largest amplitude within that interval as the initial positioning point. Selecting the peak point with the largest amplitude within the interval through peak search ensures that the extracted initial positioning point has the highest signal-to-noise ratio, thereby improving the reliability of the positioning.

[0030] S104. Based on the spatial distribution relationship of the multiple measurement areas, the initial positioning points belonging to the same target vehicle are associated as candidate trajectory segments.

[0031] It should be noted that: by utilizing the spatial distribution relationship of each test area along the road direction, the initial positioning points of different test areas are associated, thereby connecting the initial positioning points belonging to the same target vehicle in the order of the test areas to form several candidate trajectory segments.

[0032] S105. Calculate the Poisson distribution parameters based on the candidate trajectory segments, and filter the valid trajectory segments according to the Poisson distribution parameters.

[0033] It should be noted that: Statistical characteristic analysis is performed on each candidate event segment to calculate the Poisson distribution parameter of each candidate trajectory segment. The Poisson distribution parameter is used to characterize the density of the distribution of positioning points within the trajectory segment. The Poisson distribution parameter of each candidate trajectory segment is compared with a preset discrimination threshold. Candidate trajectory segments with a Poisson distribution parameter greater than or equal to the preset threshold are retained as valid trajectory segments, while candidate trajectory segments with a Poisson distribution parameter less than the preset threshold are discarded.

[0034] S106. Connect the effective trajectory segments in chronological order to form the driving trajectory of the target vehicle within the coverage area of ​​the multi-measurement area.

[0035] It should be noted that the selected effective trajectory segments are connected sequentially in chronological order to generate the complete driving trajectory of the target vehicle within the coverage area of ​​multiple test areas, thereby achieving accurate reconstruction from road vibration response to vehicle driving path.

[0036] In summary, the vehicle positioning and tracking method for highways provided by this invention first preprocesses the road vibration signal, then uses a dynamic threshold to compare the amplitude of the feature sequence and divide it into candidate target intervals. This allows for more flexible adaptation to signal amplitude differences between different measurement areas and time periods, improving the stability and reliability of the detection results. Furthermore, within each candidate target interval, the sample point with the largest amplitude is selected as the initial positioning point, improving the temporal accuracy of the positioning point and reducing positioning deviations caused by signal waveform fluctuations. Moreover, by utilizing the spatial distribution relationship between multiple measurement areas and the vehicle motion continuity constraint, the initial positioning points in different measurement areas are associated and matched. This effectively clusters positioning points belonging to the same target vehicle into candidate trajectory segments, enhancing the coherence and consistency of trajectory segments and reducing the probability of incorrect positioning point association. Furthermore, by leveraging the spatial distribution relationships among multiple measurement areas to jointly match dispersed initial positioning points, initial positioning points belonging to the same vehicle event can be effectively clustered to obtain candidate trajectory segments, reducing the probability of incorrect positioning point association. Statistical characteristic analysis and calculation of corresponding distribution parameters on the associated candidate trajectories can effectively identify and eliminate false segments caused by noise or adjacent lane interference from a large number of candidate segments, significantly improving the purity and reliability of the output trajectory. Finally, the effective trajectory segments are sequentially connected to ensure the integrity and continuity of the output trajectory, thereby improving the accuracy and reliability of vehicle driving positioning and tracking under complex conditions such as multi-lane and multi-vehicle types.

[0037] In some embodiments of the present invention, such as Figure 2 As shown, step S101 includes: S201. Obtain the road vibration signal collected by the distributed optical fiber vibration sensing system.

[0038] It should be noted that: Optical cables were buried under the surface of a highway to obtain road vibration signals. In practical applications, such as... Figure 3As shown, a DAS (Distributed Acoustic Sensing) system is used to sense road vibration signals generated when a vehicle passes by via an UWFBG (Ultra-Weak Fiber Bragg Grating) array. The DAS system includes an unbalanced Michelson interferometer and an ultra-weak fiber grating sensing array. The unbalanced Michelson interferometer consists of two Faraday mirrors, a 3x3 coupler, and a delay fiber of length L. The UWFBG sensing array consists of multiple fiber gratings spaced at intervals of L. Continuous light emitted from a narrow-linewidth laser is modulated into a sequence of optical pulses by an EOM (Electro-Optic Modulator), amplified by an EDFA (Erbium-Doped Fiber Amplifier), and then injected into the UWFBG sensing array. The pulsed light reflected from the UWFBG sensor array enters an unbalanced Michelson interferometer through two circulators. Within the unbalanced Michelson interferometer, a time-delay fiber compensates for the optical path difference between adjacent ultra-weak fiber gratings, and the amplitude of the time-domain road vibration signal is recovered by demodulating the phase change caused by the optical pulse variation between adjacent ultra-weak fiber gratings in the demodulator. Finally, three optical signals from a 3×3 coupler are acquired by three photodetectors, and after photoelectric conversion, a processable raw electrical signal is obtained as the road vibration signal.

[0039] The deployment of DAS systems on highways, such as Figure 4 As shown, the optical fiber cables are uniformly arranged along the superstructure beneath each lane of the highway (each lane corresponds to an independent optical fiber channel, or different sensing segments within the same optical fiber). Each lane is independently monitored by its corresponding optical fiber channel, and multiple optical fiber monitoring areas are organized in a specific arrangement to form an optical fiber monitoring zone. Through this deployment scheme, the sensing channels corresponding to each lane can respectively sense the road surface vibration signals generated when vehicles pass through that lane. The vibration intensity distribution within the optical fiber monitoring zone obtained over a certain time period is shown below. Figure 5 As shown, the trajectories of multiple vehicles are represented by bright white line segments. In contrast, the darker lines are a side effect caused by interference from adjacent lanes during the data acquisition phase; these lines are considered interference with the current lane. Figure 5 The darker line segments #1 and #2 can be considered as heavy-duty vehicles on channel one, while #3 and #4 represent the corresponding accompanying trajectories on channel two, with an intensity greater than that of trajectory #5 formed by a small vehicle traveling on channel two. This demonstrates that the intensity of the accompanying interference may even be greater than the actual signal from a light vehicle, causing significant interference to vehicle position detection. Vehicle positioning within the sensing area is achieved based on data collected from a single measurement area. Figure 5Taking a certain side area of ​​the Second Central Passage as an example, the relationship between the road surface vibration intensity collected during this time period and the time period is as follows: Figure 6 As shown, when a vehicle passes through the test area, there is a large amplitude, but the intensity of accompanying interference (such as interference caused by vibration coupling from heavy-load vehicles in adjacent lanes) may be greater than the signal strength generated by the vehicle, and its waveform is similar to the real signal, making it impossible to distinguish directly. Therefore, it is necessary to extract signal features and convert the original amplitude signal into a feature representation that is more sensitive to local energy fluctuations, thereby improving the signal-to-noise ratio of the target signal generated by vehicle vibration. The road vibration signal collected by the system is sampled at a frequency of 1000 Hz. Numerous engineering experiments have shown that road vibration signals are concentrated in the low-frequency band. To improve the real-time performance of signal processing and reduce subsequent computational overhead, the original road vibration signal needs to be preprocessed according to the embodiments described below.

[0040] S202. The road vibration signal is subjected to median filtering to obtain the first intermediate signal sequence.

[0041] It should be noted that: Centered on the current sampling point, a sampling window of total length 2r+1 is selected, including the current sampling point and the r sampling points before and after it. All sampling points within the sampling window are sorted by amplitude, and the median of the sorted amplitudes is taken as the filtered amplitude of the current sampling point. This operation is the median filtering operation. The expression for the first intermediate signal sequence is shown below:

[0042] in, The first intermediate signal sequence after median filtering is the first... Let x[n] be the amplitude of the nth sampling point in the road vibration signal sequence. The amplitude of each sampling point This is the index number of the original sampling sequence. The operator for finding the median of a set. This is the radius of one side of the median filter window.

[0043] S203. The signal sequence obtained after median filtering is downsampled to obtain the second intermediate signal sequence.

[0044] It should be noted that: according to the preset downsampling factor From the signal sequence obtained after median filtering (the first intermediate signal sequence), every... One sampling point is extracted from each sampling point to form the downsampled signal sequence (the second intermediate signal sequence). The expression for the second intermediate signal sequence is as follows:

[0045] in, The second intermediate signal sequence after downsampling The amplitude of each sampling point The downsampling factor represents the percentage of samples taken. One sample is selected from each original sample as the output.

[0046] S204. Extract the local vibration intensity of the second intermediate signal sequence to obtain the feature sequence.

[0047] It should be noted that after median filtering and downsampling, the local vibration intensity feature extraction operation is performed on the obtained second intermediate signal sequence to obtain a feature sequence with high contrast.

[0048] In this embodiment, road vibration signals are acquired through a DAS system. Median filtering is performed on the road vibration signals to obtain a first intermediate signal sequence, which effectively suppresses abnormal noise such as pulse spikes mixed in the road vibration signals and reduces the interference of amplitude abrupt changes caused by non-vehicle factors on subsequent feature extraction, thereby improving the purity of the input data. Furthermore, downsampling is performed on the first intermediate signal sequence to obtain a second intermediate signal sequence, which significantly compresses the data size while retaining the main low-frequency features of vehicle vibration, reducing the computational overhead and storage burden of subsequent steps, thereby improving the real-time processing efficiency of the entire detection process. Furthermore, the local vibration intensity features of the first intermediate signal sequence are extracted, significantly enhancing the contrast between vibration events caused by vehicle passage and background noise, thereby improving the accuracy and robustness of the entire vehicle position detection.

[0049] In some embodiments of the present invention, such as Figure 7 As shown, step S204 includes: S701. A sliding window is constructed based on the second intermediate signal sequence. The sliding window includes the currently processed sample point in the second intermediate signal sequence and each of its N / 2 neighboring sample points on both sides, where N is a preset positive integer.

[0050] It should be noted that: the second intermediate signal sequence The current processing sample point in m Centered on the current sample point, take samples from both sides of the current sample point in the sequence. N / 2 neighboring sample points (left side) N / 2, right side N / Two sample points (including the currently processed sample point itself) constitute a sliding window, totaling N sample points. N The value is a preset positive integer, which determines the size of the sliding window. The total length of the sliding window is denoted as . N +1, the offset value range is from N / 2 to N / 2.

[0051] S702. Calculate the arithmetic mean of all the neighboring sample points within the sliding window.

[0052] It should be noted that: for the currently processed sample points m A central sliding window, containing all... N The arithmetic mean of the neighborhood sample points μ [ m The calculation formula for ] is as follows:

[0053] in, The second intermediate signal sequence is the first The amplitude of each sampling point.

[0054] S703. Calculate the moving variance feature value corresponding to the current processing sample point based on the amplitude of each neighboring sample point within the sliding window and the arithmetic mean. The moving variance feature value is used to characterize the local vibration intensity feature at the current processing sample point.

[0055] It should be noted that, to further highlight the local disturbance changes caused by vehicle traffic and reduce the influence of temperature drift, moving variance is needed to represent the signal as a feature. Moving variance is more sensitive to local energy fluctuations and is suitable for real-time feature extraction of non-stationary vibration signals. Specifically, this involves obtaining the arithmetic mean of all neighboring sample points within the sliding window. μ [ m After that, the moving variance eigenvalue corresponding to the current processed sample point is calculated according to the following formula. :

[0056] in, The first feature sequence obtained after calculating the moving variance 1 eigenvalue, N The sliding window length is the moving variance. μ[m] represents the offset of a sample within the sliding window relative to the center position, and μ[m] represents the arithmetic mean (local mean) of all neighboring sample points within the sliding window.

[0057] S704. Traverse all currently processed sample points in the second intermediate signal sequence, and arrange the moving variance feature values ​​corresponding to each currently processed sample point in order to obtain the feature sequence.

[0058] It should be noted that, starting from the initial sample point of the second intermediate signal sequence, each sample point is sequentially taken as the current processing sample point, and operations S1201 to S1203 are repeated until all current processing sample points in the sequence have been traversed, thus obtaining their respective moving variance eigenvalues. Arrange the moving variance eigenvalues ​​corresponding to each currently processed sample point in the index order of the original sequence to obtain the complete feature sequence.

[0059] In this embodiment, by forming a sliding window with N / 2 neighboring sample points on both sides of the current processing sample point as the center, feature extraction can be performed point-by-point following the time changes of the signal, thereby improving the precision and adaptability of feature extraction. Furthermore, by calculating the arithmetic mean of all neighboring sample points within the sliding window and calculating the moving variance feature value based on the algorithm's average, the local dispersion of the signal within the sliding window can be quantified, effectively highlighting local disturbances caused by vehicle passage while suppressing the influence of slow drift components such as temperature drift, thus improving the discriminability of the target signal in the feature domain. Finally, by traversing the entire second intermediate signal sequence and arranging each moving variance feature value in order, a complete feature sequence can be formed, providing high-quality data input for subsequent adaptive dynamic threshold detection, thereby enhancing the accuracy and robustness of the entire vehicle position detection.

[0060] In some embodiments of the present invention, such as Figure 8 As shown, step S102 includes: S801. Taking the current detection sample point in the feature sequence as the center, select Nc / 2 reference sample points on both sides of the current detection sample point in the feature sequence to form a reference window, where Nc is a preset positive integer.

[0061] It should be noted that the OS-CFAR (Ordered Statistics Constant False Alarm Rate) algorithm is introduced to calculate the dynamic threshold. The core idea of ​​the OS-CFAR algorithm is to select a certain number of reference sample points around each target sample point, and estimate the current background noise level by performing ordered statistical analysis on the amplitudes of these reference sample points, thereby generating a dynamic threshold that adapts to the background noise. Specifically, for example... Figure 9As shown, with the current detection sample point (i.e., the unit to be tested) m in the feature sequence as the center, Nc / 2 reference sample points are selected on both sides of the current detection sample point in the feature sequence (Nc / 2 on the left are the left reference units, and Nc / 2 on the right are the right reference units), for a total of Nc reference sample points forming a reference window. Here, Nc is a preset positive integer that determines the size of the reference window. A protection unit is set between the current detection sample point and the reference window. This protection unit does not participate in background noise estimation and is used to prevent the target signal from leaking into the reference window and contaminating the noise estimation. The current detection sample point itself is not included in the reference window to ensure that the background noise estimation is not affected by the signal of the current unit to be tested. The reference window slides point by point along the time axis of the feature sequence. Each time it moves to a new detection sample point position, the reference window is redefined with that point as the center, realizing point-by-point adaptive threshold detection of the entire feature sequence.

[0062] S802. Sort all the reference sample points in the reference window according to the magnitude of their amplitudes.

[0063] It should be noted that: after constructing the reference window, as... Figure 9 As shown, all Nc reference sample points within the reference window are sorted in ascending order of amplitude, resulting in the following sorted sequence:

[0064] in, To be the minimum value, This is the maximum value.

[0065] S803. Obtain the background noise power estimate based on the magnitude of the kth reference sample point after sorting; k is a preset positive integer and k≤Nc / 2.

[0066] It should be noted that after sorting, the amplitude of the kth reference sample point after sorting is selected as the estimated value of background noise power. Where k is a preset positive integer and k ≤ Nc / 2. The formula for calculating the background noise power estimate is as follows:

[0067] in, This is the estimated background noise power value. This represents the amplitude of the k-th reference sample point in the sorted sequence. Background noise power estimate. This reflects the current background noise level within the reference window. In practical engineering, The value is usually greater than, for example Figure 9 The number of protection units shown is to ensure that the selected sample is not contaminated by the target signal.

[0068] S804. Obtain a dynamic threshold corresponding to the current detection sample point based on the background noise power estimation value and a threshold factor; wherein the threshold factor is determined according to a preset system false alarm rate.

[0069] It should be noted that OS-CFAR performs ordered statistical processing on the reference window around the target signal to calculate the dynamic threshold. The signal samples in the reference window are sorted by size, and the k-th value after sorting is selected as the background noise power estimation value. This estimation value reflects the intensity of background noise, and calculates an adaptive dynamic threshold for each unit under test. Through the threshold factor and the background noise power estimation value the dynamic threshold is calculated T as shown below:

[0070] wherein, is the dynamic threshold, is the threshold factor (scaling coefficient), calculated from the false alarm probability (i.e., the system false alarm rate), which is used to control the sensitivity of the detection threshold, is the preset ordinal number of the sorted sequence after the OS-CFAR algorithm performs sorting. The system false alarm rate refers to the probability that background noise is erroneously judged as a target signal when there is no target signal. The lower the system false alarm rate, the higher the detection threshold, and the less likely the system is to generate false alarms; the higher the system false alarm rate, the lower the detection threshold, the easier it is for the system to detect the target, but false alarms will also increase accordingly.

[0071] and the relationship between them is determined by the following formula:

[0072] wherein, is the preset system false alarm rate, represents the total number of combinations for selecting from the reference samples , is the total number of samples in the OS-CFAR reference window.

[0073] In practical engineering, the threshold factor is pre-calculated and determined according to the preset system false alarm rate through formula (6) in the system design stage, and is directly used as a known fixed parameter in the real-time detection process without online calculation. After solving the threshold factor through formula (6) , it is also necessary to verify whether the actual detection performance under this parameter configuration meets the system design requirements. To this end, the following formula is further used to calculate the target detection probability , so as to evaluate the selected parameters (includingk , , Whether the combination of (etc.) can achieve sufficient detection sensitivity under the expected signal-to-noise ratio conditions. If the calculated If it's too low, it needs to be adjusted. k , or Recalculate the threshold factor using parameters such as [parameter name]. until It meets the system's detection performance requirements.

[0074] Based on the target signal's SNR (Signal-to-Noise Ratio) and the reference window length Calculate the target detection probability The formula is shown below:

[0075] in, The preset false alarm probability, This represents the signal-to-noise ratio (SNR) of the target signal. The target signal refers to the signal that the detector expects to identify from background noise indicating the passage of a vehicle. The target signal differs from the raw road vibration signal, which is unprocessed time-domain data directly acquired by the DAS system. The target signal, however, refers to the effective vibration component generated when a vehicle passes through the test area, exhibiting a significant peak in the feature domain after moving variance transformation. Target detection probability. Used to verify during the system design phase whether the selected parameters meet the expected detection performance requirements.

[0076] S805. Compare the amplitude of the current detection sample point with the dynamic threshold. If the amplitude of the current detection sample point is greater than the dynamic threshold, determine the current detection sample point as a candidate point.

[0077] It should be noted that: This involves obtaining the dynamic threshold corresponding to the current detection sample point. T Then, the amplitude v[m] of the current detected sample point is compared with the dynamic threshold. T Compare. If v[m] > T This indicates that the local vibration intensity at the current detection sample point is significantly higher than the background noise level, suggesting the possibility of a vehicle passing through. Therefore, this sample point is marked as a candidate point; if v[m]≤ T This indicates that the amplitude of the current detection sample point is within the normal fluctuation range of the background noise, and no vehicles are passing by, so it is marked as a non-candidate point.

[0078] S806. The sample segment formed by a series of candidate points in the feature sequence is taken as the candidate target interval.

[0079] It should be noted that after point-by-point detection of the entire feature sequence, sample points with amplitudes greater than the dynamic threshold are marked as candidate points. A sample segment consisting of consecutive candidate points in the feature sequence is defined as a candidate target interval. Each candidate target interval corresponds to a potential vehicle passage event.

[0080] like Figure 10 As shown, for a certain test area, the OS-CFAR adaptive dynamic threshold detection method is used to calculate the dynamic threshold point by point in the feature sequence. Figure 10 The diagram clearly illustrates the correspondence between the moving variance signal curve and the dynamic threshold curve (i.e., the OS-CFAR detection threshold). The moving variance eigenvalue corresponding to the vehicle target forms a distinct peak on the time axis, while the dynamic threshold calculated by OS-CFAR adaptively changes with the background noise level, falling below the eigenvalue peak when the vehicle target passes and remaining roughly equal to the eigenvalue when there is no target. Continuous sample segments where the moving variance signal exceeds the dynamic threshold constitute candidate target intervals. Peak search is performed within each candidate target interval, and the peak point with the largest amplitude is extracted as the initial positioning point of the target vehicle in that test area. Figure 10 (marked with a dot in the middle). Figure 10 The spikes that were not detected at the initial positioning point were accompanying (interference) signals generated by vibration transmission from adjacent lanes.

[0081] In this embodiment, by selecting an equal number of reference sample points on both sides of the current detection sample point to form a reference window, a precise local reference range can be defined for each sample point to be detected. This allows the dynamic threshold to adaptively adjust point by point according to the time changes of the signal, thereby improving the adaptability of vehicle position detection and tracking. Furthermore, by arranging all reference sample points in the reference window in ascending order of amplitude and selecting the k-th value as the background noise power estimate, the contamination of noise estimation by strong interfering targets mixed in the reference window can be effectively avoided, thereby enhancing the robustness of the detector in multi-target scenarios. Furthermore, by calculating the dynamic threshold based on the background noise power estimate and the threshold factor, the detection threshold adapts to the background noise, avoiding missed detections caused by noise fluctuations in different test areas and time periods due to a fixed threshold. Finally, by comparing the amplitude of the current detection sample point with the dynamic threshold and marking candidate points, sample points that may contain targets can be efficiently separated from the feature sequence, reducing the search range and computational overhead of subsequent processing, and improving the detection accuracy and computational efficiency of the entire vehicle position detection method.

[0082] In some embodiments of the present invention, such as Figure 11 As shown, step S104 includes: S1101. Based on the spatial distribution relationship of the multiple measurement areas, the initial positioning points are associated and matched.

[0083] It should be noted that, through the above embodiments, the initial positioning points of each survey area have been extracted, such as... Figure 12 As shown, however, these initial positioning points are currently discrete, making it impossible to determine which initial positioning points belong to the same vehicle's trajectory. Therefore, the extracted initial positioning points are correlated and matched based on the spatiotemporal constraints of the vehicle trajectory. The spatial distribution relationship of multiple measurement areas refers to the arrangement, adjacent spacing, and relative positional relationship of multiple sensing areas deployed along the highway lane direction in physical space. For example... Figure 4 As shown, the optical cables are evenly distributed along the superstructure beneath each lane of the highway, and multiple optical cable monitoring zones are organized in a specific arrangement to form an optical cable monitoring area. The spatial spacing between adjacent monitoring zones is a known system deployment parameter (e.g., ...). Figure 3 The spacing between adjacent gratings is L. Thus, the measurement area information and time information corresponding to each initial positioning point are obtained. The measurement area information includes which measurement area the positioning point originates from (i.e., the measurement area number), and the time information includes the timestamp corresponding to the positioning point. Based on this, according to the spatial distribution relationship of multiple measurement areas (i.e., the spatial distance between adjacent measurement areas), combined with the physical speed range of the vehicle, pairwise correlation judgments are made between the initial positioning points of different measurement areas.

[0084] S1102. The initial positioning points that meet the preset spatiotemporal constraints are assigned to the same candidate trajectory segment; the preset spatiotemporal constraints are that the time difference between the target vehicle and adjacent test areas is within the theoretical passage time interval.

[0085] It should be noted that: since the spatial distance between adjacent test areas is known, and vehicle speeds on highways have upper and lower limits (e.g., highway speed limits), the preset spatiotemporal constraint is that the time difference between the target vehicle and adjacent test areas must be within a reasonable theoretical transit time interval. That is, the time required for a vehicle to travel from test area A to adjacent test area B should be within the theoretical transit time interval, the lower limit of which is determined by the maximum vehicle speed (t). min =L / v max The upper limit is determined by the minimum vehicle speed (t). max =L / v min In summary, for any two initial positioning points from adjacent survey areas, calculate the time difference between them Δt = |t|. B t A |,t A Let t be the time when the vehicle passes through survey area A. B This is the moment when the vehicle passes through test area B. If L / v max ≤Δt≤L / v minThe two initial positioning points are determined to have spatiotemporal consistency and belong to the same target vehicle's trajectory. All initial positioning points that meet the above-mentioned preset spatiotemporal constraints are assigned to the same candidate trajectory segment. Each candidate trajectory segment consists of initial positioning points from multiple consecutive measurement areas arranged in chronological order, representing a possible trajectory of the target vehicle within the monitoring area.

[0086] In this embodiment, the initial positioning points are associated and matched based on the spatial distribution relationship of multiple measurement areas. This fully utilizes the inherent spatial layout relationship of the distributed fiber optic sensing system, connecting discrete initial positioning points in series according to the vehicle's movement patterns, thereby improving the accuracy of the association between positioning points. Furthermore, based on preset spatiotemporal constraints, false positioning point combinations that do not meet the vehicle's motion patterns can be effectively eliminated. This enhances the reliability of trajectory associations belonging to the same vehicle, reduces erroneous associations caused by random noise or interference from adjacent lanes, and effectively distinguishes between the vehicle's true trajectory and its accompanying trajectory, thereby improving the detection accuracy and robustness of vehicle trajectory tracking in complex traffic scenarios.

[0087] In some embodiments of the present invention, such as Figure 13 As shown, the calculation of the Poisson distribution parameters based on the candidate trajectory segments includes: S1301. Count the number of initial positioning points of the candidate trajectory segment within a preset statistical time window.

[0088] It should be noted that although each candidate trajectory segment is composed of initial positioning points that satisfy the spatiotemporal constraints arranged in chronological order, however, as... Figure 14 As shown, the candidate trajectory segments still contain two different types of trajectories: the location points of the real vehicle trajectory are relatively dense and continuous, while the location points of the accompanying interference trajectory are relatively sparse and have discontinuities. Visual observation alone is insufficient to filter out the real and interference-free effective trajectory segments. For a real vehicle, as it passes through multiple continuous measurement areas at a constant speed, the arrival of the location points on the time axis exhibits a dense and uniform statistical regularity. However, the location points of the accompanying interference trajectory originate from random coupled vibrations of non-target lanes, resulting in sparse and fragmented outliers after spatiotemporal correlation. Therefore, a Poisson distribution can be used to filter out effective trajectory segments. That is, for each candidate trajectory segment, a fixed-length statistical time window (e.g., 5 seconds) is preset. This statistical time window is then slid along the time axis of the candidate trajectory segment, and the number of initial location points contained in that candidate trajectory segment within each statistical time window is counted.

[0089] S1302. Using the maximum likelihood estimation algorithm, the Poisson distribution parameters of the candidate trajectory segments are obtained based on the quantity.

[0090] It should be noted that for each candidate trajectory segment, the occurrence of its initial positioning point within the statistical time window is considered as a sequence of random events, which follows a Poisson distribution. The Poisson distribution parameters of the candidate trajectory segment are estimated based on the quantity calculated (i.e., the count of the initial positioning points of the candidate trajectory segment within the statistical time window). The probability mass function of the Poisson distribution is shown below:

[0091] in, To appear within the observation time window The probability of each location point Let be a random variable, representing the total number of arrivals at the location point within the observation window. The actual number of locations reached within the observation window. Here, represents the intensity parameter (average arrival rate) of the Poisson distribution, indicating the average number of times a location point appears per unit time, and representing the total number of arrivals of a location point within the observation window. The length of the observation time window, is the base of the natural logarithm. This is the factorial operator. The result is calculated as follows: Then, the Poisson distribution parameter λ of the candidate trajectory segment can be calculated using the maximum likelihood estimation algorithm.

[0092] In this embodiment, by counting the number of initial positioning points of candidate trajectory segments within a preset statistical time window, the abstract trajectory density can be quantified into specific count data, avoiding the uncertainty caused by relying solely on subjective experience or simple thresholds for trajectory judgment. Furthermore, by employing the maximum likelihood estimation algorithm to estimate the Poisson distribution parameters, the objectivity and accuracy of trajectory authenticity judgment can be improved.

[0093] In some embodiments of the present invention, such as Figure 15 As shown, the step of filtering valid trajectory segments based on the Poisson distribution parameters includes: S1501. Compare the Poisson distribution parameters of each candidate trajectory segment with a preset threshold.

[0094] It should be noted that each candidate trajectory segment has already obtained a corresponding Poisson distribution parameter. .Should The quantitative value reflects the average occurrence density of the initial positioning points of the candidate trajectory segment per unit time. The larger the value, the denser the initial positioning points and the more continuous the trajectory. The smaller the value, the sparser the initial positioning points, and the more likely the trajectory will have breaks. However, based solely on... It is impossible to directly determine whether the trajectory segment is the true trajectory or a trajectory accompanied by interference, as traffic flow and vehicle speed vary in different survey areas and at different times. The numerical range will change accordingly. Therefore, a unified judgment standard needs to be established. For example... Figure 16 As shown, by performing statistical analysis on a large amount of manually labeled sample data, the Poisson distribution parameters of the true trajectory and the accompanying trajectory can be obtained respectively. Distribution curve. Poisson distribution parameters of the true trajectory. Concentrated in higher numerical ranges (densely packed with location points), accompanied by the Poisson distribution parameters of the trajectory. The distribution is concentrated in the lower numerical range (sparse location points), and there is a clear dividing region between the two distribution curves. Based on this statistical regularity, an optimal dividing point is selected within this dividing region as the preset threshold λ. th This minimizes the distinction error between the true trajectory and the accompanying trajectory. A preset threshold λ is determined. th Then, the Poisson distribution parameters of each candidate trajectory segment are... Each one is compared with the preset threshold λ th Compare them.

[0095] S1502. If the Poisson distribution parameter of the candidate trajectory segment is greater than or equal to the preset threshold, the candidate trajectory segment is determined to be the valid trajectory segment.

[0096] It should be noted that: when the Poisson distribution parameters of the candidate trajectory segment satisfy ≥λ th This indicates that the density of positioning points in the trajectory segment has reached the level expected of a real vehicle trajectory. When the vehicle passes through multiple consecutive measurement areas at a constant speed, the distribution of its positioning points on the time axis is dense and uniform, corresponding to the Poisson distribution parameters. The value is relatively large. Therefore, this candidate trajectory segment is determined to be a valid trajectory segment (i.e., the actual vehicle trajectory).

[0097] In this embodiment, by comparing the Poisson distribution parameters of candidate trajectory segments with a preset threshold, the abstract statistical characteristics (density of location points) can be transformed into a quantifiable and operable numerical comparison process. This ensures that the determination of the properties of each candidate trajectory segment has a unified objective standard, thereby avoiding the uncertainty caused by relying on human experience or subjective judgment, and improving the standardization and repeatability of trajectory decision-making. Furthermore, determining candidate trajectory segments with Poisson distribution parameters greater than or equal to the preset threshold as valid trajectory segments can effectively distinguish between real vehicle trajectories and accompanying interference trajectories without relying on additional hardware equipment. This significantly reduces false detections and pseudo-trajectory outputs caused by vibration coupling between adjacent lanes, enhances the detection accuracy and reliability of vehicle positions in complex traffic scenarios, and improves the overall performance of vehicle position detection under complex conditions of multiple lanes and vehicle types on highways.

[0098] In some embodiments of the present invention, it further includes: If the Poisson distribution parameter of the candidate trajectory segment is less than the preset threshold, the candidate trajectory segment is determined to be an interference trajectory segment and the interference trajectory segment is deleted.

[0099] It should be noted that: when the Poisson distribution parameters of the candidate trajectory segment satisfy <λ th This indicates that the location point density of the trajectory segment does not reach the level that should be present in a real vehicle trajectory. Therefore, it is determined that... <λ th The candidate trajectory segments are considered interference trajectory segments. After identifying the interference trajectory segments, they are removed from the candidate trajectory segment set. For example... Figure 17 The image shows the complete driving trajectory of real vehicles in each lane. The horizontal axis represents the measurement area number, and the vertical axis represents time (seconds). The effective vehicle trajectories of each lane are accurately output, and false and interfering positioning points are systematically eliminated. Figure 17 The trajectories of heavy-load vehicles #1 and #2 in Channel 1 were accurately preserved, while the accompanying interference trajectories of #3 and #4 in Channel 2 were effectively eliminated. The actual trajectory of the car in Channel #5 in Channel 2 was accurately preserved. False and interfering location points caused by noise and vibration coupling from adjacent lanes were effectively identified and eliminated in each channel.

[0100] In this embodiment, candidate trajectory segments with Poisson distribution parameters less than a preset threshold are identified as interference trajectory segments. This accurately identifies pseudo-vehicle trajectories caused by accompanying interference, avoiding uncertainties arising from reliance on human experience or subjective judgment, and improving the scientific rigor and reliability of trajectory decision-making. Furthermore, by deleting interference trajectory segments, accompanying interference caused by vibration coupling from heavily loaded vehicles in adjacent lanes and false positioning points caused by environmental noise can be systematically eliminated without relying on additional hardware equipment. This significantly reduces false detections and pseudo-trajectory outputs, enhances the accuracy and reliability of vehicle position detection in complex traffic scenarios, and improves the overall performance of vehicle position detection under complex conditions involving multiple lanes and vehicle types on highways.

[0101] In all specific embodiments of this application, when processing data related to user identity or characteristics, such as user information, user behavior data, user historical data, and user location information, user permission or consent is obtained first. Furthermore, the collection, use, and processing of this data comply with relevant laws, regulations, and standards. Additionally, when embodiments of this application require access to sensitive personal information, separate permission or consent from the user is obtained through pop-ups or redirects to confirmation pages. Only after obtaining the user's separate permission or consent is the necessary user-related data required for the proper functioning of these embodiments acquired.

[0102] To better implement the vehicle location tracking method for highways in this embodiment of the invention, based on the vehicle location tracking method for highways, correspondingly, as follows: Figure 18 As shown, this embodiment of the invention also provides a vehicle location tracking system 1800 for highways, the vehicle location tracking system 1800 for highways comprising: The processing module 1801 is used to preprocess the road vibration signals obtained from multiple measurement areas to obtain a feature sequence reflecting the local vibration intensity. Detection module 1802 is used to obtain multiple candidate target intervals based on the feature sequence; the candidate target intervals are intervals where the continuous amplitude exceeds the dynamic threshold; Extraction module 1803 is used to determine the sample point with the largest amplitude in each of the candidate target intervals as the initial positioning point; The association module 1804 is used to associate the initial positioning points belonging to the same target vehicle as candidate trajectory segments according to the spatial distribution relationship of the multiple measurement areas. The filtering module 1805 is used to calculate the Poisson distribution parameters based on the candidate trajectory segments and to filter valid trajectory segments according to the Poisson distribution parameters. The generation module 1806 is used to connect the effective trajectory segments in chronological order to form the driving trajectory of the target vehicle within the coverage area of ​​the multi-measurement area.

[0103] The vehicle location tracking system 1800 for highways provided in the above embodiments can realize the technical solutions described in the above embodiments of the vehicle location tracking method for highways. The specific implementation principles of each module or unit can be found in the corresponding content in the above embodiments of the vehicle location tracking method for highways, and will not be repeated here.

[0104] like Figure 19 As shown, the present invention also provides an electronic device 1900. The electronic device 1900 includes a processor 1901, a memory 1902, and a display 1903. Figure 19 Only some components of the electronic device 1900 are shown, but it should be understood that it is not required to implement all of the components shown, and more or fewer components may be implemented instead.

[0105] In some embodiments, processor 1901 may be a central processing unit (CPU), microprocessor, or other data processing chip for running program code stored in memory 1902 or processing data, such as the vehicle location tracking method for highways in this invention.

[0106] In some embodiments, processor 1901 may be a single server or a group of servers. The server group may be centralized or distributed. In some embodiments, processor 1901 may be local or remote. In some embodiments, processor 1901 may be implemented on a cloud platform. In one embodiment, the cloud platform may include a private cloud, public cloud, hybrid cloud, community cloud, distributed cloud, internal cloud, multi-cloud, etc., or any combination thereof.

[0107] In some embodiments, memory 1902 may be an internal storage unit of electronic device 1900, such as a hard disk or memory of electronic device 1900. In other embodiments, memory 1902 may also be an external storage device of electronic device 1900, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc., provided on electronic device 1900.

[0108] Furthermore, the memory 1902 may include both internal storage units of the electronic device 1900 and external storage devices. The memory 1902 is used to store application software and various types of data installed on the electronic device 1900.

[0109] In some embodiments, display 1903 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen. Display 1903 is used to display information from electronic device 1900 and to display a visual user interface. Components 1901-1903 of electronic device 1900 communicate with each other via a system bus.

[0110] In one embodiment, when processor 1901 executes a vehicle location tracking program for a highway in memory 1902, the following steps can be performed: Preprocessing of road vibration signals obtained from multiple measurement areas yields a characteristic sequence reflecting local vibration intensity; Multiple candidate target intervals are obtained based on the feature sequence; the candidate target intervals are intervals where the continuous amplitude exceeds the dynamic threshold. The sample point with the largest amplitude in each of the candidate target intervals is determined as the initial positioning point; Based on the spatial distribution relationship of the multiple measurement areas, the initial positioning points belonging to the same target vehicle are associated as candidate trajectory segments; Calculate the Poisson distribution parameters based on the candidate trajectory segments, and then filter the valid trajectory segments according to the Poisson distribution parameters; The effective trajectory segments are connected in chronological order to form the driving trajectory of the target vehicle within the coverage area of ​​multiple measurement zones.

[0111] It should be understood that when the processor 1901 executes the vehicle location tracking program for the highway in the memory 1902, in addition to the functions mentioned above, it can also perform other functions, as detailed in the description of the corresponding method embodiments above.

[0112] Furthermore, this embodiment of the invention does not specifically limit the type of electronic device 1900 mentioned. Electronic device 1900 can be a mobile phone, tablet computer, personal digital assistant (PDA), wearable device, laptop computer, or other portable electronic device. Exemplary embodiments of portable electronic devices include, but are not limited to, portable electronic devices running iOS, Android, Microsoft, or other operating systems. The aforementioned portable electronic device can also be other portable electronic devices, such as a laptop computer with a touch-sensitive surface (e.g., a touch panel). It should also be understood that in some other embodiments of the invention, electronic device 1900 may not be a portable electronic device, but rather a desktop computer with a touch-sensitive surface (e.g., a touch panel).

[0113] Those skilled in the art will understand that all or part of the processes of the methods described in the above embodiments can be implemented by a computer program instructing related hardware (such as a processor, controller, etc.), and the computer program can be stored in a computer-readable storage medium. The computer-readable storage medium may be a disk, optical disk, read-only memory, or random access memory, etc.

[0114] The present invention has provided a detailed description of the vehicle location tracking method, system, and electronic equipment for highways. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, those skilled in the art will recognize that there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A vehicle location tracking method for highways, characterized in that, include: Preprocessing of road vibration signals obtained from multiple measurement areas yields a characteristic sequence reflecting local vibration intensity; Multiple candidate target intervals are obtained based on the feature sequence; the candidate target intervals are intervals where the continuous amplitude exceeds the dynamic threshold. The sample point with the largest amplitude in each of the candidate target intervals is determined as the initial positioning point; Based on the spatial distribution relationship of the multiple measurement areas, the initial positioning points belonging to the same target vehicle are associated as candidate trajectory segments; Calculate the Poisson distribution parameters based on the candidate trajectory segments, and then filter the valid trajectory segments according to the Poisson distribution parameters; The effective trajectory segments are connected in chronological order to form the driving trajectory of the target vehicle within the coverage area of ​​multiple measurement zones.

2. The vehicle location tracking method for highways according to claim 1, characterized in that, The preprocessing of road vibration signals acquired from multiple measurement areas to obtain a feature sequence reflecting local vibration intensity includes: Acquire the road vibration signal collected by the distributed fiber optic vibration sensing system; The road vibration signal is subjected to median filtering to obtain a first intermediate signal sequence; The signal sequence obtained after median filtering is downsampled to obtain the second intermediate signal sequence; The local vibration intensity of the second intermediate signal sequence is extracted to obtain the feature sequence.

3. The method according to claim 2, characterized in that, The step of calculating the moving variance of the second intermediate signal sequence and extracting the local vibration intensity features of the second intermediate signal sequence to obtain a feature sequence includes: A sliding window is constructed based on the second intermediate signal sequence. The sliding window includes the currently processed sample point in the second intermediate signal sequence and each of its N / 2 neighboring sample points on both sides, where N is a preset positive integer. Calculate the arithmetic mean of all the neighboring sample points within the sliding window; Based on the amplitude of each neighboring sample point within the sliding window and the arithmetic mean, the moving variance feature value corresponding to the current processing sample point is calculated. The moving variance feature value is used to characterize the local vibration intensity characteristics at the current processing sample point. Traverse all currently processed sample points in the second intermediate signal sequence, and arrange the moving variance feature values ​​corresponding to each currently processed sample point in order to obtain the feature sequence.

4. The vehicle location tracking method for highways according to claim 1, characterized in that, The process involves obtaining multiple candidate target intervals based on the feature sequence. The candidate target interval is the interval where the continuous amplitude exceeds the dynamic threshold, including: Taking the current detection sample point in the feature sequence as the center, Nc / 2 reference sample points are selected on both sides of the current detection sample point in the feature sequence to form a reference window, where Nc is a preset positive integer; Sort all the reference sample points within the reference window according to their amplitude. The background noise power estimate is obtained based on the magnitude of the kth reference sample point after sorting; k is a preset positive integer and k≤Nc / 2; Based on the background noise power estimate and the threshold factor, the dynamic threshold corresponding to the current detection sample point is obtained; the threshold factor is determined according to the preset system false alarm rate. The amplitude of the current detection sample point is compared with the dynamic threshold. If the amplitude of the current detection sample point is greater than the dynamic threshold, the current detection sample point is determined as a candidate point. The sample segment consisting of multiple consecutive candidate points in the feature sequence is taken as the candidate target interval.

5. The vehicle location tracking method for highways according to claim 1, characterized in that, The step of associating the initial positioning points belonging to the same target vehicle as candidate trajectory segments based on the spatial distribution relationship of the multiple measurement areas includes: Based on the spatial distribution relationship of the multiple measurement areas, the initial positioning points are associated and matched; The initial positioning points that meet the preset spatiotemporal constraints are assigned to the same candidate trajectory segment; the preset spatiotemporal constraints are that the time difference between the target vehicle and adjacent test areas is within the theoretical passage time interval.

6. The vehicle location tracking method for highways according to claim 1, characterized in that, The calculation of Poisson distribution parameters based on the candidate trajectory segments includes: The number of initial positioning points within a preset statistical time window for the candidate trajectory segment is counted. The Poisson distribution parameters of the candidate trajectory segments are obtained by using the maximum likelihood estimation algorithm based on the number of parameters.

7. The vehicle location tracking method for highways according to any one of claims 1 to 6, characterized in that, The step of filtering valid trajectory segments based on the Poisson distribution parameters includes: The Poisson distribution parameters of each candidate trajectory segment are compared with a preset threshold. If the Poisson distribution parameter of the candidate trajectory segment is greater than or equal to the preset threshold, the candidate trajectory segment is determined to be the valid trajectory segment.

8. The vehicle location tracking method for highways according to claim 7, characterized in that, Also includes: If the Poisson distribution parameter of the candidate trajectory segment is less than the preset threshold, the candidate trajectory segment is determined to be an interference trajectory segment and the interference trajectory segment is deleted.

9. A vehicle location tracking system for highways, characterized in that, include: The processing module is used to preprocess the road vibration signals obtained from multiple measurement areas to obtain a feature sequence reflecting the local vibration intensity. The detection module is used to obtain multiple candidate target intervals based on the feature sequence; the candidate target intervals are intervals where the continuous amplitude exceeds a dynamic threshold. The extraction module is used to determine the sample point with the largest amplitude in each of the candidate target intervals as the initial positioning point; The association module is used to associate the initial positioning points belonging to the same target vehicle as candidate trajectory segments based on the spatial distribution relationship of the multiple measurement areas; The filtering module is used to calculate the Poisson distribution parameters based on the candidate trajectory segments, and to filter valid trajectory segments according to the Poisson distribution parameters. The generation module is used to connect the effective trajectory segments in chronological order to form the driving trajectory of the target vehicle within the coverage area of ​​multiple measurement zones.

10. An electronic device, characterized in that, Including memory and processor, among which, The memory is used to store programs; The processor, coupled to the memory, is used to execute the program stored in the memory to implement the steps in the vehicle location tracking method for highways according to any one of claims 1 to 7.

Images