Highway dynamic weighing system and multi-lane parallel data acquisition method thereof
By constructing a waveform data-vehicle weight feature mapping system and a parallel processing mechanism, the problems of data asynchrony and low computational efficiency in traditional dynamic weighing systems in multi-lane scenarios are solved, realizing high efficiency, accuracy and real-time parallel weighing of multiple lanes on highways, and reducing hardware resource consumption.
Patent Information
- Application Number
- CN202511438961.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-10-10
- Publication Date
- 2026-02-24
AI Technical Summary
Traditional dynamic weighing systems suffer from problems such as data asynchrony in multi-lane scenarios, poor adaptability to complex driving conditions, insufficient interference suppression, large computational load, and slow processing speed, resulting in wasted hardware resources and increased costs, making it difficult to meet the non-stop traffic demand of high-volume highways.
A waveform data-vehicle weight feature mapping system is constructed. Vehicle pressure waveform data is collected synchronously through a bar sensor group. The correspondence between waveform features and vehicle driving status and weight is established. Parallel processing and verification mechanisms are adopted to eliminate invalid data and simplify the calculation logic. Parallel calculation and data transmission are performed using FPGA chips to achieve fast and accurate weight acquisition.
It improves the stability and efficiency of multi-lane parallel processing, reduces hardware resource consumption, ensures weighing accuracy and real-time performance, and meets the non-stop passage requirements of high traffic volume on highways.
Smart Images
Figure CN121564987A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of non-stop weighing technology, specifically to a multi-lane parallel data acquisition method for a highway dynamic weighing system. Background Technology
[0002] In highway operation and management, the Dynamic Weighing System (WIM) is a core infrastructure for achieving non-stop toll collection and managing overloaded vehicles. Its performance directly affects road traffic efficiency and safety. With the annual growth rate of highway traffic volume in my country exceeding 10%, multi-lane parallel traffic scenarios have become the norm. The traditional "serial processing" architecture of the WIM is no longer suitable. After independent data collection for each lane, data filtering and calculation need to be completed sequentially. When vehicles pass through multiple lanes simultaneously, the data processing queue is prone to backlog, and the weighing time for a single vehicle often exceeds 500ms, far exceeding the 100ms threshold required for "non-stop passage," which can easily cause lane congestion during peak hours.
[0003] The complex driving conditions of vehicles in multi-lane scenarios further exacerbate the computational burden on traditional systems. In actual traffic, vehicles often engage in S-shaped lane changes, continuous lane changes, and high-speed emergency braking. To ensure accuracy, traditional systems need to perform indiscriminate calculations on all waveform data (including redundant noise signals and secondary features). The feature extraction stage alone requires processing more than 10 dimensions of parameters, and weight calculation relies on complex nonlinear fitting models, with a single calculation workload more than three times that of simplified models. This "full-scale calculation" mode not only prolongs processing time but is also prone to signal loss due to data overload, affecting weighing accuracy.
[0004] Inefficient computing logic also leads to wasted hardware resources and escalating costs in traditional systems. Existing solutions often alleviate the pressure by increasing the number of CPU cores and expanding memory, but they do not fundamentally optimize the computing logic—invalid data is not removed and models are not simplified, resulting in hardware resource utilization of less than 40%. This increases equipment procurement and maintenance costs, and fails to completely solve real-time issues, creating a vicious cycle of "hardware upgrade - resource waste - further upgrade," making it difficult to meet the long-term usage needs of multi-lane, high-traffic scenarios. Summary of the Invention
[0005] This application mainly addresses the technical problems of data asynchrony, poor adaptability to complex driving conditions, insufficient interference suppression, large computational load, and slow processing speed in existing multi-lane dynamic weighing technologies.
[0006] This application provides a multi-lane parallel data acquisition method for a highway dynamic weighing system. The highway dynamic weighing system includes a bar sensor group, which comprises N rows of bar sensor subunits arranged on multiple parallel lanes. Each bar sensor subunit includes M bar sensors, where M and N are both positive integers. The multi-lane parallel data acquisition method for the highway dynamic weighing system includes: Establish a waveform data-vehicle weight feature mapping system; When the target vehicle triggers the bar sensor array, acquire the target's original waveform data from the bar sensor. Based on the waveform data-vehicle weight feature mapping system, the target weight of the target vehicle is obtained through the original waveform data of the target vehicle. The target original waveform data includes target waveform features and target waveform parameters. The waveform data-vehicle weight feature mapping system includes at least one waveform feature and a correspondence between the waveform features and the vehicle driving state, as well as at least one set of waveform data and a correspondence between the waveform data and the vehicle weight. The at least one waveform feature and at least one set of waveform data included in the waveform data-vehicle weight feature mapping system both cover the target waveform features and target waveform parameters.
[0007] In the aforementioned multi-lane parallel data acquisition method for the highway dynamic weighing system, as a preferred embodiment, the establishment of the waveform data-vehicle weight feature mapping system includes: Establish a waveform feature-driving state mapping table that includes the correspondence between waveform features of the bar sensor group and vehicle driving state; the driving state includes parallel passage, S-shaped passage, edge passage, continuous lane change passage, high-speed passage, slow passage, and rapid speed change passage; Based on different waveform characteristics, waveform parameter-weight conversion tables representing the correspondence between waveform parameters and vehicle weight are constructed respectively.
[0008] In the aforementioned multi-lane parallel data acquisition method for the highway dynamic weighing system, as a preferred embodiment, the waveform features in the waveform feature-driving state mapping table include at least the following: The time dimension features include the trigger time difference and the variation pattern of the time difference of the N-row bar sensor sub-units; Spatial dimension characteristics, including the signal amplitude distribution and trigger position distribution of M strip sensors within the same strip sensor subunit; Multi-lane association characteristics include the triggering timing association of adjacent lane bar sensors and the degree of signal cross-interference.
[0009] In the aforementioned multi-lane parallel data acquisition method for the highway dynamic weighing system, as a preferred embodiment, the method of obtaining the target weight of the target vehicle through the target original waveform data based on the waveform data-vehicle weight feature mapping system includes: The extracted target waveform features are compared with the waveform feature-driving state mapping table to determine the driving state of the target vehicle. The original waveform data of the target is validated based on the determined driving state. The waveform data that passes the validation is considered valid data. Based on the vehicle model information of the target vehicle, the conversion relationship of the corresponding vehicle model is called in the waveform parameter-weight conversion relationship table. The target waveform parameters in the valid data are substituted into the conversion relationship to calculate the axle weight and total weight of the target vehicle.
[0010] In the aforementioned multi-lane parallel data acquisition method for a highway dynamic weighing system, as a preferred embodiment, the step of validating the target raw waveform data based on the determined driving state includes: If the driving state is parallel passage, verify the temporal continuity of waveform data within a single lane and the independence of waveform data in adjacent lanes; If the driving state is S-shaped passage, edge-pressing passage, or continuous lane change passage, verify the spatial correlation and splicing capability of waveform data from multiple rows of sensors; If the driving state is high-speed passage, slow passage, or rapid speed change passage, verify the consistency between the time characteristics of the waveform data and the speed change law; For waveform data that fails verification, cross-verification is performed by retrieving waveform data from other bar sensors in the same lane or adjacent lanes.
[0011] In the aforementioned multi-lane parallel data acquisition method for the highway dynamic weighing system, as a preferred embodiment, after calculating the axle load and total weight of the target vehicle, the method further includes: If multiple rows of bar sensor subunits in the same lane collect valid data, the multiple weight calculation results are fused together, and the fused result is used as the target weight. The target weight, corresponding axle load data, lane markings, and collection time are integrated into structured data and transmitted to the highway business system. If the target weight exceeds the normal weight range of the corresponding vehicle model, a secondary verification process is triggered. If the error persists after multiple verifications, an alarm message is pushed.
[0012] In the aforementioned multi-lane parallel data acquisition method for a highway dynamic weighing system, as a preferred embodiment, acquiring the target raw waveform data from the bar sensor includes: The pressure waveform signal of the target vehicle is synchronously collected by the bar sensor group of multiple lanes through a unified clock control system. The raw waveform signals collected from each lane are transmitted in parallel to the data processing unit via a high-speed transmission bus; Each lane is allocated an independent storage area within the data processing unit to store the target raw waveform data for that lane separately.
[0013] In the aforementioned multi-lane parallel data acquisition method for the highway dynamic weighing system, as a preferred embodiment, before obtaining the target weight of the target vehicle through the target raw waveform data based on the waveform data-vehicle weight feature mapping system, the method further includes: The original waveform data of the target is preprocessed, and signal noise is removed by filtering algorithm; Target waveform features and target waveform parameters are extracted from the preprocessed waveform data. The target waveform features include time features, spatial features, and multi-lane correlation features related to the vehicle's driving state. The target waveform parameters include the peak voltage features and duration features of the waveform.
[0014] In the aforementioned multi-lane parallel data acquisition method for a highway dynamic weighing system, as a preferred embodiment, the preprocessing of the target raw waveform data and the extraction of target waveform features and target waveform parameters from the preprocessed waveform data include: A multi-threaded parallel processing approach is adopted, with an independent processing thread assigned to the raw waveform data of each lane to perform filtering processing; The target waveform features and target waveform parameters of each lane are extracted synchronously using hardware units with parallel computing capabilities. The filtering process employs Kalman filtering algorithms or wavelet denoising algorithms, and the hardware unit for parallel computing capabilities includes an FPGA chip.
[0015] This application also provides a highway dynamic weighing system for implementing the multi-lane parallel data acquisition method of the highway dynamic weighing system described above, the system comprising: The sensing unit includes a bar sensor group and a unified clock control system. The bar sensor group is used to collect vehicle pressure waveform signals, and the unified clock control system is used to control the synchronous collection of signals by multiple lane sensors. The transmission and storage unit includes a high-speed transmission bus and independent storage modules for parallel transmission and separate storage of the original waveform data of each lane. The data processing unit includes a preprocessing module, a feature extraction module, and a weight calculation module. The preprocessing module is used for waveform data noise reduction, the feature extraction module is used for extracting waveform features and parameters, and the weight calculation module is used for calculating the vehicle weight based on a mapping system. The interaction unit includes a data integration module and an alarm module, which are used to generate structured data and transmit it to the business system, and to issue alarms for abnormal weight results. According to the above embodiment of the multi-lane parallel data acquisition method for the highway dynamic weighing system, the highway dynamic weighing system constructs a waveform data-vehicle weight feature mapping system, establishing in advance the correspondence between waveform features and vehicle driving states, and between waveform data and vehicle weight. This solves the problem of traditional systems struggling to identify complex driving states, accurately matching various states such as parallel passage and S-shaped passage, and specifically verifying data validity to avoid weighing deviations caused by state misjudgment. Furthermore, it replaces the traditional indiscriminate calculation mode of full data, requiring only the use of pre-set correspondences within the system to derive weight, significantly reducing computational load. This method utilizes N rows and M lanes on multiple parallel lanes... The layout and synchronous acquisition design of the shape sensor subunit solves the problem of data asynchrony across multiple lanes, ensuring that the target waveform characteristics and parameter timing contained in the original waveform data of each lane are consistent, avoiding data misalignment interference from parallel vehicles, and improving the stability of multi-lane parallel processing. Through the synergy of the mapping system and synchronous acquisition, the problems of large computational load and slow processing speed in traditional systems are finally solved. There is no need to build complex calculation models in real time; the target weight can be quickly obtained by directly matching the mapping relationship based on the original target waveform data. While ensuring weighing accuracy, the processing efficiency is significantly improved, meeting the core requirements of high traffic volume and non-stop passage on highways, while reducing hardware resource consumption and operation and maintenance costs. Attached Figure Description
[0016] Figure 1 A flowchart (I) of a multi-lane parallel data acquisition method for a highway dynamic weighing system provided in this application embodiment. Figure 2 A flowchart (II) of the multi-lane parallel data acquisition method for the highway dynamic weighing system provided in this application embodiment. Figure 3 A flowchart (III) of the multi-lane parallel data acquisition method for the highway dynamic weighing system provided in this application embodiment. Figure 4 This is a schematic diagram of the dynamic weighing system for highways provided in an embodiment of this application. Detailed Implementation
[0017] The present application will now be described in further detail with reference to the accompanying drawings and specific embodiments. Similar elements in different embodiments are referred to by related similar element reference numerals. In the following embodiments, many details are described to facilitate a better understanding of the present application. However, those skilled in the art will readily recognize that some features may be omitted in different situations, or may be replaced by other elements, materials, or methods. In some cases, certain operations related to the present application are not shown or described in the specification. This is to avoid obscuring the core parts of the present application with excessive description. For those skilled in the art, detailed description of these related operations is not necessary; they can fully understand the related operations based on the description in the specification and general technical knowledge in the art.
[0018] Furthermore, the features, operations, or characteristics described in the specification can be combined in any suitable manner to form various embodiments. At the same time, the steps or actions in the method description can be rearranged or adjusted in a manner obvious to those skilled in the art. Therefore, the various orders in the specification and drawings are only for the clear description of a particular embodiment and do not imply a necessary order, unless otherwise stated that a particular order must be followed.
[0019] The serial numbers assigned to components in this document, such as "first" and "second," are used only to distinguish the described objects and have no sequential or technical meaning. The terms "connection" and "linkage" used in this application, unless otherwise specified, include both direct and indirect connections (linkages).
[0020] Please refer to Figure 1 To address the technical problems of data asynchrony, poor adaptability to complex driving conditions, insufficient interference suppression, high computational load, and slow processing speed in existing multi-lane dynamic weighing systems, this application provides a multi-lane parallel data acquisition method for a highway dynamic weighing system. The highway dynamic weighing system includes a bar sensor group, which comprises N rows of bar sensor subunits arranged on multiple parallel lanes. Each bar sensor subunit includes M bar sensors, where M and N are both positive integers. The multi-lane parallel data acquisition method includes the following steps: Step S101: Establish a waveform data-vehicle weight feature mapping system. A corresponding rule base for "data-results" is built in advance to avoid complex modeling during subsequent real-time processing. This system associates abstract waveform data with specific vehicle driving states and weights, providing a basis for rapid matching and calculation, while ensuring that subsequent collected actual data can find corresponding analytical logic within the system, avoiding situations where "data has no rules to follow."
[0021] Step S102: When the target vehicle triggers the bar sensor array, acquire the target raw waveform data of the bar sensor. Emphasis is placed on "trigger-based acquisition" rather than continuous acquisition; data recording only begins when the vehicle touches the sensor, significantly reducing the amount of invalid data during periods without a vehicle. The acquired "target raw waveform data" contains two types of key information—"target waveform characteristics" reflecting the vehicle's driving state (such as the sensor trigger time difference) and "target waveform parameters" reflecting weight-related information (such as peak voltage), providing a complete data source for subsequent analysis.
[0022] Step S103: Based on the waveform data-vehicle weight feature mapping system, the target weight of the target vehicle is obtained through the target's original waveform data. By comparing the collected raw data with the correspondence in the mapping system, the vehicle's driving status is first determined to confirm the validity of the data. Then, the weight conversion rules are called to calculate the weight. This ensures accuracy while significantly reducing the amount of computation and improving processing speed, meeting the real-time requirements of non-stop passage on highways.
[0023] The target original waveform data includes target waveform features and target waveform parameters. The waveform data-vehicle weight feature mapping system includes at least one waveform feature and a correspondence between the waveform features and the vehicle driving state, as well as at least one set of waveform data and a correspondence between the waveform data and the vehicle weight. The at least one waveform feature and at least one set of waveform data included in the waveform data-vehicle weight feature mapping system both cover the target waveform features and target waveform parameters.
[0024] In the aforementioned multi-lane parallel data acquisition method for the highway dynamic weighing system, as a preferred embodiment, the establishment of the waveform data-vehicle weight feature mapping system includes the following: Step S201: Establish a waveform feature-driving state mapping table that includes the correspondence between waveform features of the bar sensor group and vehicle driving states. The driving states include parallel passage, S-shaped passage, lane-keeping passage, continuous lane-changing passage, high-speed passage, slow-moving passage, and rapid acceleration / deceleration passage. By clearly defining seven common complex driving states on highways, each state is bound to a specific waveform feature (e.g., parallel passage corresponds to "independent triggering sequence of adjacent lanes"), forming a visual and callable mapping table. Subsequently, only waveform feature comparison is needed to quickly determine the vehicle driving state, solving the problem of traditional systems struggling to identify complex states.
[0025] Step S202: Based on different waveform characteristics, construct waveform parameter-weight conversion tables to represent the correspondence between waveform parameters and vehicle weight. Considering that the correspondence between waveform parameters and weight differs under different driving conditions (e.g., the waveform duration is short when passing at high speed, resulting in a different conversion factor with weight), conversion tables are constructed separately based on different waveform characteristics (i.e., different driving conditions). This ensures that weight calculation under each condition has its own specific and accurate rules, avoiding accuracy deviations caused by a unified conversion model, and providing a clear basis for subsequent quick calls.
[0026] In the above-mentioned multi-lane parallel data acquisition method for the highway dynamic weighing system, as a preferred embodiment, the waveform features in the waveform feature-driving state mapping table include at least: time dimension features, including the trigger time difference and time difference variation law of N rows of bar sensor subunits; spatial dimension features, including the signal amplitude distribution and trigger position distribution of M bar sensors within the same bar sensor subunit; and multi-lane association features, including the trigger timing association of adjacent lane bar sensors and the degree of signal cross-interference.
[0027] For example, if the trigger time difference of N rows of sensors is uniform, it indicates that the vehicle is traveling at a constant speed; if the time difference suddenly decreases, it indicates that the vehicle is accelerating (passing through a sudden change of speed). These features can quickly capture the vehicle's time-dimensional driving status, providing key basis for subsequent status determination.
[0028] For example, if only the M edge sensors in a single row have signals, it indicates that the vehicle has passed over the edge; if the signal amplitudes of the left and right sensors fluctuate alternately, it indicates that the vehicle has passed in an S-shape. These features can accurately reflect the spatial driving status of the vehicle within the lane, supplementing the deficiencies of time-dimensional features.
[0029] For example, if adjacent lane sensors are triggered sequentially, it indicates that vehicles are changing lanes continuously; if adjacent lane signals do not cross each other, it indicates that vehicles are passing in parallel. These features are specifically designed for multi-lane scenarios, solving the problem that traditional single-lane features cannot determine cross-lane behavior and ensuring the accuracy of multi-lane parallel processing.
[0030] In the aforementioned multi-lane parallel data acquisition method for the highway dynamic weighing system, as a preferred embodiment, the method of obtaining the target weight of the target vehicle through the target original waveform data based on the waveform data-vehicle weight feature mapping system includes: Step S301: Compare the extracted target waveform features with the waveform feature-driving state mapping table to determine the driving state of the target vehicle. By comparing the target waveform features (such as time difference and amplitude distribution) extracted from the original data with the waveform feature-driving state mapping table one by one, the most matching driving state (such as slow passage or continuous lane change passage) is found, providing a state basis for subsequent targeted processing and avoiding inefficiency caused by indiscriminate processing.
[0031] Step S302: Verify the validity of the target's original waveform data based on the determined driving state. Waveform data that passes the verification is considered valid data. The criteria for valid data differ under different driving states (e.g., continuous lane changes require verification of data splicability, while slow-moving passage requires verification of data continuity). Therefore, based on the determined driving state, the corresponding verification rules are invoked to remove invalid data such as noise and disconnections, ensuring that high-quality data is used for subsequent weight calculations and improving the accuracy of the results.
[0032] Step S303: Based on the vehicle model information, the conversion relationship for the corresponding vehicle model is called from the waveform parameter-weight conversion table. The target waveform parameters in the valid data are substituted into the conversion relationship to calculate the axle load and total weight of the target vehicle. Considering that the correspondence between weight and waveform parameters varies greatly for different vehicle models (such as small cars and heavy vehicles), the vehicle model is determined first. Then, the conversion relationship for the corresponding vehicle model and driving state is called from the waveform feature-driving state mapping table. The parameters in the valid data (such as peak voltage) are substituted into the calculation. At the same time, the axle load (weight of a single axle) and the total weight (cumulative weight of all axles) are calculated to ensure the accuracy and completeness of the weight results.
[0033] In the aforementioned multi-lane parallel data acquisition method for the highway dynamic weighing system, as a preferred embodiment, the validity verification of the target raw waveform data based on the determined driving state includes the following scenarios: If the driving state is parallel passage, verify the temporal continuity of waveform data within a single lane and the independence of waveform data in adjacent lanes. "Temporal continuity" verifies that there is no interruption in the data of multiple rows of sensors within a single lane (i.e., no situation where a certain row of sensors has no signal), and "independence of adjacent lanes" verifies that there is no overlap in the data of adjacent lanes (i.e., no situation where the same vehicle triggers the sensors in adjacent lanes at the same time). Passing these two verifications ensures the data quality when passing in parallel.
[0034] If the driving state is S-shaped, edge-crossing, or continuous lane changing, verify the spatial correlation and stitching capability of the waveform data from multiple rows of sensors. "Spatial correlation" verifies that the trigger positions of multiple rows of sensors conform to the vehicle's driving trajectory (e.g., the position is S-shaped when passing in an S-shape). "Sticking capability" verifies that the data from adjacent lanes can be stitched together to form a complete trajectory during continuous lane changes, avoiding data fragmentation caused by spatial offset and ensuring that the data can be used for subsequent calculations.
[0035] If the driving conditions are high-speed passage, slow passage, or rapid speed change passage, verify the consistency between the time characteristics of the waveform data and the speed change pattern. When passing at high speed, the time characteristics should conform to the pattern of "small time difference and short waveform duration"; when passing through rapid speed change, the time difference change should conform to the speed change curve of acceleration / deceleration. Through this consistency verification, abnormal data that does not conform to the speed pattern (such as abnormal time difference caused by sensor failure) are eliminated.
[0036] For waveform data that fails verification, cross-validation is performed by retrieving waveform data from other bar sensors in the same lane or adjacent lanes. When data from a certain lane fails verification due to local sensor failure, transient interference, or other reasons, it is not discarded directly. Instead, cross-validation is performed by retrieving data from M-1 other sensors in the same lane (to supplement data from local faults) or data from adjacent lanes (to supplement cross-lane data). If the cross-validation passes, the data is still considered valid, thus improving data utilization.
[0037] In the aforementioned multi-lane parallel data acquisition method for the highway dynamic weighing system, as a preferred embodiment, after calculating the axle load and total weight of the target vehicle, the following scenarios are also included: If multiple rows of bar sensor subunits in the same lane collect valid data, the multiple weight calculation results are fused together, and the fused result is used as the target weight. When N rows of sensors in the same lane have valid data, N sets of weight calculation results will be obtained. By fusion processing (such as taking the average or weighted average), the random errors of individual rows of sensors are reduced, making the final target weight more accurate. For example, if the data from three rows are calculated to be 23.6 tons, 23.5 tons, and 23.7 tons respectively, the average value of the fused data is 23.6 tons, improving the reliability of the result.
[0038] The target weight, corresponding axle load data, lane markings, and collection time are integrated into structured data and transmitted to the highway business system. The weight results are integrated with key auxiliary information (axle load, lane, time) into structured data (such as JSON format) to ensure data format uniformity. This allows the data to be directly recognized and accessed by business systems such as toll station charging systems and overload control platforms, achieving seamless integration of "weighing data - business applications" and supporting actual business operations such as toll calculation and overload determination.
[0039] If the target weight exceeds the normal weight range for the corresponding vehicle type, a secondary verification process is triggered. If the weight remains abnormal after multiple verifications, an alarm message is sent. The normal weight range for each vehicle type is preset (e.g., small vehicles ≤ 5 tons). If the calculated weight exceeds the range, a secondary verification is triggered (retrieving the original data for review) to avoid misjudgments caused by single calculation errors. If the weight remains abnormal after multiple verifications (e.g., heavy vehicles exceeding the 49-ton limit), an alarm message is sent to staff, prompting manual intervention (e.g., intercepting overloaded vehicles) to ensure road safety.
[0040] In the above-mentioned multi-lane parallel data acquisition method of the highway dynamic weighing system, as a preferred embodiment, the acquisition of the target raw waveform data of the bar sensor includes: triggering the bar sensor group of multiple lanes to synchronously acquire the pressure waveform signal of the target vehicle through a unified clock control system; transmitting the raw waveform signals acquired by each lane in parallel to the data processing unit through a high-speed transmission bus; allocating an independent storage area for each lane in the data processing unit to store the target raw waveform data of the corresponding lane separately.
[0041] In the above-mentioned multi-lane parallel data acquisition method of the highway dynamic weighing system, as a preferred solution, before obtaining the target weight of the target vehicle through the target original waveform data based on the waveform data-vehicle weight feature mapping system, the method further includes: preprocessing the target original waveform data and removing signal noise through a filtering algorithm; extracting target waveform features and target waveform parameters from the preprocessed waveform data, wherein the target waveform features include time features, spatial features and multi-lane correlation features related to the vehicle driving state, and the target waveform parameters include the peak voltage features and duration features of the waveform.
[0042] In the aforementioned multi-lane parallel data acquisition method for a highway dynamic weighing system, as a preferred embodiment, the preprocessing of the target raw waveform data and the extraction of target waveform features and target waveform parameters from the preprocessed waveform data include: employing a multi-threaded parallel processing approach, allocating an independent processing thread to perform filtering processing on the raw waveform data of each lane; and synchronously extracting the target waveform features and target waveform parameters of each lane using a hardware unit with parallel computing capabilities; wherein the filtering processing employs a Kalman filter-type algorithm or a wavelet denoising algorithm, and the hardware unit with parallel computing capabilities includes an FPGA chip.
[0043] This application also provides a highway dynamic weighing system for implementing the multi-lane parallel data acquisition method of the highway dynamic weighing system described above, the system comprising: The sensing unit 41 includes a strip sensor group 411 and a unified clock control system 412. The strip sensor group is used to collect vehicle pressure waveform signals, and the unified clock control system is used to control the synchronous collection of multi-lane sensors. The transmission and storage unit 42 includes a high-speed transmission bus 421 and an independent storage module 422, which is used to transmit in parallel and store the original waveform data of each lane separately. The data processing unit 43 includes a preprocessing module 431, a feature extraction module 432, and a weight calculation module 433. The preprocessing module is used for waveform data noise reduction processing, the feature extraction module is used for extracting waveform features and parameters, and the weight calculation module is used for calculating vehicle weight based on a mapping system. The interaction unit 44 includes a data integration module 441 and an alarm module 442, which are used to generate structured data and transmit it to the business system, and to issue alarms for abnormal weight results.
[0044] The bar sensor group 411, through the layout of N rows and M sensors per row on multiple parallel lanes, can comprehensively capture the pressure changes of the vehicle in the time (multi-row triggering sequence) and space (triggering position within a single row) dimensions, forming raw waveform signals, ensuring that the collected information covers the entire trajectory of the vehicle and avoiding data bias caused by insufficient collection points.
[0045] The unified clock control system 412 ensures that sensors in different lanes and rows start / stop collecting data at the same time by sending a unified clock trigger signal to all bar sensors. This avoids data misalignment of parallel vehicles caused by independent timing of each sensor (e.g., data for lane A corresponds to vehicle 1, while data for lane B corresponds to vehicle 2, but is misjudged as the same vehicle due to time difference). This lays the foundation for data synchronization in multi-lane parallel processing and ensures the accuracy of subsequent data correlation analysis.
[0046] The high-speed transmission bus 421 adopts a high-speed transmission protocol (such as PCIe 4.0) to support the simultaneous transmission of raw waveform data from each lane to the data processing unit, instead of serial queuing, which greatly shortens the data transmission time. For example, data from 8 lanes can be transmitted synchronously, with a single lane transmission rate of over 10Mbps, avoiding processing delays caused by data accumulation during high traffic volumes and meeting the real-time requirements of "non-stop passage".
[0047] The independent storage module 422 allocates a separate storage area for each lane (such as a Redis-based distributed cache partition) to store only the original waveform data of the corresponding lane. This not only prevents data from different lanes from overlapping (such as lane 1 data being overwritten by lane 2 data), but also allows subsequent data processing to directly locate the target lane storage area to retrieve data without having to filter from the full dataset, thus improving data reading and processing efficiency.
[0048] The preprocessing module 431 uses algorithms such as Kalman filtering and wavelet denoising to remove irrelevant signals such as road vibration and electromagnetic interference (e.g., irregular small fluctuations in the original waveform), making the waveform signal smoother and more in line with the actual pressure changes of the vehicle. Without denoising, noise may lead to errors in subsequent feature extraction (e.g., misjudging the peak position), which in turn affects the accuracy of weight calculation. This module provides high-quality "clean data" for subsequent processing.
[0049] The feature extraction module 432 accurately extracts two types of key information from the denoised waveform: first, the "waveform features" reflecting the driving state (such as the trigger time difference of multiple rows of sensors and the amplitude distribution within a single row), which provides a basis for determining whether the vehicle is driving in parallel / changing lanes / passing at high speed; second, the "waveform parameters" associated with the weight (such as the peak voltage value and the waveform duration), which provide core input values for subsequent weight calculation. This module is a key bridge connecting "data acquisition" and "weight calculation".
[0050] The weight calculation module 433 first calls the correspondence between waveform features and driving status to determine the vehicle driving status and verify the validity of the data; then, combined with the vehicle model information, it calls the weight conversion rules of the corresponding vehicle model and substitutes the extracted waveform parameters into the calculation of axle load and total weight; if there is data from multiple rows of sensors, the results will be fused and optimized (such as taking the average value) and finally outputting the accurate target weight. This module directly realizes the core transformation "from data to weighing result".
[0051] The integration module 441 integrates information such as target weight, axle load, lane markings, and collection time into structured data (such as JSON format), ensuring that the data format meets the interface requirements of highway business systems (such as toll collection systems and overload control platforms), and achieving seamless integration between "weighing data" and "actual business"; for example, it directly provides weight data to the toll collection system for calculating fees, and provides data to the overload control platform for determining whether to intercept vehicles.
[0052] The alarm module 442 is pre-set with the normal weight range for each vehicle type. When the calculated target weight exceeds this range, a secondary verification is first triggered (data is retrieved and rechecked) to avoid misjudgment due to single calculation error. If multiple verifications are still abnormal (such as heavy vehicles exceeding the 49-ton limit), alarm information is pushed to staff through sound and light, system pop-ups, etc., to prompt manual intervention (such as intercepting overloaded vehicles) to ensure road traffic safety and reduce the safety risks caused by abnormal vehicles.
[0053] Those skilled in the art will understand that all or part of the functions of the various methods in the above embodiments can be implemented by hardware or by computer programs. When all or part of the functions in the above embodiments are implemented by computer programs, the program can be stored in a computer-readable storage medium, which may include: read-only memory, random access memory, disk, optical disk, hard disk, etc., and the program is executed by a computer to achieve the above functions. For example, the program can be stored in the memory of a device, and when the program in the memory is executed by the processor, all or part of the above functions can be achieved. In addition, when all or part of the functions in the above embodiments are implemented by computer programs, the program can also be stored in a server, another computer, disk, optical disk, flash drive, or external hard drive, etc., and can be downloaded or copied to the memory of a local device, or the system of the local device can be updated. When the program in the memory is executed by the processor, all or part of the functions in the above embodiments can be achieved.
[0054] The above examples illustrate this application only to aid understanding and are not intended to limit its scope. Those skilled in the art to which this application pertains can make various simple deductions, modifications, or substitutions based on the ideas presented.
Claims
1. A method for multi-lane parallel data acquisition in a highway dynamic weighing system, the highway dynamic weighing system comprising a bar sensor group, the bar sensor group comprising N rows of bar sensor subunits arranged on multiple parallel lanes, each bar sensor subunit comprising M bar sensors, where M and N are both positive integers, characterized in that, The multi-lane parallel data acquisition method of the highway dynamic weighing system includes: Establish a waveform data-vehicle weight feature mapping system; When the target vehicle triggers the bar sensor array, acquire the target's original waveform data from the bar sensor. Based on the waveform data-vehicle weight feature mapping system, the target weight of the target vehicle is obtained through the original waveform data of the target vehicle. The target original waveform data includes target waveform features and target waveform parameters. The waveform data-vehicle weight feature mapping system includes at least one waveform feature and a correspondence between the waveform features and the vehicle driving state, as well as at least one set of waveform data and a correspondence between the waveform data and the vehicle weight. The at least one waveform feature and at least one set of waveform data included in the waveform data-vehicle weight feature mapping system both cover the target waveform features and target waveform parameters.
2. The multi-lane parallel data acquisition method for the highway dynamic weighing system as described in claim 1, characterized in that, The establishment of the waveform data-vehicle weight feature mapping system includes: Establish a waveform feature-driving state mapping table that includes the correspondence between waveform features of the bar sensor group and vehicle driving state; the driving state includes parallel passage, S-shaped passage, edge passage, continuous lane change passage, high-speed passage, slow passage, and rapid speed change passage; Based on different waveform characteristics, waveform parameter-weight conversion tables representing the correspondence between waveform parameters and vehicle weight are constructed respectively.
3. The multi-lane parallel data acquisition method for the highway dynamic weighing system as described in claim 2, characterized in that, The waveform features in the waveform feature-driving state mapping table include at least the following: The time dimension features include the trigger time difference and the variation pattern of the time difference of the N-row bar sensor sub-units; Spatial dimension characteristics, including the signal amplitude distribution and trigger position distribution of M strip sensors within the same strip sensor subunit; Multi-lane association characteristics include the triggering timing association of adjacent lane bar sensors and the degree of signal cross-interference.
4. The multi-lane parallel data acquisition method for the highway dynamic weighing system as described in claim 2, characterized in that, The waveform data-vehicle weight feature mapping system obtains the target weight of the target vehicle through the target's original waveform data, including: The extracted target waveform features are compared with the waveform feature-driving state mapping table to determine the driving state of the target vehicle. The original waveform data of the target is validated based on the determined driving state. The waveform data that passes the validation is considered valid data. Based on the vehicle model information of the target vehicle, the conversion relationship of the corresponding vehicle model is called in the waveform parameter-weight conversion relationship table. The target waveform parameters in the valid data are substituted into the conversion relationship to calculate the axle weight and total weight of the target vehicle.
5. The multi-lane parallel data acquisition method for the highway dynamic weighing system as described in claim 4, characterized in that, The validity verification of the target raw waveform data based on the determined driving state includes: If the driving state is parallel passage, verify the temporal continuity of waveform data within a single lane and the independence of waveform data in adjacent lanes; If the driving state is S-shaped passage, edge-pressing passage, or continuous lane change passage, verify the spatial correlation and splicing capability of waveform data from multiple rows of sensors; If the driving state is high-speed passage, slow passage, or rapid speed change passage, verify the consistency between the time characteristics of the waveform data and the speed change law; For waveform data that fails verification, cross-verification is performed by retrieving waveform data from other bar sensors in the same lane or adjacent lanes.
6. The multi-lane parallel data acquisition method for the highway dynamic weighing system as described in claim 4, characterized in that, After calculating the axle load and total weight of the target vehicle, the process also includes: If multiple rows of bar sensor subunits in the same lane collect valid data, the multiple weight calculation results are fused together, and the fused result is used as the target weight. The target weight, corresponding axle load data, lane markings, and collection time are integrated into structured data and transmitted to the highway business system. If the target weight exceeds the normal weight range of the corresponding vehicle model, a secondary verification process is triggered. If the error persists after multiple verifications, an alarm message is pushed.
7. The multi-lane parallel data acquisition method for the highway dynamic weighing system as described in claim 1, characterized in that, The acquisition of the target raw waveform data of the bar sensor includes: The pressure waveform signal of the target vehicle is synchronously collected by the bar sensor group of multiple lanes through a unified clock control system. The raw waveform signals collected from each lane are transmitted in parallel to the data processing unit via a high-speed transmission bus; Each lane is allocated an independent storage area within the data processing unit to store the target raw waveform data for that lane separately.
8. The multi-lane parallel data acquisition method for the highway dynamic weighing system as described in claim 1, characterized in that, Before obtaining the target weight of the target vehicle from the original target waveform data in the waveform data-vehicle weight feature mapping system, the following steps are also included: The original waveform data of the target is preprocessed, and signal noise is removed by filtering algorithm; Target waveform features and target waveform parameters are extracted from the preprocessed waveform data. The target waveform features include time features, spatial features, and multi-lane correlation features related to the vehicle's driving state. The target waveform parameters include the peak voltage features and duration features of the waveform.
9. The multi-lane parallel data acquisition method for the highway dynamic weighing system as described in claim 8, characterized in that, The preprocessing of the target's original waveform data and the extraction of target waveform features and parameters from the preprocessed waveform data include: A multi-threaded parallel processing approach is adopted, with an independent processing thread assigned to the raw waveform data of each lane to perform filtering processing; The target waveform features and target waveform parameters of each lane are extracted synchronously using hardware units with parallel computing capabilities. The filtering process employs Kalman filtering algorithms or wavelet denoising algorithms, and the hardware unit for parallel computing capabilities includes an FPGA chip.
10. A dynamic weighing system for highways, used to implement the multi-lane parallel data acquisition method of the dynamic weighing system for highways according to any one of claims 1-9, characterized in that, The system includes: The sensing unit includes a bar sensor group and a unified clock control system. The bar sensor group is used to collect vehicle pressure waveform signals, and the unified clock control system is used to control the synchronous collection of signals by multiple lane sensors. The transmission and storage unit includes a high-speed transmission bus and independent storage modules for parallel transmission and separate storage of the original waveform data of each lane. The data processing unit includes a preprocessing module, a feature extraction module, and a weight calculation module. The preprocessing module is used for waveform data noise reduction, the feature extraction module is used for extracting waveform features and parameters, and the weight calculation module is used for calculating the vehicle weight based on a mapping system. The interaction unit includes a data integration module and an alarm module, which are used to generate structured data and transmit it to the business system, and to issue alarms for abnormal weight results.