Folding height-limiting machine self-detection and overload control system and method

Abstract

The invention relates to the technical field of control systems, and provides a folding type height limiting machine self-detection and overload control system and method, through close cooperation of a sensing fusion module, an overload control module, a self-detection module and a control management module, the precision and reliability of overload control are jointly improved, overload control is more scientific and efficient, and the system and method are suitable for popularization and application. The overload control system can automatically sense vehicle information, adjust control strategies, detect equipment faults and achieve remote control, manual intervention is reduced, the labor cost is reduced, the intelligent level of management is improved, meanwhile, the functions of early warning in advance and timely maintenance are achieved, the service life of equipment is prolonged, the maintenance cost of the equipment is reduced, and the service life of the equipment is prolonged. Through precise overload control work, the over-high vehicle is effectively prevented from entering the road, the traffic accident risk caused by illegal driving of the vehicle is reduced, and the road traffic safety is guaranteed.

Description

Technical Field

[0001] This application relates to the technical field of control systems, and particularly to a self-detection and overloading control system and method for a folding height limiter. Background Art

[0002] On national and provincial roads, the road height difference is an objective factor that cannot be ignored. The road height difference will have a significant impact on the height limit risk of large trucks. However, when the existing height limiters conduct height limit detection, they often do not fully consider the impact of the road height difference and still judge according to a fixed height limit standard. This leads to the situation that trucks that are not actually over-height may be misjudged as over-height when encountering road bumps or depressions due to the road height difference, or trucks that are actually over-height may not be detected due to the road height difference, thus increasing road safety hazards and reducing the accuracy and effectiveness of overloading control work.

[0003] At the same time, when the currently common folding height limiters sense vehicle information, they often rely only on a single sensor, such as a simple height sensor or weight sensor, making the obtained vehicle data incomplete and difficult to accurately identify key information such as the vehicle's contour, type, and driving state. Moreover, the overloading control strategy of traditional folding height limiters is relatively fixed and difficult to dynamically adjust according to actual environmental factors and real-time vehicle conditions. At the same time, the self-detection function of folding height limiters is relatively weak and cannot monitor the operating parameters of each component in real time and detect potential faults in a timely manner. Summary of the Invention

[0004] In view of the deficiencies of the prior art, this application provides a self-detection and overloading control system and method for a folding height limiter.

[0005] In a first aspect, this application provides a self-detection and overloading control system for a folding height limiter, which includes: a perception fusion module, an overloading control module, a self-detection module, and a control and management module;

[0006] The perception fusion module is used to obtain the perception data of the target vehicle in the detection area through a sensor network and perform data association and alignment on the perception data of the target vehicle;

[0007] The overloading control module is used to receive the perception data of the target vehicle, obtain the road slope in the detection area, extract the features of the perception data as contour features and state features, extract the feature of the road slope as the road height difference, perform correlation analysis on the road height difference with the contour features and state features, and obtain the analysis result to dynamically adjust the overloading control strategy of the folding height limiter;

[0008] The self-detection module is used to monitor the operating parameters of each component of the folding height limiter during the overloading control process to identify the fault modes of the folding height limiter and trigger feedback information;

[0009] The control and management module is used to control the folding height limiter according to the feedback information and synchronously perform the power supply management of the folding height limiter.

[0010] As an alternative implementation, the perception data includes contour data, visual data, and status data;

[0011] The logic for obtaining the perception data of the target vehicle in the detection area includes:

[0012] Deploy a sensor network according to the detection area of the folding height limiter. When the target vehicle enters the detection area, a signal acquisition is triggered;

[0013] After the sensor network receives the acquisition signal, it obtains the contour data, visual data, and status data of the target vehicle;

[0014] Perform data association and alignment on the contour data, visual data, and status data of the target vehicle to obtain the perception data of the target vehicle.

[0015] As an alternative implementation, the sub-logic of the data association and alignment includes:

[0016] Add a timestamp and associate and store the timestamp with the contour data, visual data, and status data of the target vehicle;

[0017] Define a unified spatial coordinate system based on the detection area of the folding height limiter, and convert the contour data, visual data, and status data of the target vehicle to the spatial coordinate system and perform spatial registration and alignment to obtain the perception data of the target vehicle;

[0018] Verify and correct the perception data that has completed timestamp association and spatial registration alignment.

[0019] As an alternative implementation, the adjustment logic of the overloading control strategy includes:

[0020] Receive the perception data of the target vehicle, obtain the environmental data of the detection area, and integrate the perception data of the target vehicle and the environmental data of the detection area into vehicle environmental data;

[0021] Perform data analysis on the vehicle environmental data to generate an analysis result;

[0022] Generate an overloading control strategy based on the analysis result;

[0023] Send the overloading control strategy to the folding height limiter to control the lifting action of the height limit bar, the working state of the audible and visual alarm, and the prompt content of the information display screen;

[0024] Collect the effect feedback data during the overloading control process to dynamically adjust the overloading control strategy.

[0025] As an alternative implementation, the generation sub-logic of the analysis result includes:

[0026] Extract the environmental features of the detection area in the vehicle environmental data, where the environmental features include road height differences;

[0027] Extract the contour features and status features of the perception data of the target vehicle in the vehicle environmental data;

[0028] Perform correlation analysis on the road height difference, contour features, and status features based on a graph neural network;

[0029] Quantify the vehicle risk based on the results of the correlation analysis.

[0030] As an alternative implementation, the sub-logic of the correlation analysis includes:

[0031] Use the road slope of the detection area and the perception data of the target vehicle as the nodes of the graph to construct an initial graph structure;

[0032] Perform feature fusion on the road height difference, contour features, and status features through an attention mechanism to obtain fused features, and input the fused features into the graph neural network to obtain a correlation result;

[0033] Mine the causal relationship between the road slope of the detection area and the behavior of the target vehicle through a causal inference algorithm to obtain a causal inference result;

[0034] Verify the correlation result and the causal inference result to generate the result of the correlation analysis.

[0035] As an alternative implementation, the trigger logic of the feedback information includes:

[0036] Monitor the operating parameters of each component of the folding height limiter during the over-limit control process;

[0037] Judge whether the folding height limiter fails by comparing the operating parameters of each component with thresholds, and identify the failure mode of the folding height limiter;

[0038] Generate feedback information according to the failure mode of the folding height limiter.

[0039] As an alternative implementation, the monitoring sub-logic of the operating parameters includes:

[0040] Construct the deployment layout of the sensor network according to the mechanical structure of the folding height limiter and the environmental data of the detection area;

[0041] Determine the state space, action space, and reward function of the sensor nodes;

[0042] Train the sensor nodes through a reinforcement learning algorithm, gradually optimize to obtain the deployment plan of the sensor network, and monitor the operating parameters of each component of the folding height limiter through the sensor network according to the deployment plan of the sensor network.

[0043] As an alternative implementation, the deployment logic of the sensor network includes:

[0044] Determine a combination of at least two lidars, three cameras, and two millimeter-wave radars to obtain contour data, visual data, and status data of the target vehicle;

[0045] According to the actual installation environment of the folding height limiter and the road slope in the detection area, plan the installation positions and angles of the lidars, cameras, and millimeter-wave radars;

[0046] Adaptive adjustment of the operating parameters of the lidars, cameras, and millimeter-wave radars.

[0047] In a second aspect, the present application provides a self-detection and over-limit control method for a folding height limiter. The method includes: S1. Obtain the perception data of the target vehicle in the detection area through the sensor network;

[0048] S2. Receive the perception data of the target vehicle, and dynamically adjust the over-limit control strategy of the folding height limiter in combination with the environmental data of the detection area;

[0049] S3. Monitor the operating parameters of each component of the folding height limiter during the over-limit control process to identify the fault modes of the folding height limiter and trigger feedback information;

[0050] S4. Control the folding height limiter according to the feedback information and synchronously execute the power supply management of the folding height limiter.

[0051] Compared with the prior art, the beneficial effects of the present application are as follows: Through the close cooperation of the perception fusion module, over-limit control module, self-detection module, and control management module, the accuracy and reliability of over-limit control are jointly improved, making over-limit control more scientific and efficient. The over-limit control system can automatically sense vehicle information, adjust control strategies, detect equipment failures, and achieve remote control, reducing manual intervention, lowering labor costs, and improving the intelligent level of management. At the same time, the functions of early warning and timely maintenance extend the service life of the equipment, reduce equipment maintenance costs. Through precise over-limit control work, it effectively prevents over-height vehicles from entering the road, reduces the risk of traffic accidents caused by vehicle violations, and ensures road traffic safety. Reliable equipment operation and power supply management ensure that the height limiter can work normally under various conditions, avoiding potential safety hazards caused by equipment failures or power supply problems, and further improving the safety of the entire traffic system.

[0052] The sensor network in the perception fusion module can obtain multi-dimensional information such as the contour data, visual data, and status data of the target vehicle. The collaborative work of multiple sensors and data fusion technology greatly improves the accuracy and comprehensiveness of vehicle information acquisition, effectively reduces misjudgment and missed judgment situations, and provides reliable data support for subsequent over-limit control.

[0053] The over-limit control module can dynamically adjust the over-limit control strategy of the folding height limiter in combination with environmental data, automatically optimize the over-limit control strategy according to different weather, road, and traffic conditions, enable the folding height limiter to maintain an efficient over-limit control effect in various complex situations, and at the same time ensure the normal passage of traffic to the greatest extent, improving the overall operation efficiency of the road.

[0054] The self-detection module real-time monitors the operating parameters of each component of the folding height limiter, can timely detect potential faults of the equipment, and trigger feedback information. The early warning function enables maintenance personnel to make preparations before the occurrence of faults, perform repairs and maintenance in a timely manner, reduces the impact of equipment failures on over-limit control work, reduces maintenance costs, and improves the reliability and service life of the equipment.

[0055] The control and management module realizes the remote control function and timely reminder function. Managers can conveniently control and manage the folding height limiter through the remote management terminal, improving work efficiency; in terms of power supply, not only is a backup power supply configured, but also energy recovery technology is adopted to convert the mechanical energy during the lifting and lowering of the height limit bar into electrical energy for storage, and the recovered electrical energy is preferentially used when the power supply is interrupted, ensuring the normal operation of the folding height limiter in emergency situations, improving the reliability and stability of the power supply of the folding height limiter, and at the same time realizing the effective utilization of energy. Brief Description of the Drawings

[0056] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings. Among them:

[0057] Figure 1 It is the system structure diagram of a folding height limiter self-detection and over-limit control system provided by the embodiment of the present application;

[0058] Figure 2 It is the associated analysis sub-logic diagram of a folding height limiter self-detection and over-limit control system provided by the embodiment of the present application;

[0059] Figure 3Schematic diagram of the folding height limiter structure of a self - detection and over - load control system provided by an embodiment of the present application;

[0060] Figure 4 Method flowchart of a self - detection and over - load control method for a folding height limiter provided by an embodiment of the present application. Detailed implementation manners

[0061] To make the objectives, technical solutions, and advantages of the embodiments of the present application more apparent and understandable, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings of the specification. Apparently, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments.

[0062] Embodiment 1

[0063] As Figure 1 shown, an embodiment of the present application provides a system structure diagram of a self - detection and over - load control system for a folding height limiter. The system includes a perception fusion module, an over - load control module, a self - detection module, and a control and management module.

[0064] The target vehicle in this embodiment represents a large truck.

[0065] The perception fusion module is used to obtain the perception data of the target vehicle in the detection area through a sensor network.

[0066] The perception data includes contour data, visual data, and status data.

[0067] The logic for obtaining the perception data of the target vehicle in the detection area includes:

[0068] Deploy a sensor network according to the detection area of the folding height limiter. When the target vehicle enters the detection area, a signal acquisition trigger is activated;

[0069] After receiving the acquisition signal, the sensor network obtains the contour data, visual data, and status data of the target vehicle;

[0070] Perform data association and alignment on the contour data, visual data, and status data of the target vehicle to obtain the perception data of the target vehicle.

[0071] The deployment logic of the sensor network includes:

[0072] Determine a combination of at least two lidars, three cameras, and two millimeter - wave radars to obtain the contour data, visual data, and status data of the target vehicle;

[0073] According to the actual installation environment of the folding height limiter and the road slope in the detection area, plan the installation positions and angles of the lidars, cameras, and millimeter - wave radars;

[0074] Adaptively adjust the working parameters of lidar, camera, and millimeter-wave radar.

[0075] According to the detection requirements of the target vehicle, select appropriate sensors. Currently, it is necessary to obtain the contour data, visual data, and status data of the target vehicle. Then, use lidar to obtain contour data such as the contour and height of the target vehicle, use the camera to identify visual data such as the type and license plate of the target vehicle, and use millimeter-wave radar to monitor the speed and position of the target vehicle in real time. Measure based on the actual installation environment of the folding height limiter and the road slope of this section within the detection area, and reasonably plan the installation positions and angles of the above sensors. For example, install the lidar at a high position in front of the folding height limiter so that the lidar can cover the vertical and horizontal spaces where the target vehicle approaches from a distance to obtain the omnidirectional contour data of the target vehicle. Install another lidar on the side of the folding height limiter to supplement the detailed data of the side contour and ensure that the scanning range of the lidar can effectively cover the space where the target vehicle appears. Install three cameras at different angles to ensure that the front, side, and rear of the target vehicle can be photographed, and the cameras can completely cover the passing area of the target vehicle without visual blind spots. The millimeter-wave radar needs to be installed in key areas where the speed and position of the target vehicle can be effectively monitored, and the detection direction of the millimeter-wave radar is adapted to the driving direction of the target vehicle to ensure that the detection ranges of each sensor complement each other and cover the driving area of the target vehicle without dead angles, ensuring the comprehensiveness and effectiveness of the acquisition of perception data.

[0076] During the startup phase of the overloading control system, perform initialization operations on all sensor networks and read the default parameters of the sensor networks, such as the scanning frequency and resolution of lidar, the exposure time and frame rate of the camera, and the working frequency band of the millimeter-wave radar. According to the actual application scenario and detection accuracy requirements, configure the working parameters of the sensor network specifically. For example, adjust the ranging accuracy and scanning angle range of the lidar according to the height limit range and detection accuracy requirements of the folding height limiter; set the automatic exposure and dynamic white balance functions for the camera in an environment with large light changes; optimize the detection range and frequency of the millimeter-wave radar for the detection requirements of vehicles with different speeds.

[0077] In reality, there may be situations of light and electromagnetic interference in the detection area of a folding height limiter. It is necessary to adaptively adjust the working parameters of lidar, cameras, and millimeter-wave radars. For example, the light intensity sensor built into the lidar collects ambient light intensity data in real time. When the sunlight is strong during the day, the light intensity sensor detects a high light intensity, and the control system will increase the scanning frequency of the lidar to ensure that the contour data of the target vehicle can be obtained even in strong light environments. When the light is dim in the evening, the light intensity sensor detects a low light intensity, and the control system will reduce the scanning frequency of the lidar and increase the resolution to enhance the recognition ability of the target vehicle in low light environments. At the same time, according to the prediction of the road slope and the driving trajectory of the target vehicle in the detection area, when the target vehicle is driving on a section with a large slope, the control system will automatically adjust the scanning angle range of the lidar to ensure effective coverage of the target vehicle at different slopes and target vehicle driving positions.

[0078] The camera uses image histogram analysis technology to analyze the brightness distribution of the collected images in real time. When the overall image is too dark, the control system automatically increases the exposure time and appropriately increases the gain value to restore the image brightness to normal. When the overall image is too bright, the control system reduces the exposure time and shrinks the aperture size to avoid overexposure of the image. Combining the road slope and the driving direction of the target vehicle, when the target vehicle is in a climbing state, the control system dynamically adjusts the focal length and shooting angle of the camera to ensure clear shooting of the target vehicle at different slopes and target vehicle driving postures.

[0079] The spectrum monitoring module of the millimeter-wave radar scans the electromagnetic spectrum of the detection area in real time. When an interference signal similar to its own working frequency band is detected, the control system will adjust the working frequency band of the millimeter-wave radar. At the same time, according to the changes in the driving speed of the target vehicle and the road slope, when the target vehicle is driving slowly on a steep slope, the control system appropriately increases the transmission power of the millimeter-wave radar to ensure accurate monitoring of the speed and position of the target vehicle at a long distance. When the target vehicle is driving at high speed on a flat section, the control system will adjust the detection distance of the millimeter-wave radar to meet the detection requirements for the high-speed driving of the target vehicle.

[0080] A geomagnetic sensor is set at the entrance of the detection area of the folding height limiter. When the target vehicle enters the detection area, it triggers the acquisition signal of the sensor network. When the sensor network receives the acquisition signal, it immediately starts the data acquisition process. The lidar quickly scans the contour and height of the target vehicle, the camera starts to continuously capture the visual images of the target vehicle, and the millimeter-wave radar monitors the speed and position changes of the target vehicle in real time to ensure comprehensive acquisition of the perception data of the target vehicle during the process of the target vehicle passing through the detection area.

[0081] During the data acquisition process, preliminary screening is performed on the original data. For example, in the point cloud data obtained by lidar scanning, outliers generated due to environmental interference are filtered out. In the visual images obtained by the camera, visual images that are significantly blurred and have abnormal exposure are removed. And in the vehicle speed and vehicle position data obtained by the millimeter-wave radar, outliers beyond the reasonable range are eliminated. Then, preprocessing is performed on the preliminarily screened data. The lidar point cloud data is subjected to coordinate transformation and normalization processing, the visual images are subjected to grayscale conversion, noise reduction, and contrast enhancement processing, and the millimeter-wave radar data is subjected to filtering and signal amplification processing to improve the quality and usability of the overall perception data and lay a foundation for subsequent data association and alignment.

[0082] Specifically, the sub-logic of data association and alignment includes:

[0083] Add a timestamp and associate and store the timestamp with the contour data, visual data, and status data of the target vehicle.

[0084] Define a unified spatial coordinate system based on the detection area of the folding height limiter, and convert the contour data, visual data, and status data of the target vehicle to the spatial coordinate system and perform spatial registration and alignment to obtain the perception data of the target vehicle.

[0085] Verify and correct the perception data that has completed timestamp association and spatial registration and alignment.

[0086] During each data acquisition, after each sensor network obtains the perception data of the target vehicle, an accurate timestamp is immediately added. The accuracy of the timestamp should meet the requirements of the control system for time synchronization, such as reaching the millisecond level or microsecond level. The timestamp is associated and stored with the obtained contour data, visual data, and status data of the target vehicle to form a perception data record with time tags. For example, for the contour data depicted by the lidar point cloud data and the status data of the vehicle speed and position measured by the millimeter-wave radar, a timestamp field is added to the data structure. For the visual image data obtained by the camera, the timestamp is saved as the metadata information of the visual image.

[0087] According to the actual installation position and detection area of the folding height limiter, define a unified spatial coordinate system. For example, take a certain fixed point of the folding height limiter as the coordinate origin of the spatial coordinate system and determine the direction of the coordinate axes, where the horizontal direction is the X-axis, the vertical direction is the Y-axis, and the vehicle driving direction is the Z-axis. For each sensor network, determine its installation position and attitude parameters in the spatial coordinate system. These parameters include the three-dimensional coordinates (X, Y, Z) of the sensor network and the rotation angles around the three coordinate axes. The rotation angles include the pitch angle, yaw angle, and roll angle.

[0088] By defining the installation location and attitude parameters of the sensor network and the spatial coordinate system, a conversion relationship from the local coordinate system of the sensor network to the unified spatial coordinate system is established. For example, for a camera, the internal and external parameters of the camera are obtained using camera calibration technology to establish a conversion matrix from the camera coordinate system to the spatial coordinate system. For lidar and millimeter-wave radar, according to their installation location and attitude parameters, the corresponding coordinate conversion formulas are determined.

[0089] Spatial registration and alignment are performed on the contour data, visual data, and status data of different sensor networks converted to the spatial coordinate system. By comparing the position information of the data obtained by different sensor networks in the spatial coordinate system, it is checked whether there are deviations. If there are deviations, the contour data, visual data, and status data are fine-tuned through a spatial registration algorithm to make the data obtained by different sensor networks accurately aligned in space, ensuring that the contour data, visual data, and status data obtained by each sensor network correspond to the target vehicle information at the same position.

[0090] After completing the timestamp association and spatial registration and alignment, the obtained perception data is verified. By comparing the characteristic information of the data obtained by different sensor networks (such as the contour characteristics and speed characteristics of the target vehicle, etc.), it is checked whether the perception data after timestamp association and spatial registration and alignment accurately corresponds to the same target vehicle; if it is found that the perception data after timestamp association and spatial registration and alignment is incorrect or inconsistent, it is corrected through manual intervention, such as by an artificial operator making judgments and adjustments according to the actual situation, ensuring the accuracy of data association and alignment.

[0091] The overloading control module is used to receive the perception data of the target vehicle, obtain the road slope in the detection area, extract the characteristics of the perception data as contour characteristics and status characteristics, extract the characteristics of the road slope as the road height difference, and perform correlation analysis on the road height difference with the contour characteristics and status characteristics to obtain the analysis result to dynamically adjust the overloading control strategy of the folding height limiter.

[0092] The adjustment logic of the overloading control strategy includes:

[0093] Receive the perception data of the target vehicle, obtain the environmental data of the detection area, and integrate the perception data of the target vehicle and the environmental data of the detection area into vehicle environmental data;

[0094] Perform data analysis on the vehicle environmental data to generate an analysis result;

[0095] Generate an overloading control strategy based on the analysis result;

[0096] Send the overloading control strategy to the folding height limiter to control the lifting action of the height limit bar, the working state of the audible and visual alarm, and the prompt content of the information display screen;

[0097] Collect the effect feedback data in the process of overloading control to dynamically adjust the overloading control strategy.

[0098] In order to comprehensively understand the driving conditions of the target vehicle and the environmental conditions of the detection area, it is necessary to integrate the perception data of the target vehicle and the environmental data of the detection area to provide a basis for subsequent data analysis. Real-time receive the finally obtained perception data of the target vehicle, including the contour data of the target vehicle to judge whether the size of the target vehicle is abnormal, the visual data of the target vehicle to distinguish the height and weight limits of different types of vehicles, and the status data of the target vehicle to judge whether the target vehicle is speeding or its position is abnormal; at the same time, obtain the environmental data in the detection area, such as environmental factors such as weather conditions, road surface slipperiness, road gradient, and surrounding traffic flow. Integrate these perception data of the target vehicle and the environmental data of the detection area and store them in a unified data structure. When a large truck enters the detection area, the lidar quickly obtains the contour data of the large truck, the camera captures the external image of the large truck, the millimeter-wave radar real-time monitors the speed and position of the large truck, and at the same time, the road height difference on the national and provincial roads is obtained in real time through the terrain measurement equipment, and the bumps and depressions of the road sections in the detection area are recorded, providing comprehensive data support for subsequent data analysis, ensuring the comprehensiveness and accuracy of data analysis, and avoiding analysis errors caused by missing or incomplete data.

[0099] Specifically, the generation sub-logic of the analysis result includes:

[0100] Extract the environmental characteristics of the environmental data of the detection area in the vehicle environmental data, and the environmental characteristics include the road height difference;

[0101] Extract the contour characteristics and status characteristics of the perception data of the target vehicle in the vehicle environmental data;

[0102] Perform correlation analysis on the road height difference, contour characteristics and status characteristics based on the graph neural network;

[0103] Quantify the vehicle risk based on the results of the correlation analysis.

[0104] Perform normalization processing on the obtained environmental data, such as weather conditions, road surface slipperiness, road gradient, and surrounding traffic flow, to unify the environmental data with different dimensions to the same scale range for convenient subsequent analysis. At the same time, use a filtering algorithm to perform noise reduction processing on the environmental data to remove outliers caused by sensor errors or environmental interference and improve the accuracy of the environmental data.

[0105] Extract the environmental features of environmental data through deep learning feature extraction methods, with a focus on extracting the feature of road slope as the road height difference. For example, in a certain convex section, the road height difference reaches 3 meters, and in a certain concave section, the road height difference is -2 meters. For the surrounding traffic flow data, extract the peak and trough period features and the traffic flow distribution features of different sections, etc., to provide more representative environmental features for subsequent correlation analysis.

[0106] Analyze the contour data of the target vehicle through a convolutional neural network. By training the convolutional neural network to analyze the contour data of the target vehicle, extract contour features such as the height of the cargo platform and the length of the vehicle body of the target vehicle, accurately judge whether the target vehicle has an over-height risk. At the same time, combined with semantic segmentation technology, separate the contour of the target vehicle from the background, further refine the analysis of the vehicle size, and improve the detection accuracy.

[0107] By analyzing the state change of the speed and acceleration of the target vehicle over a period of time, predict the driving state of the target vehicle, so as to obtain the state features of the perception data. For example, when it is predicted that the vehicle speed of the target vehicle drops too fast and the acceleration is abnormal when climbing a slope, combined with whether the height difference of the road slope in the detection area is convex or concave, judge whether the target vehicle has an over-height problem, and successfully extract the environmental features and vehicle features, which prepares for subsequent correlation analysis and improves the efficiency and accuracy of data analysis.

[0108] Furthermore, the sub-logic of the correlation analysis is as Figure 2 shown, specifically including:

[0109] Use the road slope of the detection area and the perception data of the target vehicle as the nodes of the graph to construct an initial graph structure;

[0110] Through the attention mechanism, perform feature fusion on the road height difference, contour features, and state features to obtain the fusion features, and input the fusion features into the graph neural network to obtain the correlation result;

[0111] Use the causal inference algorithm to mine the causal relationship between the road slope of the detection area and the behavior of the target vehicle to obtain the causal inference result;

[0112] Verify the correlation result and the causal inference result to generate the result of the correlation analysis.

[0113] Taking the road slope of the detection area and the perception data of the target vehicle as the nodes of the graph, classify them according to the category and nature of the data. For example, the road height difference is used as an environmental node, and vehicle factors such as vehicle type, contour size, and speed are used as perception nodes. Initialize the connection relationship between the environmental node and the perception node according to the existing knowledge in the relevant field, and construct an initial graph structure. For example, based on the actual situation, it is known that the road height difference will directly affect the relative height between the large truck and the height limit pole. Therefore, establish the connection between the environmental node and the height node and the body length node of the large truck's cargo platform, so that the correlation between the environmental data and the perception data is clearer and convenient for in-depth analysis.

[0114] Fuse the road height difference, contour features, and state features, and automatically learn the importance of different features in the correlation analysis through the attention mechanism. For example, when analyzing that the target vehicle bumps due to the road height difference during driving, resulting in the target vehicle being overheight, the attention mechanism will pay more attention to the contour features (height of the large truck's cargo platform), state features (such as speed and acceleration), and the features of the road height difference of the target vehicle. When there are bumps or depressions on the road, the changes in the height of the large truck's cargo platform and the road height difference will directly affect whether the large truck is overheight. Then the attention mechanism will assign higher weights to these features to obtain the fused features, which are used as the input of the graph neural network. Use a large amount of historical road slopes and perception data to train the graph neural network model. At the same time, during the training process, adopt optimization algorithms such as stochastic gradient descent to continuously adjust the connection weights of the graph and the model parameters of the graph neural network to minimize the loss function between the predicted value and the true value. For example, use the actual driving stability indicators of the vehicle (such as roll angle and acceleration change) as the true value, and the driving stability predicted by the graph neural network model as the predicted value. Through multiple iterative trainings, improve the accuracy and generalization ability of the graph neural network model to obtain the correlation result, so as to reveal the internal relationship between the road height difference and the contour features and state features of the large truck, and accurately analyze the impact of the road height difference on the height limit of the large truck, providing a reliable basis for risk quantification; for example, when there is a bump on the national and provincial roads, the graph neural network will pay more attention to the height of the large truck's cargo platform because the bump on the road will make the distance between the large truck and the height limit pole smaller, increasing the risk of overheight. When there is a depression on the road, the graph neural network will comprehensively consider the speed and the height of the large truck's cargo platform because the speed change of the large truck when driving on the depression section will affect the stability of the large truck and also affect the relative height between the large truck and the height limit pole.

[0115] Use causal inference algorithms (such as propensity score matching, PSM) to comprehensively identify the confounding factors in the causal relationship between road height differences and vehicle over-height risks, such as the type of target vehicle (empty load and full load), the type of road height difference (hump or depression), and the changes under traffic flow (peak or trough). Through stratified analysis and interaction analysis, obtain causal inference results. For example, when fully loaded, the cargo platform height of large trucks is relatively high, and when there are humps on the road, it is more likely to exceed the height limit. Thus, through causal inference algorithms, the direct and indirect impacts of road height differences on the height limit of target vehicles, as well as the variation laws under different conditions, are clarified, providing theoretical support for formulating more effective over-height control strategies.

[0116] Verify the correlation results and causal inference results. Through the method of cross-validation, divide the data set into a training set and a test set. Train the graph neural network on the training set, and calculate metrics such as accuracy, recall rate, and mean square error on the test set to verify the correlation results and causal inference results, and evaluate the performance of the graph neural network model and the reliability of the causal relationship. For example, verify the accuracy of the graph neural network model in predicting the over-height risks of vehicles under different types of road height differences, and the degree of conformity between the causal inference results and the actual situation. During the verification process, it is found that when the road height difference changes significantly, the accuracy of the graph neural network model in evaluating the over-height risks of large trucks decreases. By adjusting the parameters of the graph neural network model and optimizing the algorithm, the adaptability of the graph neural network model under different road height difference conditions is improved, thus ensuring the accuracy and reliability of the analysis results and providing a guarantee for formulating effective over-height control strategies.

[0117] Based on the results of the correlation analysis, establish a risk assessment index system to quantitatively evaluate the over-height risks of target vehicles. Comprehensively consider factors such as road height differences, the contour features and status features of large trucks, and quantify the over-height risks of large trucks. During the risk quantification process, particularly consider the corrective effect of road height differences on the over-height limit of large trucks. For example, when there is a hump on the road and the cargo platform height of the large truck is relatively high, according to the preset weight coefficient, increase the over-height risk level of the large truck. When there is a depression on the road and the large truck is traveling at a relatively high speed, the risk assessment of the large truck colliding with the over-height pole due to instability will be increased. Calculate a risk score for each target vehicle and set a risk threshold to determine different risk levels, such as low risk, medium risk, and high risk, in order to classify and manage target vehicles, clearly understand the over-height risk status of large trucks, and provide a clear reference for formulating over-height control strategies. When it is detected that the determined risk level of the target vehicle is medium risk and high risk, the over-height control system issues an abnormal warning and broadcasts a prompt through an audible and visual alarm to take further measures to ensure road safety.

[0118] Based on the vehicle risks quantified according to the results of association analysis, combined with preset overloading control rules, multiple overloading control strategies are generated. For example, when it is detected that the road height difference is a bump and the target vehicle is in a high-risk state, the overloading control strategy is to send out a warning signal in advance to remind the truck driver to pay attention to the height limit and extend the detection time of the folding height limit machine; when there is a depression on the road and the large truck is moving at a high speed, control the height limit bar to rise a certain height in advance to prevent the large truck from colliding with the height limit bar due to jolting; when the traffic flow is large, the lifting and lowering frequency of the folding height limit machine can be appropriately adjusted to optimize the vehicle passing efficiency; use intelligent decision-making algorithms, such as evaluating and screening the multiple overloading control strategies generated through decision trees. The factors considered in the evaluation and screening include the safety, feasibility of the overloading control strategy, and its impact on traffic fluency, etc., and finally determine the optimal overloading control strategy.

[0119] Send the determined overloading control strategy to the folding height limit machine to control the lifting and lowering actions of the height limit bar, the working state of the sound and light alarm, and the prompt content of the information display screen. For example, when it is determined that the target vehicle is overweight, the height limit bar remains in the raised state, the sound and light alarm sounds an alarm at the same time, and the information display screen shows "Overweight, prohibited from passing", which improves the safety of road traffic and the accuracy of overloading control work, and reduces traffic accidents and traffic jams caused by height limit violations of large trucks.

[0120] The structural schematic diagram of the folding height limit machine is as Figure 3 shown. The folding height limit machine includes a support column, a height limit bar, a height limit sign, an information display screen, a sound and light alarm, and a height limit machine main box body; the support column plays a decorative role; the height limit bar plays a role in restricting the lane height; the height limit sign is used to prompt passing vehicles that vehicles with a lane limit lower than this height can pass; the information displayed on the information display screen can be set, such as "Height limit 2.5 meters" or "Height limit 2.8 meters", and "Prohibited from passing" is displayed during the height limit conversion process; the sound and light alarm plays a role of sound and light alarm during the height limit conversion process to prompt vehicles passing through the lane to pay attention when passing; the height limit machine main box body is installed with a height limit machine core, and the height limit machine core is composed of a motor, a linkage mechanism, and a height limit machine controller. The controller receives the information from the overloading control system, controls the operation of the motor, and the motor drives the linkage mechanism to realize the lifting and lowering of the height limit bar; when the height limit bar is lifted, the height limit bar rotates from vertically downward to horizontally, and the folding height limit machine is in the height limit state, and only vehicles below 2.5 or 2.8 meters can pass normally in the lane. When the height limit bar is lowered, the height limit bar rotates from horizontally to vertically downward, and the folding height limit machine is in the non-height limit state, and vehicles of various types can pass through the lane.

[0121] During the implementation of the overloading control strategy, real-time feedback data on the implementation effect is collected, such as the response of target vehicles and changes in traffic flow. These feedback data on the effect are compared and analyzed with the expected effect. If it is found that the actual effect does not match the expectation, the overloading control strategy is adjusted in a timely manner to form a closed-loop control, continuously optimizing the overloading control process, so that the overloading control strategy can better adapt to the ever-changing actual situation and improve the accuracy and effectiveness of overloading control.

[0122] The self-detection module is used to monitor the operating parameters of each component of the folding height limiter during the overloading control process to identify the fault modes of the folding height limiter and trigger feedback information.

[0123] The triggering logic of the feedback information includes:

[0124] Monitor the operating parameters of each component of the folding height limiter during the overloading control process;

[0125] Judge whether the folding height limiter has a fault by comparing the operating parameters of each component with thresholds, and identify the fault modes of the folding height limiter;

[0126] Generate feedback information according to the fault modes of the folding height limiter.

[0127] Specifically, the monitoring sub-logic of the operating parameters includes:

[0128] Construct the deployment layout of the sensor network based on the mechanical structure of the folding height limiter and the environmental data of the detection area;

[0129] Determine the state space, action space and reward function of the sensor nodes;

[0130] Train the sensor nodes through the reinforcement learning algorithm, gradually optimize to obtain the deployment scheme of the sensor network, and monitor the operating parameters of each component of the folding height limiter through the sensor network according to the deployment scheme of the sensor network.

[0131] Precisely model the mechanical structure of the folding height limiter, including the positions, dimensions and connection relationships of each component such as the core motor and the linkage mechanism. At the same time, simulate the operating environment of the folding height limiter to determine the distribution of physical fields such as the temperature field and the electromagnetic field. Use this information as the environmental input of the reinforcement learning algorithm to determine the deployment layout of the sensor network, where each sensor network represents a sensor node.

[0132] Determine the state space of the deployment layout of the sensor network, including information such as the position, monitoring direction and coverage range of each sensor node. For example, use parameters such as the coordinate position and angular direction of each sensor node as state variables to form a multi-dimensional state space.

[0133] Define the actions that a sensor node can perform, such as moving, rotating, turning on or off the monitoring function, etc. Each action corresponds to a change in the state space of the sensor node. For example, the moving action can change the coordinate position of the sensor node, and the rotating action can change the monitoring direction of the sensor node.

[0134] Taking the comprehensiveness, accuracy, and real-time nature of the operating parameter acquisition as the optimization goal, design a reward function. For example, when the operating parameters monitored by the sensor can more comprehensively reflect the operating states of the various components of the folding height limiter, or when the data accuracy improves and the transmission delay decreases, a higher reward is given; conversely, a lower reward is given.

[0135] Apply the reinforcement learning algorithm to enable the sensor node (i.e., the agent) to continuously explore and learn in the environment. Through interaction with the environment, the agent adjusts its behavior strategy according to the reward feedback and gradually finds the optimal deployment plan for the sensor network, achieving the best monitoring effect of the operating parameters under the constraints of the mechanical structure, operating environment, and transmission reliability of the operating parameters of the folding height limiter.

[0136] According to the design specifications, historical operating data, and industry standards of the various components of the folding height limiter, set the threshold range for normal operation for each operating parameter. For example, set the normal operating current range for the motor. When the current exceeds this range, it means that there are faults such as overload or short circuit in the motor; by comparing the real-time monitored operating parameters with the set threshold range for normal operation, once a certain operating parameter exceeds the threshold range, immediately trigger a preliminary abnormal alarm to provide a signal for subsequent in-depth fault analysis.

[0137] When a certain operating parameter exceeds the threshold range, perform in-depth analysis on the abnormal operating parameter through the convolutional neural network in deep learning. For example, by analyzing the abnormal change combinations of the motor current, voltage, and speed, determine specific fault modes such as winding short circuit, bearing failure, or power supply problems in the motor; for the connecting rod mechanism, based on the abnormal conditions of its force, displacement, and motion trajectory, identify whether there are faults such as loose connections, component wear, or jamming, and for the sensor monitoring the operating parameters, determine whether the abnormal output is caused by signal interference, component aging, or hardware damage.

[0138] Generate detailed feedback information according to the identified fault mode. The feedback information includes the name of the faulty component, the description of the fault type, the time of fault occurrence, the impact of the fault, etc. For example, if it is determined that the motor of the height limiter mechanism has a winding short - circuit fault, the feedback information will clearly state that "the motor of the height limiter mechanism, winding short - circuit fault, occurred at 12 noon, which will cause the height - limiting rod to fail to lift and lower normally, affecting the over - load control work"; classify and code the feedback information for subsequent information transmission and processing. For example, classify the faults into emergency faults, important faults, and general faults according to the severity of the faults, and assign different codes for quick identification and response.

[0139] The control and management module is used to control the folding height limiter according to the feedback information and synchronously perform the power supply management of the folding height limiter.

[0140] After receiving the feedback information, judge the fault mode of the folding height limiter and feedback it to the remote management terminal, so that the management personnel can understand the operation status of the folding height limiter in real time, in order to troubleshoot the faults in time and remotely control the lifting and lowering operation of the height - limiting rod.

[0141] Specifically, the logic of power supply management includes:

[0142] During the lifting and lowering process of the height - limiting rod, recover the mechanical energy of the height - limiting rod and convert it into electrical energy, and store it in the energy storage device.

[0143] Real - time monitor the power supply status of the folding height limiter, and determine the abnormal situation of the power supply status through threshold judgment.

[0144] When it is determined that the power supply status of the folding height limiter is abnormal, switch the energy storage device to perform emergency power supply.

[0145] During the lifting and lowering process of the height - limiting rod, the energy recovery device installed at key parts such as the connecting rod mechanism starts to work. When the height - limiting rod descends, using hydraulic energy storage technology, convert the mechanical energy of the height - limiting rod into electrical energy. For example, install a small generator at the rotating joint of the connecting rod to convert the rotational mechanical energy of the connecting rod into electrical energy and store it in a dedicated energy storage device.

[0146] Use devices such as voltage sensors and current sensors to real - time monitor the power supply status of the folding height limiter, including power supply parameters such as voltage stability and current magnitude, and at the same time monitor information such as the power and health status of the backup power supply, as well as the energy storage situation of the energy recovery device.

[0147] Set the normal power supply parameter threshold. Once the power supply parameters exceed the threshold range, such as too low or too high voltage, abnormal current fluctuation, immediately trigger a warning signal to prepare for the emergency power supply switch.

[0148] If the backup power supply and the energy recovery device meet the switching conditions, the energy recovery device is preferentially started for power supply. If the energy recovery device has insufficient power or cannot work properly, the backup power supply is enabled. If there are problems such as faults or insufficient power in the backup power supply, an alarm message is sent to the remote management terminal in a timely manner to prompt the maintenance personnel to handle it as soon as possible.

[0149] Embodiment 2

[0150] As Figure 4 shown, the embodiment of the present application provides a method flowchart of a self-detection and over-limit control method for a folding height limiter. The method includes:

[0151] S1. Obtain the perception data of the target vehicle in the detection area through the sensor network, and perform data association and alignment on the perception data of the target vehicle;

[0152] S2. Receive the perception data of the target vehicle, obtain the road slope in the detection area, extract the features of the perception data as contour features and state features, extract the feature of the road slope as the road height difference, perform correlation analysis on the road height difference with the contour features and state features, and obtain the analysis result to dynamically adjust the over-limit control strategy of the folding height limiter;

[0153] S3. Monitor the operating parameters of each component of the folding height limiter during the over-limit control process to identify the fault mode of the folding height limiter and trigger feedback information;

[0154] S4. Control the folding height limiter according to the feedback information and synchronously execute the power supply management of the folding height limiter.

[0155] Since the principle of solving problems by the method in the embodiment of the present application is similar to that of the above-mentioned system in the embodiment of the present application, the implementation of the method can refer to the implementation of the system, and the repeated parts will not be described again.

Claims

1. A foldable height limiter self-detection and overload control system, characterized in that: include: Perception fusion module, overload control module, self-detection module and control management module; The perception fusion module is used to obtain the perception data of the target vehicle in the detection area through the sensor network, and to associate and align the perception data of the target vehicle; The overload control module is used to receive the perception data of the target vehicle, obtain the road slope in the detection area, extract the characteristics of the perception data as contour characteristics and state characteristics, extract the characteristics of the road slope as road height difference, perform correlation analysis on the road height difference and the contour characteristics and state characteristics, and obtain the analysis results to dynamically adjust the overload control strategy of the foldable height limiter; The self-detection module is used to monitor the operating parameters of various components of the foldable height limiter during the overload control process to identify the failure mode of the foldable height limiter and trigger feedback information; The control management module is used to control the foldable height limiter according to the feedback information and simultaneously perform power supply management of the foldable height limiter.

2. A foldable height limiter self-detection and overload control system as claimed in claim 1, characterized in that: The perception data includes contour data, visual data and state data; The logic of obtaining the perception data of the target vehicle in the detection area includes: Deploy a sensor network according to the detection area of ​​the foldable height limiter, and trigger an acquisition signal when the target vehicle enters the detection area; After receiving the acquisition signal, the sensor network obtains the contour data, visual data and status data of the target vehicle; The contour data, visual data and status data of the target vehicle are associated and aligned to obtain the perception data of the target vehicle.

3. A foldable height limiter self-detection and overload control system as claimed in claim 2, characterized in that: The sub-logic of data association and alignment includes: Add a timestamp and associate the timestamp with the target vehicle's contour data, visual data, and status data for storage; A unified spatial coordinate system is defined based on the detection area of ​​the foldable height limiter, and the contour data, visual data and state data of the target vehicle are converted into the spatial coordinate system and spatially aligned to obtain the perception data of the target vehicle; Verify and correct the perception data that has completed timestamp association and spatial registration alignment.

4. A foldable height limiter self-detection and overload control system as claimed in claim 1, characterized in that: The adjustment logic of the overload control strategy includes: Receive the perception data of the target vehicle, obtain the environmental data of the detection area, and integrate the perception data of the target vehicle and the environmental data of the detection area into vehicle environmental data; Perform data analysis on vehicle environment data and generate analysis results; Generate overload control strategies based on analysis results; Send the control strategy to the foldable height limiter to control the lifting and lowering of the height limiter, the working status of the sound and light alarm, and the prompt content of the information display screen; Collect feedback data on the effects of overweight vehicle control in the process to dynamically adjust the overweight vehicle control strategy.

5. A foldable height limiter self-detection and overload control system as claimed in claim 4, characterized in that: The generation sub-logic of the analysis result includes: Extracting environmental features of environmental data of a detection area from vehicle environmental data, the environmental features including road height difference; Extracting contour features and state features of the perception data of the target vehicle in the vehicle environment data; Based on graph neural network, the correlation analysis between road height difference, contour features and state features is carried out; Quantify vehicle risk based on the results of association analysis.

6. A foldable height limiter self-detection and overload control system as claimed in claim 5, characterized in that: The sub-logic of the association analysis includes: The road slope of the detection area and the perception data of the target vehicle are used as the nodes of the graph to construct the initial graph structure; The road height difference, contour features and state features are fused through the attention mechanism to obtain fused features, and the fused features are input into the graph neural network to obtain the association results; The causal relationship between the road slope in the detection area and the behavior of the target vehicle is mined through the causal inference algorithm to obtain the causal inference result; The association results and causal inference results are verified to generate the results of the association analysis.

7. A foldable height limiter self-detection and overload control system as claimed in claim 1, characterized in that: The triggering logic of the feedback information includes: Monitor the operating parameters of each component of the foldable height limiter during the overload control process; By comparing the operating parameters of each component with a threshold value, it is determined whether the foldable height limiter fails, and the failure mode of the foldable height limiter is identified; Feedback information is generated based on the failure mode of the foldable height limiter.

8. A foldable height limiter self-detection and overload control system as claimed in claim 7, characterized in that: The monitoring sub-logic of the operating parameters includes: The deployment layout of the sensor network is constructed based on the mechanical structure of the foldable height limiter and the environmental data of the detection area; Determine the state space, action space and reward function of the sensor node; The sensor nodes are trained by reinforcement learning algorithm, and the deployment plan of the sensor network is gradually optimized. According to the deployment plan of the sensor network, the operating parameters of each component of the foldable height limiter are monitored through the sensor network.

9. A foldable height limiter self-detection and overload control system as claimed in claim 8, characterized in that: The deployment logic of the sensor network includes: Determine a combination of at least two laser radars, three cameras, and two millimeter wave radars to obtain contour data, visual data, and status data of the target vehicle; Plan the installation positions and angles of the laser radar, camera, and millimeter-wave radar based on the actual installation environment of the foldable height limiter and the road slope in the detection area; Adaptively adjust the working parameters of lidar, camera and millimeter-wave radar.

10. A self-detection and overload control method for a foldable height limiter, implemented based on a self-detection and overload control system for a foldable height limiter according to any one of claims 1 to 9, characterized in that: include: S1. Acquire the perception data of the target vehicle in the detection area through the sensor network, and associate and align the perception data of the target vehicle; S2. Receive the perception data of the target vehicle and obtain the road slope in the detection area, extract the characteristics of the perception data as contour characteristics and state characteristics, extract the characteristics of the road slope as road height difference, perform correlation analysis on the road height difference and the contour characteristics and state characteristics, and obtain the analysis results to dynamically adjust the overload control strategy of the foldable height limiter; S3. monitoring the operating parameters of various components of the foldable height limiter during the overload control process to identify the failure mode of the foldable height limiter and trigger feedback information; S4. Operate the foldable height limiter according to the feedback information, and simultaneously perform power supply management of the foldable height limiter.

Images