Urban traffic management system and method based on intelligent sensor network driving

Abstract

The invention discloses an urban traffic management system and method based on intelligent sensor network driving, and particularly relates to the technical field of urban traffic management, and the method comprises the following steps: collecting traffic data in real time through intelligent sensor network nodes deployed on urban roads, generating a road segment sensor network node use information set, related features are extracted for analysis; inputting an analysis result into a trained machine learning model, calculating a road section data quality coefficient, and judging whether a secondary decision is needed or not; in a secondary decision-making stage, combining a feature analysis result and a node use proportion, and inputting a second model to judge an optimization direction type; according to the optimization direction, the use condition of the sensing network nodes of the road section is improved, and the traffic data quality of the intelligent sensing network nodes is ensured; according to the invention, through continuous optimization of the sensor network, the stability of urban traffic management can be ensured, the representativeness and accuracy of traffic data are ensured, and a guarantee is provided for the decision-making capability of an intelligent traffic management system.

Description

Technical Field

[0001] The present invention relates to the technical field of urban traffic management, and more specifically, to an urban traffic management system and method driven by an intelligent sensor network. Background Art

[0002] With the acceleration of urbanization, urban traffic problems have become increasingly prominent. Traffic jams, road congestion, frequent traffic accidents and other problems have seriously affected people's daily life and work efficiency, and even restricted the sustainable development of cities. The traditional urban traffic management model mainly relies on manual scheduling and fixed traffic light control, which often has the disadvantages of low efficiency, slow response, and inability to adjust in real time, and cannot effectively respond to dynamic changes in traffic flow and emergencies.

[0003] In order to improve traffic management efficiency, reduce traffic congestion, and reduce energy consumption and environmental pollution, more and more cities are beginning to try to adopt intelligent traffic management systems. These systems achieve the goal of optimizing traffic flow and reducing traffic accidents by collecting real-time traffic data, analyzing traffic flow, adjusting signal timing, and performing traffic dispatch. However, existing intelligent traffic management systems still face some problems, such as incomplete data acquisition, inflexible response, inability to effectively process large-scale dynamic data, and inefficient coordination between devices and systems.

[0004] The emergence of intelligent sensor networks provides a new approach to solving these problems. Intelligent sensor networks collect information such as traffic flow, vehicle speed, and traffic status in real time through a large number of distributed sensors, and upload the data to the cloud platform or central control system through wireless communication technology, thus realizing real-time monitoring and management of traffic conditions. In addition, combined with big data analysis and artificial intelligence technology, it can better predict traffic flow, identify potential traffic bottlenecks and accident risks, and thus improve the accuracy and efficiency of traffic dispatch. However, there are still some bottleneck problems: if the sensor layout is not dense enough or the distribution is unreasonable, it may lead to incomplete monitoring data, especially in remote areas or complex traffic scenes (such as viaducts and underground tunnels), and the quality of the data source for urban traffic management cannot be guaranteed. Therefore, an urban traffic management system and method driven by intelligent sensor networks are proposed here to solve the above problems. Summary of the invention

[0005] To achieve the above object, the present invention provides the following technical solutions: The urban traffic management method based on intelligent sensor network driving includes the following steps: By deploying intelligent sensor network nodes on urban roads, traffic data is collected in real time, and the usage of sensor network nodes in different sections of urban roads is obtained to generate a section sensor network node usage information set; Based on the usage information set of the road section sensor network nodes, the usage-related features of the sensor network nodes are extracted and feature analysis is performed. The feature analysis results are then input into a pre-trained machine learning model to generate a road section data quality coefficient, and then decide whether to make a secondary decision. When making a secondary decision, the feature analysis results and the usage ratio of the sensor network nodes in the road section are substituted into the pre-trained machine learning model 2 to determine the type of optimization direction; Based on the type of optimized direction, the usage of sensor network nodes in road sections is improved to ensure the quality of real-time traffic data collected by intelligent sensor network nodes on urban roads.

[0006] In a preferred embodiment, extracting sensor network node usage-related features based on the road segment sensor network node usage information set refers to: The effectiveness and coverage of the current deployment scheme of the road section sensor network nodes are respectively obtained to obtain a sensor coverage feature data group and a sensor deployment feature data group.

[0007] In a preferred embodiment, when performing feature analysis, a sensor coverage index is generated based on the sensor coverage feature data set, and a layout uniformity index is generated based on the sensor layout feature data set.

[0008] In a preferred embodiment, the logic for obtaining the sensor coverage index is: First calculate the effective coverage area of ​​a single sensor in use: ; Indicates the sensing radius of the i-th sensor in use, reflecting the coverage of the sensor. It represents the environmental adaptability factor, which reflects the coverage capability of the i-th sensor in use in the actual environment, and its value range is from zero to one. represents the effective coverage area of ​​the i-th sensor in use; Then calculate the priority weights for the coverage areas: ; represents the importance score of the area covered by the i-th sensor in use, N represents the number of sensors in use, j is the sensor number index in use, represents the importance score of the jth sensor coverage area in use, represents the priority weight of the i-th sensor in use; Correct the duplicate coverage contribution value: ; represents the overlapping area of ​​the coverage area of ​​the i-th and j-th sensors in use, Indicates that the overlap area of ​​the i-th and j-th sensors in use corresponds to a preset proportionality coefficient; Divide the road section where the sensor network nodes are deployed into M sub-areas, calculate the sum of the monitoring areas of the M sub-areas, and get the total monitoring area ; Finally calculate the sensor coverage index: ; Represents the sensor coverage index of all sensors in use.

[0009] In a preferred embodiment, the environmental adaptability factor The acquisition logic is: Introducing obstacle factors , represents the impact of obstacles on the coverage area of ​​the i-th sensor in use, which is obtained by calculating the ratio of the obstacle area to the coverage area; Introducing temperature factor , represents the current ambient temperature of the i-th sensor in use; Introducing the light factor , represents the current light intensity of the i-th sensor in use state; Substitute into the following calculation formula: ; Indicates the preset maximum obstacle density. Indicates the preset maximum temperature deviation range, Indicates the optimal operating temperature of the sensor. Indicates the optimal working light intensity of the sensor. It represents the influence weight of the kth environmental factor on the adaptability factor, and the values ​​of k are 1, 2, and 3.

[0010] In a preferred embodiment, the logic for obtaining the layout uniformity index is: The road section where the sensor network nodes are deployed is divided into M sub-areas, and the sub-areas are numbered as r. The standard deviation of the distance between the sensors in use in the rth sub-area is calculated first. , and then calculate the average distance between sensors in use in the rth sub-area ; According to the coverage deviation degree of the sensors in use in the rth sub-area, the coverage penalty coefficient is introduced to reflect the matching between the sensor layout and the actual demand. : ; The deployment capacity of the sensors in use in the r-th sub-area is calculated by calculating the sum of the effective coverage areas of the sensors in use in the r-th sub-area and then subtracting the total overlapping area in the sub-area; The ideal deployment capacity of the sensors in the r-th sub-area in use is the sum of the coverage areas of the sensors in the r-th sub-area in use under ideal conditions without obstacles and interference; The calculation formula of layout uniformity index is: ; Indicates the uniformity index of all sensors in use.

[0011] In a preferred embodiment, the feature analysis result is input into a pre-trained machine learning model 1 to generate a road section data quality coefficient, and then deciding whether to make a secondary decision refers to: The sensor coverage index and the layout uniformity index are input into a pre-trained machine learning model. The output result is the road section data quality coefficient. The road section data quality coefficient is compared with the preset quality threshold. If the road section data quality coefficient is greater than or equal to the preset quality threshold, no secondary decision is made. If the road section data quality coefficient is less than the preset quality threshold, a secondary decision is made.

[0012] In a preferred embodiment, when making a secondary decision, fuzzy logic is used to take the sensor coverage index, the layout uniformity index, and the sensor network node usage ratio of the road section as input data, and the type of optimization direction is taken as the output data of the fuzzy logic.

[0013] In a preferred embodiment, the urban traffic management system driven by the intelligent sensor network includes: The data collection module is composed of a data collection unit 1 and a data collection unit 2. The data collection unit 1 collects traffic data in real time through intelligent sensor network nodes deployed on urban roads. At the same time, the data collection unit 2 obtains the usage of sensor network nodes in different sections of urban roads and generates a section sensor network node usage information set. A feature extraction module extracts features related to the usage of sensor network nodes based on the usage information set of the road segment sensor network nodes; The feature analysis module performs feature analysis operations based on the features related to the usage of the sensor network nodes to obtain feature analysis results; The primary decision module inputs the feature analysis results into the pre-trained machine learning model 1 to generate the road section data quality coefficient, and then decides whether to make a secondary decision; The secondary decision module, when making a secondary decision, substitutes the feature analysis results and the usage ratio of the sensor network nodes of the road section into the pre-trained machine learning model 2 to determine the type of optimization direction; The quality assurance module improves the usage of sensor network nodes on road sections based on the type of optimization direction, ensuring the quality of real-time traffic data collected by intelligent sensor network nodes on urban roads.

[0014] Technical effects and advantages of the present invention: The present invention deploys intelligent sensor network nodes in urban roads and collects traffic data (such as vehicle flow, speed, congestion, etc.) in real time. This method can provide timely and accurate data support in urban traffic management. These data are of great significance for optimizing traffic flow, reducing congestion, and improving road utilization efficiency. By acquiring traffic data of road sections in real time, a more effective decision-making basis can be provided for the traffic management system. By performing feature analysis on the information set used by the road section sensor network nodes and combining the machine learning model to quantitatively evaluate the coverage quality and layout uniformity of the sensors, the quality of the network layout can be effectively identified and evaluated. Through the generated "road section data quality coefficient", the system can determine whether there are problems with the current layout and make timely adjustments and optimizations. This dynamic evaluation and optimization process can promptly discover problems during the sensor layout process, thereby improving the effectiveness of the network and the quality of data collection.

[0015] The convolutional neural network (CNN) model analyzes the coverage index and layout uniformity index of the sensor, generates a data quality coefficient and determines whether a secondary decision is needed, which can automatically and efficiently evaluate the layout quality of the current network. If the data quality coefficient is lower than the set threshold, the system will automatically make a secondary decision, combine the fuzzy reasoning model for further analysis and give specific optimization directions (such as adding sensors, adjusting the layout location, etc.), so as to ensure that the quality of traffic data collection is always in the best state.

[0016] Through the secondary decision-making stage of the fuzzy reasoning model, it is possible to determine whether the sensor layout is uniform based on the sensor coverage index, layout uniformity index and sensor usage ratio, and provide specific optimization directions for improving the layout strategy. For example, if it is found that the number of sensors is insufficient or the layout is uneven, the system will recommend increasing the number of sensors, adjusting the positions of existing sensors, and even using higher-precision sensors to replace some low-quality sensors when necessary. Through this optimization, the system can achieve reasonable allocation of resources, avoid over-concentration or over-redundant layout, and improve overall network efficiency.

[0017] The "environmental adaptability factor" used in the present invention dynamically adjusts the effective coverage area of ​​the sensor and can compensate for the sensor performance according to the actual environment (such as obstacles, weather, temperature and other factors). This dynamic adjustment based on the actual environment can make the traffic management system more adaptable, not be disturbed by changes in the external environment, and ensure the reliability of data collection.

[0018] By continuously optimizing the sensor network, the present invention can ensure the long-term stable and efficient operation of the urban traffic management system. As the urban traffic conditions change, the system can flexibly adjust the sensor deployment plan, promptly solve the blind spots or excessive overlap problems of sensor data collection, ensure the representativeness and accuracy of traffic data, and further enhance the decision-making ability of the intelligent traffic management system.

[0019] By combining machine learning models with fuzzy reasoning models, this method can not only automatically evaluate and optimize sensor networks, but also dynamically adjust deployment strategies based on actual traffic needs. Through automated evaluation and decision support, manual intervention and judgment are reduced, and the adaptive ability of sensor networks is improved. With the improvement of sensor deployment quality and the optimization of network performance, the maintenance cost of the overall traffic management system has also been effectively reduced. Through reasonable deployment and refined management, urban traffic management departments can maintain and manage traffic sensor networks more economically and efficiently. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to facilitate understanding by those skilled in the art, the present invention is further described below in conjunction with the accompanying drawings; Figure 1 This is a schematic diagram of the urban traffic management method based on intelligent sensor network drive in the present invention.

[0021] Figure 2 This is a schematic diagram of the urban traffic management system driven by an intelligent sensor network in the present invention. DETAILED DESCRIPTION

[0022] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0023] Reference Figure 1 - Figure 2 The following embodiments are obtained: Example 1: Urban traffic management driven by intelligent sensor networks is a management method that uses intelligent sensor network (ISN) technology to monitor, collect and analyze urban traffic operation conditions in real time to optimize traffic resource allocation, improve traffic efficiency, reduce congestion and improve safety.

[0024] Intelligent Sensor Network (ISN): A network consisting of multiple distributed sensor nodes that work together through wireless communication technology to monitor data in the physical environment (such as traffic flow, vehicle speed, pollution concentration, etc.) in real time.

[0025] Urban traffic management: Through data collection, analysis and control measures, urban traffic operations are planned, regulated and optimized to achieve efficient, safe and sustainable traffic operations.

[0026] The core of urban traffic management driven by intelligent sensor networks is to use sensor nodes to monitor traffic conditions (such as traffic flow, vehicle speed, vehicle type, etc.), transmit real-time data to the traffic management system through the network, analyze the data and generate optimization strategies (such as signal light control, traffic diversion scheme, etc.) through the existing traffic management system, and improve real-time traffic flow through feedback mechanisms.

[0027] For example, sensors are deployed at key road nodes to obtain real-time information such as traffic flow, average vehicle speed, and road occupancy rate. The data is transmitted to the cloud or edge computing nodes through wireless communication technology (such as Wi-Fi, 5G). However, there are still some urgent problems that need to be solved, such as insufficient density or unreasonable distribution of sensors, which may lead to incomplete monitoring data, especially in remote areas or complex traffic scenarios (such as viaducts and underground tunnels). There is a lack of dynamic deployment capabilities. When environmental conditions change, it is impossible to quickly determine the optimization direction of the data collection plan for the monitoring area. The use of fixed sensors may also result in waste of resources or fail to ensure the quality of traffic data collection.

[0028] Therefore, the present invention proposes an urban traffic management method based on intelligent sensor network driving, comprising the following steps: The intelligent sensor network nodes deployed on urban roads collect traffic data in real time, and at the same time obtain the usage of sensor network nodes in different sections of urban roads to generate a section sensor network node usage information set; the purpose of this step is to collect traffic data in real time by deploying intelligent sensor network nodes on urban roads, such as vehicle flow, speed, congestion, etc. At the same time, the usage status of each section sensor node (such as whether it is working, etc.) is monitored and recorded, forming a section sensor network node usage information set to provide basic data for subsequent analysis.

[0029] Based on the information set of the road section sensor network node usage, the relevant features of the sensor network node usage are extracted, and the feature analysis is performed. Then, the feature analysis results are input into the pre-trained machine learning model 1 to generate the road section data quality coefficient, and then decide whether to make a secondary decision; in this step, the features related to the sensor network node usage are extracted from the data collected in the first step, and the quality of the traffic data is evaluated by analyzing these features. The analysis results will be input into a pre-trained machine learning model. The goal of the model is to generate a "road section data quality coefficient" to represent the data quality level of the road section. According to this coefficient, it can be judged whether further measures need to be taken (such as whether a secondary decision needs to be made). The machine learning model is a convolutional neural network model, which outputs the road section data quality coefficient. It can also be trained by outputting 0 or 1 when deciding whether to make a secondary decision. The result is the layout quality classification: high-quality layout (1) or low-quality layout (0). When the layout is low-quality (0), there are defects in the layout and it needs to be optimized.

[0030] When making a secondary decision, the feature analysis results and the usage ratio of the sensor network nodes in the road section are substituted into the pre-trained machine learning model 2 to determine the type of optimization direction; if the data quality coefficient shows that the data quality is poor or needs to be improved, it will enter the secondary decision stage. In this stage, the feature analysis results and the usage ratio of the sensor network nodes in the current road section (the ratio of the number of active nodes to the total nodes) are passed as input to another pre-trained machine learning model, which will determine the optimization direction of the road section based on the input data.

[0031] The second machine learning model is a fuzzy inference model. According to the ratio of the number of sensors currently in use to the total number of sensors planned for the road section, the following strategies are adopted: Insufficient number of sensors (low coverage + high uniformity): If the sensor coverage index (SCI) is low and the deployment uniformity index (SDUI) is high, the current number of sensors is insufficient.

[0032] Optimization measures: Increase the number of sensors and prioritize their deployment in blind spots and key traffic areas (such as congested roads and intersections).

[0033] Uneven deployment (high coverage + low uniformity): If the sensor deployment uniformity index (SDUI) is low but the coverage index (SCI) is high, the sensors are unevenly deployed. Optimization measures: Adjust the position of existing sensors and re-deploy them to fill in uncovered areas to avoid over-concentration in certain areas and waste of resources.

[0034] Insufficient quantity and uniformity (low coverage + low uniformity): If both indexes are low, the number of sensors deployed is insufficient and the distribution is unreasonable. Optimization measures: Increase the number of sensors and optimize the deployment strategy at the same time, give priority to increasing coverage in key areas, and evenly deploy them in secondary areas.

[0035] Redundant quantity but poor effect (high quantity ratio + low quality): If the Ratio is close to 1 (the number of sensors is close to the planned value), but the deployment is still low-quality, there may be a waste of resources or an error in the deployment strategy. Optimization measures: Consider replacing some existing sensors with higher-precision sensors.

[0036] Based on the type of optimization direction, the usage of sensor network nodes on the road section is improved. The improvement measures have been explained in the above content. To ensure the quality of real-time traffic data collected by intelligent sensor network nodes on urban roads, it means that the layout quality classification result is high-quality layout (1). In this way, the collected traffic data can ensure that it has sufficient data collection quality in the current application environment. In the future, the traffic management system in the existing technology can better analyze the data and generate optimization strategies (such as traffic light control, traffic diversion solutions, etc.).

[0037] Extracting the relevant features of the sensor network node usage based on the road section sensor network node usage information set means: respectively obtaining the effectiveness and coverage of the current deployment scheme of the road section sensor network nodes, and obtaining the sensor coverage feature data group and the sensor deployment feature data group. When performing feature analysis, a sensor coverage index is generated based on the sensor coverage feature data group, and a deployment uniformity index is generated based on the sensor deployment feature data group. Sensor coverage feature data group: This data group reflects the characteristics of the road section sensor network in terms of regional coverage, and mainly focuses on whether the sensor effectively covers all areas that need to be monitored. For example, whether the coverage range is uniform, whether there are blind spots, whether the distribution of sensors is reasonable, etc. Sensor deployment feature data group: This data group focuses on the distribution characteristics of sensors during the deployment process, including the density, spacing, and uniformity of the deployment of sensors. This data group can be used to evaluate whether the spatial optimization is considered when the sensors are deployed, and whether there is an over-dense or over-sparse situation. By extracting the coverage and deployment features of sensors, the design of the existing sensor network can be optimized to ensure the effective coverage and reasonable deployment of the network. Specifically, the coverage index and layout uniformity index help identify the weak links in the network and make targeted adjustments and optimizations.

[0038] The logic for obtaining the sensor coverage index is: First calculate the effective coverage area of ​​a single sensor in use: ; Indicates the sensing radius of the i-th sensor in use, reflecting the coverage of the sensor. It represents the environmental adaptability factor, which reflects the coverage capability of the i-th sensor in use in the actual environment (such as obstacles and weather conditions). The value range is from zero to one. Bad weather may reduce the perception capability, and the value of the environmental adaptability factor will decrease. represents the effective coverage area of ​​the i-th sensor in use; Then calculate the priority weights for the coverage areas: ; represents the importance score of the area covered by the i-th sensor in use, which depends on the area type, such as: a high score for a busy traffic area, a low score for a remote area or a secondary area, N represents the number of sensors in use, j is the index of the sensor number in use, represents the importance score of the jth sensor coverage area in use, represents the priority weight of the i-th sensor in use; When multiple sensors cover the same monitoring point, it is necessary to correct the impact of repeated coverage on the overall coverage index. The calculation formula for corrected repeated coverage contribution value is: ; represents the overlapping area of ​​the coverage area of ​​the i-th and j-th sensors in use, Indicates that the overlap area of ​​the i-th and j-th sensors in use corresponds to the preset proportionality coefficient, which measures the actual impact of repeated coverage on monitoring accuracy (for example, repeated coverage of key areas may be given a higher weight). represents the duplicate coverage contribution value; Divide the road section where the sensor network nodes are deployed into M sub-areas, calculate the sum of the monitoring areas of the M sub-areas, and get the total monitoring area ; Finally, the effective coverage area, priority weight, and repeated coverage contribution of each sensor are summarized and normalized to a ratio relative to the total area of ​​the monitoring area to calculate the sensor coverage index: ; Represents the sensor coverage index of all sensors in use. The effective coverage area represents the effective coverage capability of each sensor in the actual environment. The introduction of environmental adaptability factors makes the coverage area closer to reality. Taking into account the performance changes of sensors under obstacles, weather, etc., the priority weight reflects the importance of different monitoring areas, ensuring that key areas (such as transportation hubs and accident-prone areas) account for a larger proportion in the coverage index calculation. It indicates the contribution value of repeated coverage, avoids overestimation caused by repeated coverage, and emphasizes the efficiency of sensor deployment rather than the simple accumulation of quantity. The total monitoring area provides a normalized benchmark to ensure that the coverage index range is within a comparable numerical range, which is convenient for comparison and evaluation.

[0039] The environmental adaptability factor reflects the coverage capability of the sensor in the actual operating environment. It is affected by environmental conditions such as obstacles and lighting. The calculation formula should take these factors into consideration and normalize them based on the sensor performance model. The acquisition logic is: Introducing obstacle factors , represents the impact of obstacles on the coverage area of ​​the i-th sensor in use, that is, the density of obstacles (such as buildings, trees, etc.), which is obtained by calculating the ratio of the obstacle area to the coverage area; Introducing temperature factor , represents the current ambient temperature of the i-th sensor in use; Introducing the light factor , represents the current light intensity of the i-th sensor in use state; Substitute into the following calculation formula: ; Indicates the preset maximum obstacle density. Indicates the preset maximum temperature deviation range, Indicates the optimal operating temperature of the sensor. Indicates the optimal working light intensity of the sensor. If the temperature and light conditions in the environment deviate from the optimal working conditions, the performance will degrade more. It represents the influence weight of the kth environmental factor on the adaptability factor, k takes values ​​of 1, 2, and 3, and the quadratic term Used to emphasize the nonlinear effect of temperature deviation on performance, It provides flexible adjustment space and can adjust the importance of different factors according to the actual environmental needs. When the environment changes (such as increased obstacles and drastic temperature fluctuations), the formula can reflect the dynamic impact on sensor performance in real time.

[0040] The logic for obtaining the layout uniformity index is: The road sections where the sensor network nodes are deployed are divided into M sub-areas for comprehensive calculation of the deployment uniformity of all road sections. The larger M is, the finer the monitoring area division is. The sub-areas are numbered as r. The standard deviation of the distance between sensors in use in the rth sub-area is first calculated. , measures the fluctuation of the distance between sensors in use in the rth sub-area. The smaller the value, the more uniform the distribution of sensor spacing in the road section. Then calculate the average distance between sensors in use in the rth sub-area , used to reflect the overall spacing level of sensors in use in the rth sub-area, serving as a benchmark for uniformity calculation; According to the coverage deviation degree of the sensors in use in the rth sub-area, the coverage penalty coefficient is introduced to reflect the matching between the sensor layout and the actual demand. : ; The deployment capacity of the sensors in use in the r-th sub-area is calculated by calculating the sum of the effective coverage areas of the sensors in use in the r-th sub-area and then subtracting the total overlapping area in the sub-area; The ideal deployment capacity of the sensors in the rth sub-area in use is the sum of the coverage areas of the sensors in the rth sub-area in use under ideal conditions without obstacles and interference. The closer it is to 0, the closer the road coverage is to ideal.

[0041] The calculation formula of layout uniformity index is: ; Represents the layout uniformity index of all sensors in use. The closer the value is to 1, the higher the layout uniformity.

[0042] Inputting the feature analysis results into the pre-trained machine learning model 1 to generate the road section data quality coefficient, and then deciding whether to make a secondary decision means: The sensor coverage index and the layout uniformity index are input into a pre-trained machine learning model. The output result is the road section data quality coefficient. The road section data quality coefficient is compared with the preset quality threshold. If the road section data quality coefficient is greater than or equal to the preset quality threshold, no secondary decision is made. If the road section data quality coefficient is less than the preset quality threshold, a secondary decision is made.

[0043] Sensor coverage index: This index indicates the coverage of sensors on the road section, which may be sensor density, coverage range, or coverage uniformity. This indicator can reflect whether the data collection on the road section is sufficient and representative. Deployment uniformity index: This index measures the uniformity of sensor deployment on the road section, that is, whether the sensors are deployed too densely or too sparsely. Uniform deployment helps improve data quality and avoid data gaps or imbalances. These input features will be passed into a convolutional neural network model (CNN), which is used to process the input features, extract features and output a road section data quality coefficient to represent the data quality level of the road section.

[0044] The road section data quality coefficient is the output of the convolutional neural network, which is used to quantify the data quality of the road section. A higher value may mean that the sensor layout is reasonable and the data collection is complete, while a lower value may indicate that the layout is unreasonable and there are problems with the data quality. The quality coefficient is compared with the preset quality threshold. If the road section data quality coefficient is greater than or equal to the threshold, it can be considered that the data quality of the road section meets the standard and no secondary decision is required. Otherwise, the data quality of the road section is lower than the preset standard and a secondary decision is required. Secondary decision means further optimization or adjustment after the preliminary assessment that the road section data quality coefficient is low. For example, measures may be taken to increase the density of sensor deployment and adjust the deployment position.

[0045] Layout quality classification (high-quality layout or low-quality layout): Through the training of machine learning models, the road section data quality coefficient can not only measure the data quality, but also provide a classification for the layout quality of the road section (for example: high-quality layout = 1, low-quality layout = 0). Low-quality layout (i.e., low quality coefficient) will indicate the need for further layout optimization. For example, low-quality layout may mean that the layout of sensors is unreasonable, there are blind spots in data collection or excessive overlap, which affects the accuracy and representativeness of the data. Further analysis and optimization of sensor layout is needed to ensure that the data coverage on the road section is complete and uniform.

[0046] During the training process, through a large number of labeled data sets (including sensor deployment conditions in different road sections and corresponding quality assessments), the convolutional neural network model can learn to extract key factors affecting data quality from input features such as sensor coverage index and deployment uniformity index. The goal of the training is to enable the model to accurately generate road section data quality coefficients, and to judge the deployment quality based on the coefficients, providing a scientific basis for subsequent optimization.

[0047] When making secondary decisions, fuzzy logic is used to take the sensor coverage index, layout uniformity index, and sensor network node usage ratio of the road section as input data, and the type of optimization direction as the output data of the fuzzy logic.

[0048] Specifically: Input data fuzzification: Convert data such as sensor coverage index (SCI), deployment uniformity index (SDUI) and sensor usage ratio (Ratio) into fuzzy values. For example, SCI may be converted into fuzzy categories such as "low coverage", "medium coverage", and "high coverage", SDUI into "low uniformity", "medium uniformity" and "high uniformity", and Ratio into "low usage ratio", "medium usage ratio" and "high usage ratio". The purpose of fuzzification is to allow a certain degree of uncertainty and ambiguity in the input data without being precise, which is in line with the fact that data is often incomplete or imprecise in reality.

[0049] Rule base: A set of fuzzy rules is designed to describe the relationship between input data and output. For example: If SCI is low coverage and SDUI is high uniformity, the number of sensors is insufficient. If SCI is high coverage and SDUI is low uniformity, the sensors are unevenly distributed. If both SCI and SDUI are low, the number of sensors is insufficient and the distribution is unreasonable. These rules are set based on the experience of domain experts and the specific needs of the problem to help the reasoning system make reasonable decisions.

[0050] Based on the fuzzy rule base, the input fuzzy data is inferred to derive a fuzzy output result. During the inference process, the system will determine which rules are applicable through "rule activation", and then weight and calculate the qualified rules to obtain a fuzzy output. The core of the inference is to generate fuzzy results through the relationship between the combination rules of fuzzy logic and the input data. For example, based on the relationship between low coverage and high uniformity, the optimization suggestion of "insufficient number of sensors" can be derived.

[0051] Convert the result of fuzzy reasoning into a clear decision or numerical output. Defuzzification is to convert the fuzzy result into an actual value or category, such as "increase the number of sensors" or "adjust the sensor position" or "replace a higher performance sensor". Common defuzzification methods include the centroid method, the maximum membership method, etc.

[0052] Embodiment 2: An urban traffic management system driven by an intelligent sensor network, comprising: The data collection module is composed of a data collection unit 1 and a data collection unit 2. The data collection unit 1 collects traffic data in real time through intelligent sensor network nodes deployed on urban roads. At the same time, the data collection unit 2 obtains the usage of sensor network nodes in different sections of urban roads and generates a section sensor network node usage information set. A feature extraction module extracts features related to the usage of sensor network nodes based on the usage information set of the road segment sensor network nodes; The feature analysis module performs feature analysis operations based on the features related to the usage of the sensor network nodes to obtain feature analysis results; The primary decision module inputs the feature analysis results into the pre-trained machine learning model 1 to generate the road section data quality coefficient, and then decides whether to make a secondary decision; The secondary decision module, when making a secondary decision, substitutes the feature analysis results and the usage ratio of the sensor network nodes of the road section into the pre-trained machine learning model 2 to determine the type of optimization direction; The quality assurance module improves the usage of sensor network nodes on road sections based on the type of optimization direction, ensuring the quality of real-time traffic data collected by intelligent sensor network nodes on urban roads.

[0053] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0054] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0055] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0056] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0057] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. An urban traffic management method based on intelligent sensor network driving, characterized in that: The following steps are involved: By deploying intelligent sensor network nodes on urban roads, traffic data is collected in real time, and the usage of sensor network nodes in different sections of urban roads is obtained to generate a section sensor network node usage information set; Based on the usage information set of the road section sensor network nodes, the usage-related features of the sensor network nodes are extracted and feature analysis is performed. The feature analysis results are then input into a pre-trained machine learning model to generate a road section data quality coefficient, and then decide whether to make a secondary decision. When making a secondary decision, the feature analysis results and the usage ratio of the sensor network nodes in the road section are substituted into the pre-trained machine learning model 2 to determine the type of optimization direction; Based on the type of optimized direction, the usage of sensor network nodes in road sections is improved to ensure the quality of real-time traffic data collected by intelligent sensor network nodes on urban roads.

2. The urban traffic management method based on intelligent sensor network drive according to claim 1 is characterized in that: The features related to the usage of sensor network nodes extracted based on the usage information set of the road segment sensor network nodes refer to: The effectiveness and coverage of the current deployment scheme of the road section sensor network nodes are respectively obtained to obtain a sensor coverage feature data group and a sensor deployment feature data group.

3. The urban traffic management method based on intelligent sensor network drive according to claim 2 is characterized in that: When performing feature analysis, a sensor coverage index is generated based on the sensor coverage feature data set, and a layout uniformity index is generated based on the sensor layout feature data set.

4. The urban traffic management method based on intelligent sensor network drive according to claim 3 is characterized in that: The logic for obtaining the sensor coverage index is: First calculate the effective coverage area of ​​a single sensor in use: ; Indicates the sensing radius of the i-th sensor in use, reflecting the coverage of the sensor. It represents the environmental adaptability factor, which reflects the coverage capability of the i-th sensor in use in the actual environment, and its value range is from zero to one. represents the effective coverage area of ​​the i-th sensor in use; Then calculate the priority weights for the coverage areas: ; represents the importance score of the area covered by the i-th sensor in use, N represents the number of sensors in use, j is the sensor number index in use, represents the importance score of the jth sensor coverage area in use, represents the priority weight of the i-th sensor in use; Correct the duplicate coverage contribution value: ; represents the overlapping area of ​​the coverage area of ​​the i-th and j-th sensors in use, Indicates that the overlap area of ​​the i-th and j-th sensors in use corresponds to the preset proportionality coefficient, represents the duplicate coverage contribution value; Divide the road section where the sensor network nodes are deployed into M sub-areas, calculate the sum of the monitoring areas of the M sub-areas, and get the total monitoring area ; Finally calculate the sensor coverage index: ; Represents the sensor coverage index of all sensors in use.

5. The urban traffic management method based on intelligent sensor network drive according to claim 4 is characterized in that: Environmental adaptability factor The acquisition logic is: Introducing obstacle factors , represents the impact of obstacles on the coverage area of ​​the i-th sensor in use, which is obtained by calculating the ratio of the obstacle area to the coverage area; Introducing temperature factor , represents the current ambient temperature of the i-th sensor in use; The illumination factor is introduced to represent the current illumination intensity of the i-th sensor in use state; Substitute into the following calculation formula: ; Indicates the preset maximum obstacle density. Indicates the preset maximum temperature deviation range, Indicates the optimal operating temperature of the sensor. Indicates the optimal working light intensity of the sensor. It represents the influence weight of the kth environmental factor on the adaptability factor, and the values ​​of k are 1, 2, and 3.

6. The urban traffic management method based on intelligent sensor network drive according to claim 5 is characterized in that: The logic for obtaining the layout uniformity index is: The road section where the sensor network nodes are deployed is divided into M sub-areas, and the sub-areas are numbered as r. The standard deviation of the distance between the sensors in use in the rth sub-area is calculated first. , and then calculate the average distance between sensors in use in the rth sub-area ; According to the coverage deviation degree of the sensors in use in the rth sub-area, the coverage penalty coefficient is introduced to reflect the matching between the sensor layout and the actual demand. : ; The deployment capacity of the sensors in use in the r-th sub-area is calculated by calculating the sum of the effective coverage areas of the sensors in use in the r-th sub-area and then subtracting the total overlapping area in the sub-area; The ideal deployment capacity of the sensors in the r-th sub-area in use is the sum of the coverage areas of the sensors in the r-th sub-area in use under ideal conditions without obstacles and interference; The calculation formula of layout uniformity index is: ; Indicates the uniformity index of all sensors in use.

7. The urban traffic management method based on intelligent sensor network drive according to claim 6 is characterized in that: Inputting the feature analysis results into the pre-trained machine learning model 1 to generate the road section data quality coefficient, and then deciding whether to make a secondary decision means: The sensor coverage index and the layout uniformity index are input into a pre-trained machine learning model. The output result is the road section data quality coefficient. The road section data quality coefficient is compared with the preset quality threshold. If the road section data quality coefficient is greater than or equal to the preset quality threshold, no secondary decision is made. If the road section data quality coefficient is less than the preset quality threshold, a secondary decision is made.

8. The urban traffic management method based on intelligent sensor network drive according to claim 7 is characterized in that: When making secondary decisions, fuzzy logic is used to take the sensor coverage index, layout uniformity index, and sensor network node usage ratio of the road section as input data, and the type of optimization direction as the output data of the fuzzy logic.

9. An urban traffic management system driven by an intelligent sensor network, based on the urban traffic management method driven by an intelligent sensor network according to any one of claims 1 to 8, characterized in that: include: The data collection module is composed of a data collection unit 1 and a data collection unit 2. The data collection unit 1 collects traffic data in real time through intelligent sensor network nodes deployed on urban roads. At the same time, the data collection unit 2 obtains the usage of sensor network nodes in different sections of urban roads and generates a section sensor network node usage information set. A feature extraction module extracts features related to the usage of sensor network nodes based on the usage information set of the road segment sensor network nodes; The feature analysis module performs feature analysis operations based on the features related to the usage of the sensor network nodes to obtain feature analysis results; The primary decision module inputs the feature analysis results into the pre-trained machine learning model 1 to generate the road section data quality coefficient, and then decides whether to make a secondary decision; The secondary decision module, when making a secondary decision, substitutes the feature analysis results and the usage ratio of the sensor network nodes of the road section into the pre-trained machine learning model 2 to determine the type of optimization direction; The quality assurance module improves the usage of sensor network nodes on road sections based on the type of optimization direction, ensuring the quality of real-time traffic data collected by intelligent sensor network nodes on urban roads.

Images