Traffic flow prediction method and system based on intelligent video analysis
By configuring a multi-objective optimization model and edge computing nodes, combined with intelligent video analysis and communication protocol optimization, the problems of low accuracy in traffic flow prediction and high latency in command transmission have been solved. This has enabled accurate dynamic prediction of traffic flow and coordinated optimization of signals, thereby improving the efficiency and adaptability of the road network.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- INTELLIGENT INTER CONNECTION TECH CO LTD
- Filing Date
- 2026-03-11
- Publication Date
- 2026-06-16
AI Technical Summary
Existing technologies suffer from low accuracy in traffic flow prediction, insufficient adaptability of control schemes to complex traffic scenarios, and high latency in command transmission, resulting in limited road network efficiency and failing to meet the real-time, accuracy, and collaborative requirements of modern urban intelligent traffic management.
By acquiring traffic infrastructure parameters and operating environment parameters at intersections, configuring multi-objective optimization models, dynamically adjusting signal phase switching timing, video feature extraction frequency, and edge computing resource allocation strategies, and combining road network operation efficiency simulation, a signal collaborative optimization scheme is generated. Signal control commands are deployed through edge computing nodes, command transmission is optimized using TSN and MQTT communication protocols, and a lightweight state caching unit and a traffic anomaly event modal database are integrated to achieve accurate dynamic prediction of traffic flow and signal collaborative optimization.
It enables accurate dynamic prediction of traffic flow at urban intersections and coordinated signal control, improving road network efficiency and enhancing the system's robustness, real-time performance, adaptability, and accuracy under complex and changing conditions.
Smart Images
Figure CN122223958A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of traffic flow prediction technology, and specifically to a traffic flow prediction method and system based on intelligent video analysis. Background Technology
[0002] In the current field of urban traffic management, traffic flow prediction and signal timing control mostly rely on single-data-dimensional analysis or static timing strategies. Traditional methods do not fully integrate static parameters of intersection infrastructure with real-time dynamic environmental data, and lack multi-objective collaborative optimization model design, which easily leads to problems such as low accuracy of traffic flow prediction and disconnect between signal timing and actual traffic flow. At the same time, existing technologies mostly adopt centralized computing architectures, resulting in high command transmission latency, lack of flexibility in edge resource scheduling, and no dedicated state cache and abnormal event adaptation mechanism. Under special conditions such as communication interruption and signal switching, task state is easily lost. When facing abnormal traffic scenarios such as traffic accidents, road construction, and severe weather, signal control solutions cannot quickly and dynamically adapt, resulting in limited road network traffic efficiency and difficulty in effectively alleviating congestion problems. Furthermore, the lack of virtual-real fusion correction capabilities of digital twins makes it difficult to achieve closed-loop optimization of traffic flow prediction and signal timing, which can no longer meet the high requirements of modern urban intelligent traffic management for real-time performance, accuracy, collaboration, and emergency response capabilities.
[0003] Existing technologies suffer from low accuracy in traffic flow prediction, insufficient adaptability of control schemes to complex traffic scenarios, and high latency in command transmission. Summary of the Invention
[0004] This application provides a traffic flow prediction method and system based on intelligent video analysis, which is used to address the technical problems of low accuracy in traffic flow prediction, insufficient adaptability of control schemes to complex traffic scenarios, and high instruction transmission latency in the prior art.
[0005] In view of the above problems, this application provides a traffic flow prediction method and system based on intelligent video analysis.
[0006] The first aspect of this application provides a traffic flow prediction method based on intelligent video analysis, the method comprising:
[0007] Acquire traffic infrastructure parameters and operating environment parameters at the intersection; based on these parameters, configure a multi-objective optimization model that meets the needs of urban traffic management, dynamically adjust the signal phase switching timing, video feature extraction frequency, and edge computing resource allocation strategy; simulate the road network operating efficiency under different traffic flow conditions, and generate a signal collaborative optimization scheme under dynamic traffic flow prediction by combining the multi-objective optimization model; deploy edge computing nodes, and configure signal control command issuance according to the signal collaborative optimization scheme.
[0008] A second aspect of this application provides a traffic flow prediction system based on intelligent video analytics, the system comprising:
[0009] The system includes: an operating environment parameter acquisition module for acquiring traffic infrastructure parameters and operating environment parameters at the intersection; a multi-objective optimization model configuration module for configuring a multi-objective optimization model that meets the needs of urban traffic management based on the traffic infrastructure parameters and operating environment parameters, and dynamically adjusting the signal phase switching timing, video feature extraction frequency, and edge computing resource allocation strategy; an optimization scheme generation module for simulating road network operating efficiency under different traffic flow states, and generating a signal collaborative optimization scheme under dynamic traffic flow prediction based on the multi-objective optimization model; and a signal control command issuance module for deploying edge computing nodes and configuring signal control command issuance according to the signal collaborative optimization scheme.
[0010] One or more technical solutions provided in this application have at least the following technical effects or advantages:
[0011] The system acquires traffic infrastructure parameters and operating environment parameters at intersections; based on these parameters, it configures a multi-objective optimization model that meets urban traffic management needs, dynamically adjusting signal phase switching timing, video feature extraction frequency, and edge computing resource allocation strategies; it simulates road network operating efficiency under different traffic flow patterns, and, combined with the multi-objective optimization model, generates a signal coordination optimization scheme for dynamic traffic flow prediction; it deploys edge computing nodes and configures signal control command issuance according to the signal coordination optimization scheme. This achieves the technical effect of realizing accurate dynamic prediction of traffic flow and coordinated intelligent control of traffic signals at urban intersections, improving road network traffic efficiency. Attached Figure Description
[0012] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0013] Figure 1 A schematic diagram of the traffic flow prediction method based on intelligent video analysis provided in the embodiments of this application;
[0014] Figure 2 A schematic diagram of the traffic flow prediction system based on intelligent video analysis provided in this application embodiment.
[0015] Figure labeling: Module 10 for obtaining operating environment parameters, Module 20 for configuring multi-objective optimization model, Module 30 for generating optimization scheme, and Module 40 for issuing signal control commands. Detailed Implementation
[0016] This application provides a traffic flow prediction method and system based on intelligent video analysis, which addresses the technical problems of low accuracy in traffic flow prediction, insufficient adaptability of control schemes to complex traffic scenarios, and high instruction transmission latency in existing technologies.
[0017] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.
[0018] Example 1, as Figure 1 As shown, this application provides a traffic flow prediction method based on intelligent video analysis, the method comprising:
[0019] Step S100: Obtain the traffic infrastructure parameters and operating environment parameters of the intersection.
[0020] Specifically, through intelligent video analysis equipment and urban traffic sensing terminals, data on two core dimensions of the target intersection—traffic infrastructure parameters and operating environment parameters—are accurately collected. The traffic infrastructure parameters include the actual number of lanes at the intersection, signal phase configuration, and the specific locations of traffic detectors. The operating environment parameters cover real-time weather conditions at the intersection, special event markers within the area, and on-site lighting conditions. The acquired comprehensive and high-precision basic data provides comprehensive and reliable data support for the accurate configuration of subsequent multi-objective optimization models, dynamic prediction of traffic flow, and generation of signal collaborative optimization schemes. This is the prerequisite for realizing large-scale model-enabled intelligent traffic flow prediction and signal control.
[0021] Step S200: Based on the traffic infrastructure parameters and operating environment parameters, configure a multi-objective optimization model that meets the needs of urban traffic management, and dynamically adjust the signal phase switching timing, video feature extraction frequency and edge computing resource allocation strategy.
[0022] Specifically, based on the obtained intersection traffic infrastructure parameters and operating environment parameters, a multi-objective optimization model tailored to the actual needs of urban traffic management is configured. Signal cycle duration, green light ratio allocation, video analysis frame rate, edge AI inference load, and regional traffic capacity are incorporated into the model's core decision variables. Reasonable constraints are set for each decision variable, including minimum pedestrian crossing time, maximum queue overflow threshold, and upper limit of computing resources. Leveraging the multi-dimensional optimization capabilities of the large model, this model achieves dynamic and coordinated adjustment of signal phase switching timing, video feature extraction frequency, and edge computing resource allocation strategies. This allows signal control, video analysis, and edge computing allocation to adapt to the real-time traffic conditions and operating environment of the intersection, achieving precise matching of multiple elements and optimal resource allocation. Furthermore, this configuration method breaks through the limitations of single-objective optimization in traditional traffic control models, achieving coordinated optimization of multi-dimensional traffic control elements and improving the model's adaptability to complex traffic scenarios.
[0023] Step S300: Simulate the road network operation efficiency under different traffic flow conditions, and combine the multi-objective optimization model to generate a signal collaborative optimization scheme under dynamic traffic flow prediction.
[0024] Specifically, the core step in generating a traffic flow prediction solution based on intelligent video analytics relies on a pre-configured multi-objective optimization model. First, it performs a comprehensive simulation of the operational efficiency of target intersections and related road networks under different traffic flow patterns, such as off-peak, peak, and slow-moving traffic. This accurately calculates key indicators such as road network capacity, vehicle queuing time, and node congestion probability under different flow patterns. Then, combining the decision variable constraints and optimization objectives of the multi-objective optimization model with the core requirement of dynamic traffic flow prediction, it comprehensively generates a signal coordination optimization solution adapted to the actual traffic infrastructure and operating environment of the intersection. Simultaneously, during the solution generation process, real-time data feedback from the intersection's traffic digital twin can be used to correct traffic flow prediction deviations and signal timing errors. This ensures that the solution not only meets the computational and control constraints of the multi-objective optimization model but also accurately matches dynamically changing traffic flow patterns, achieving deep coordination between traffic flow prediction and signal control. This overcomes the technical limitations of traditional static signal timing solutions and poor adaptability, improving the solution's adaptability and feasibility for complex and ever-changing traffic scenarios.
[0025] Step S400: Deploy edge computing nodes and configure signal control commands to be issued according to the signal collaborative optimization scheme.
[0026] Specifically, the process begins by deploying edge computing nodes at core nodes of the target traffic network. Leveraging the low latency and high computing power of edge computing, hardware support is provided for the rapid processing and issuance of signal control commands. Then, based on the generated traffic flow dynamic prediction-driven signal coordination optimization scheme, various signal control commands, such as phase extension commands, phase skipping commands, and green wave coordination commands, are specifically configured and executed. Simultaneously, based on the deployed edge computing nodes, a dedicated data synchronization mechanism is built at the urban traffic data interaction interface using TSN and MQTT communication protocols. This mechanism accurately corrects latency issues in the transmission of signal control commands, ensuring the real-time and accurate issuance of commands. Furthermore, this step incorporates dynamic data feedback from the intersection's traffic digital twin to perform real-time verification and adjustment of the issued signal control commands, ensuring a high degree of adaptation between the commands and the actual traffic flow of the road network. This overcomes the technical limitations of traditional traffic signal control commands, such as high transmission latency and poor adaptability, achieving efficient and accurate implementation of traffic signal management empowered by a large model. This provides a reliable execution guarantee for the practical implementation of traffic flow prediction and signal coordination optimization.
[0027] In one possible implementation, step S100 further includes:
[0028] The traffic infrastructure parameters include the number of lanes, signal phase configuration, and detector deployment locations.
[0029] The operating environment parameters include weather conditions, special event markers, and lighting conditions.
[0030] Specifically, the collected basic data is divided into two core categories: traffic infrastructure parameters and operating environment parameters. Traffic infrastructure parameters are inherent static basic data of the intersection, including the actual number of lanes at the target intersection, the preset signal phase configuration scheme of the intersection's traffic signal controller, and the specific deployment location of traffic perception detectors at the intersection. These parameters are the basic hardware basis for configuring multi-objective optimization models and formulating signal collaborative optimization schemes, directly determining the basic adaptability of intersection traffic control. Operating environment parameters are real-time dynamic environmental data of the intersection, including real-time weather conditions monitored at the intersection, special event markers such as traffic control or large-scale events within the jurisdiction, and the actual lighting conditions at the intersection. These dynamic parameters can accurately reflect the environmental influencing factors of real-time traffic at the intersection, providing real-time basis for the dynamic adjustment of multi-objective optimization models and the adaptation adjustment of video feature extraction frequency. This combination of static basic parameters and dynamic environmental parameters makes subsequent traffic flow prediction and signal control schemes more closely aligned with the actual traffic scenarios at the intersection. It is also an important data support for large-scale models to empower intelligent traffic management and achieve dynamic and precise control.
[0031] In one possible implementation, step S200 further includes:
[0032] Step S210: Using signal cycle duration, green signal ratio allocation, video analysis frame rate, edge AI inference load, and regional traffic capacity as decision variables, set a constraint range, which includes minimum pedestrian crossing time, maximum queue overflow threshold, and upper limit of computing power resources.
[0033] Specifically, based on a multi-objective optimization model built according to traffic infrastructure parameters and operating environment parameters to meet the needs of urban traffic management, five key elements are specifically selected as core decision variables: signal cycle duration, green light ratio allocation, video analysis frame rate, edge AI inference load, and regional traffic capacity. This comprehensively covers multi-dimensional management needs such as traffic signal timing, intelligent video analysis, edge computing resource scheduling, and road network traffic control, achieving full-element coverage of multi-objective optimization. At the same time, clear constraints are scientifically set for each decision variable. These constraints include the minimum pedestrian crossing time to ensure pedestrian safety, the maximum queue overflow threshold to prevent traffic congestion at intersections, and the upper limit of computing power resources to adapt to edge computing hardware capabilities. By quantifying the constraint boundaries, problems such as extreme timing, resource overload, and traffic safety hazards that may occur during model optimization are avoided. This ensures that the model optimization results not only meet the needs of dynamic urban traffic management but also comply with the safety and hardware capability requirements of actual traffic operation. This breaks through the technical limitations of traditional traffic optimization models with single decision variables and vague constraint boundaries, improving the practicality and applicability of multi-objective optimization models.
[0034] In one possible implementation, step S210 further includes:
[0035] Step S211: Construct a digital twin of the intersection traffic, connect to the urban traffic data interaction interface, and upload vehicle trajectory data, queue length estimates, and signal status logs.
[0036] Step S212: Based on the vehicle trajectory data, queue length estimate, and signal status log, dynamically correct the traffic flow prediction deviation and signal timing error using the intersection traffic digital twin.
[0037] Specifically, relying on the comprehensive traffic perception data of intersections collected by intelligent video analysis and the acquired traffic infrastructure parameters, a digital twin of the intersection traffic is constructed, which is fully and faithfully mapped to all elements of the physical intersection. This digital twin completely replicates the physical characteristics and dynamic states of the intersection, such as lane layout, signal facilities, and traffic flow. At the same time, this digital twin is seamlessly connected to the urban traffic data interaction interface, establishing a two-way data interaction link between the physical intersection and the urban traffic management platform. Through the interface, core traffic operation data such as vehicle trajectory data collected in real time by intelligent video analysis, estimated values of vehicle queue lengths in each direction of the intersection, and signal status logs of traffic lights are uploaded and synchronized in real time, allowing the digital twin to dynamically and accurately restore the real-time traffic operation of the physical intersection.
[0038] Based on real traffic operation data such as vehicle trajectory data, queue length estimates, and signal status logs obtained from urban traffic data interaction interfaces, this data is compared in real time with theoretical traffic flow prediction data and signal timing data output by the multi-objective optimization model in the intersection traffic digital twin. The differences and error nodes in dimensions such as traffic flow efficiency, intersection queue length, and signal phase switching timing are extracted. Then, utilizing the high-fidelity simulation capabilities of the digital twin, the actual traffic flow at the intersection is replicated, and the causes of deviations and the impact of errors are simulated and deduced to clarify the direction and magnitude of model parameter adjustments. Finally, based on large-scale... The model's data analysis and optimization capabilities enable dynamic fine-tuning and strategy correction of key indicators such as the core parameters of the traffic flow prediction algorithm, the cycle length of signal timing, and the allocation of green light ratios. Simultaneously, the corrected parameters and schemes are fed back into the digital twin for simulation verification, ensuring that the corrected traffic flow prediction results and signal timing schemes are highly adapted to the actual traffic operation needs of the intersection. Through a closed-loop implementation of actual data verification, digital simulation deduction, model parameter optimization, and simulation verification, the correction of traffic flow prediction and signal timing becomes more targeted and accurate, effectively reducing the deviation between the model's theoretical values and actual traffic operation values.
[0039] In one possible implementation, step S400 further includes:
[0040] Step S410: The signal control commands include phase extension command, phase jump command, and green wave coordination command.
[0041] Step S420: Based on the edge computing node, a data synchronization mechanism is configured on the urban traffic data interaction interface using the TSN communication protocol and the MQTT communication protocol. The data synchronization mechanism is used to correct the transmission delay of the signal control command.
[0042] Specifically, based on the signal coordination optimization scheme under dynamic traffic flow prediction, three types of core signal control commands are customized for different traffic flow patterns and control needs at intersections. Each command performs its own function and is precisely adapted to different traffic scenarios: the phase extension command is suitable for scenarios where traffic flow in a certain direction of the intersection suddenly increases and vehicle queues are too long, improving traffic flow efficiency by extending the green light phase duration in that direction; the phase skipping command is suitable for scenarios where there are no vehicles passing in a certain direction of the intersection or traffic flow is extremely low, directly skipping the phase switching in that direction, reducing the waiting time of invalid signals, and improving the overall signal timing efficiency; the green wave coordination command is suitable for the coordinated control scenarios of continuous intersections such as urban arterial roads and traffic corridors, uniformly coordinating the signal phase and timing of intersections along the road to create green wave traffic belts, enabling vehicles to pass continuously on the main road, and significantly reducing the number of stops and queuing time at intersections along the road. The three types of instructions cover multiple dimensions of traffic control needs, such as dynamic control of single-point intersections and collaborative control of arterial road intersections. They enable refined, scenario-based, and collaborative control of traffic signals, breaking through the technical limitations of traditional signal control instructions being singular and poorly adaptable. They allow signal control instructions to be highly matched with the real-time traffic flow at intersections and regional traffic control needs, providing precise instruction support for the implementation of signal collaborative optimization schemes under traffic flow prediction.
[0043] Leveraging edge computing nodes already deployed at core road network nodes, and fully combining the low latency, hard real-time, and high determinism of the TSN communication protocol with the lightweight, low bandwidth, and easy-to-adapt-to-edge-scenario advantages of the MQTT communication protocol, a dedicated data synchronization mechanism for dual-protocol collaboration is established at the urban traffic data interaction interface, creating a highly reliable data transmission link adapted to intelligent traffic signal control. This mechanism deeply integrates the local computing power of edge computing nodes with the communication advantages of dual protocols. Throughout the entire transmission link from the control platform to the intersection signal controller, it collects key data such as timestamps of command transmission, data packet loss, and link latency in real time. Through collaborative verification and dynamic scheduling of the dual protocols, it accurately detects, quantifies, and corrects the latency generated during command transmission, effectively avoiding problems such as latency fluctuations, data asynchrony, and command execution lag that are prone to occur in traditional single-protocol transmission. This ensures that signal control commands such as phase extension, phase skipping, and green wave coordination can be implemented synchronously with the real-time traffic flow at the intersection. The innovative design of this dual-protocol integrated data transmission mechanism breaks through the technical limitations of insufficient latency control in traditional traffic signal command transmission. It achieves low latency, high accuracy, and high reliability in signal control command transmission, providing key communication guarantees for the efficient implementation of signal collaborative optimization schemes. This enables edge computing-enabled dynamic traffic signal control to have stronger real-time performance and operability.
[0044] In one possible implementation, step S400 further includes:
[0045] Integrated lightweight state cache unit.
[0046] During system standby, communication interruption, or signal switching, the context state of the traffic prediction task is automatically latched according to the lightweight state cache unit.
[0047] Specifically, a lightweight state caching unit is integrated into the traffic flow prediction and signal control system. This unit adopts a lightweight hardware and software co-architecture design, which enables high-speed reading and writing, real-time updating and secure storage of all core context state data of the traffic prediction task while taking into account the utilization rate of edge computing node computing resources. The cached data covers key information such as real-time traffic flow prediction results, core signal timing parameters, edge AI inference load configuration, video feature extraction frequency settings and road network operation status monitoring data. At the same time, it supports lightweight data compression and fast retrieval.
[0048] In various special operating conditions, such as when the system enters standby mode, experiences communication link interruption, or performs signal switching operations, the system automatically triggers a dedicated latching mechanism for the context state of the traffic prediction task, relying on the integrated lightweight state caching unit. This allows for fully automated processing without manual intervention. The latching operation completely and accurately retains the core context state data of the traffic prediction task, covering key information such as real-time traffic flow prediction results, current signal timing parameters, edge computing resource allocation strategies, video feature extraction frequency configuration, and real-time road network monitoring data. This ensures that the task's operational status data is not lost, tampered with, or interrupted under special conditions, effectively guaranteeing the continuity and stability of the entire process of traffic flow prediction and signal coordination optimization. This significantly improves the system's robustness, fault tolerance, and operability under complex and changing operating conditions.
[0049] In one possible implementation, step S400 further includes:
[0050] A traffic anomaly event modal database is constructed, which stores traffic accident patterns, road construction and obstruction patterns, and traffic patterns during severe weather.
[0051] Based on the traffic anomaly event modal database and combined with the context state of the traffic prediction task latched by the lightweight state cache unit, the signal collaborative optimization scheme is dynamically corrected.
[0052] Specifically, a traffic anomaly event modal database is constructed as the core data support library for adapting to various emergency traffic scenarios and dynamically correcting signal coordination optimization schemes. This database is built based on the city's full historical traffic anomaly event data, anomaly scene feature data collected in real time by intelligent video analysis, and combined with the scene modeling and pattern mining capabilities of a large model. It systematically and standardizedly stores the full-dimensional feature information of three core traffic anomaly events: traffic accident patterns, construction road occupation patterns, and severe weather traffic patterns. Among them, the traffic accident pattern covers key parameters such as the road network capacity attenuation law corresponding to different accident types, accident road occupation range, accident handling time, and traffic flow diversion adaptation strategies; the construction road occupation pattern includes core content such as the rules for allocating intersection traffic resources and phase timing adjustment standards for different construction areas, construction enclosure ranges, and construction periods; the severe weather traffic pattern includes practical information such as traffic flow speed thresholds, video feature extraction adaptation strategies, and signal timing optimization criteria under different severe weather levels such as rain, snow, fog, and haze.
[0053] Based on a traffic anomaly event modal database, the system retrieves control models and timing adjustment criteria that precisely match the actual anomaly scenarios in the current road network in real time. Simultaneously, it extracts the full contextual state data of the traffic prediction task latched by the lightweight state cache unit, including real-time traffic prediction results at the latching time, core signal timing parameters, edge computing resource allocation strategies, and video feature extraction configurations. This ensures that the scheme correction is consistent with the previous traffic control status, avoiding policy gaps. Building upon this, and combining the traffic impact patterns of anomaly scenarios in the database with the latched task context, targeted adjustments are made to the signal phase switching timing, green light ratio allocation, video feature extraction frequency, edge computing resource scheduling strategies, and signal control command types in the signal coordination optimization scheme. For example, the green light duration for traffic accident-occupied lane scenarios is increased, and the video analysis frame rate is reduced and adapted to low-speed traffic signal timing for severe weather scenarios. This ensures that the corrected signal coordination optimization scheme not only conforms to the real-time traffic anomaly conditions of the road network but also maintains the continuity and stability of traffic control.
[0054] Example 2, based on the same inventive concept as the traffic flow prediction method based on intelligent video analysis in the foregoing examples, such as... Figure 2 As shown, this application provides a traffic flow prediction system based on intelligent video analysis. The system and method embodiments in this application are based on the same inventive concept. The system includes:
[0055] The operating environment parameter acquisition module 10 is used to acquire traffic infrastructure parameters and operating environment parameters at the intersection.
[0056] The multi-objective optimization model configuration module 20 is used to configure a multi-objective optimization model that meets the needs of urban traffic management based on the traffic infrastructure parameters and operating environment parameters, and to dynamically adjust the signal phase switching timing, video feature extraction frequency and edge computing resource allocation strategy.
[0057] The optimization scheme generation module 30 is used to simulate the road network operation efficiency under different traffic flow conditions, and in combination with the multi-objective optimization model, generate a signal collaborative optimization scheme under dynamic traffic flow prediction.
[0058] The signal control command issuing module 40 is used to deploy edge computing nodes and configure the issuance of signal control commands according to the signal collaborative optimization scheme.
[0059] Furthermore, the system is also used to implement the following functions:
[0060] The traffic infrastructure parameters include the number of lanes, signal phase configuration, and detector deployment locations; the operating environment parameters include weather conditions, special event markers, and lighting conditions.
[0061] Furthermore, the system is also used to implement the following functions:
[0062] The signal cycle duration, green signal ratio allocation, video analysis frame rate, edge AI inference load, and regional traffic capacity are used as decision variables, and a constraint range is set. The constraint range includes the minimum pedestrian crossing time, the maximum queuing overflow threshold, and the upper limit of computing power resources.
[0063] Furthermore, the system is also used to implement the following functions:
[0064] A digital twin of the intersection traffic is constructed, connected to the urban traffic data interaction interface, and vehicle trajectory data, queue length estimates, and signal status logs are uploaded. Based on the vehicle trajectory data, queue length estimates, and signal status logs, the traffic flow prediction deviation and signal timing error are dynamically corrected in conjunction with the digital twin of the intersection traffic.
[0065] Furthermore, the system is also used to implement the following functions:
[0066] The signal control commands include phase extension commands, phase skipping commands, and green wave coordination commands. Based on the edge computing node, a data synchronization mechanism is configured on the urban traffic data interaction interface using the TSN and MQTT communication protocols. The data synchronization mechanism is used to correct the transmission delay of the signal control commands.
[0067] Furthermore, the system is also used to implement the following functions:
[0068] An integrated lightweight state cache unit is used to automatically latch the context state of the traffic prediction task during system standby, communication interruption, or signal switching.
[0069] Furthermore, the system is also used to implement the following functions:
[0070] A traffic anomaly event modal database is constructed, which stores traffic accident patterns, road construction and obstruction patterns, and severe weather traffic patterns. Based on the traffic anomaly event modal database and combined with the context state of the traffic prediction task latched by the lightweight state cache unit, the signal collaborative optimization scheme is dynamically corrected.
[0071] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.
[0072] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.
[0073] This specification and accompanying drawings are merely illustrative examples of this application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Therefore, if such modifications and variations fall within the scope of this application and its equivalents, this application intends to include such modifications and variations.
Claims
1. A traffic flow prediction method based on intelligent video analytics, characterized in that, The method includes: Obtain traffic infrastructure parameters and operating environment parameters at the intersection; Based on the traffic infrastructure parameters and operating environment parameters, a multi-objective optimization model that meets the needs of urban traffic management is configured to dynamically adjust the signal phase switching timing, video feature extraction frequency and edge computing resource allocation strategy. The road network operation efficiency under different traffic flow conditions is simulated, and a signal collaborative optimization scheme under dynamic traffic flow prediction is generated by combining the multi-objective optimization model. Deploy edge computing nodes and configure signal control commands to be issued according to the aforementioned signal coordination optimization scheme.
2. The traffic flow prediction method based on intelligent video analysis as described in claim 1, characterized in that, The method includes: The traffic infrastructure parameters include the number of lanes, signal phase configuration, and detector deployment locations; The operating environment parameters include weather conditions, special event markers, and lighting conditions.
3. The traffic flow prediction method based on intelligent video analysis as described in claim 1, characterized in that, Based on the aforementioned transportation infrastructure parameters and operating environment parameters, a multi-objective optimization model that meets the needs of urban traffic management is configured. The method further includes: The signal cycle duration, green signal ratio allocation, video analysis frame rate, edge AI inference load, and regional traffic capacity are used as decision variables, and a constraint range is set. The constraint range includes the minimum pedestrian crossing time, the maximum queuing overflow threshold, and the upper limit of computing power resources.
4. The traffic flow prediction method based on intelligent video analysis as described in claim 3, characterized in that, The method includes: Construct a digital twin of the intersection traffic, connect to the urban traffic data exchange interface, and upload vehicle trajectory data, queue length estimates, and signal status logs; Based on the vehicle trajectory data, queue length estimates, and signal status logs, the traffic flow prediction deviation and signal timing error are dynamically corrected using the intersection traffic digital twin.
5. The traffic flow prediction method based on intelligent video analysis as described in claim 4, characterized in that, According to the aforementioned signal coordination optimization scheme, the method for configuring signal control command issuance includes: The signal control commands include phase extension commands, phase skipping commands, and green wave coordination commands. Based on the edge computing node, a data synchronization mechanism is configured on the urban traffic data interaction interface using the TSN and MQTT communication protocols. The data synchronization mechanism is used to correct the transmission delay of the signal control commands.
6. The traffic flow prediction method based on intelligent video analysis as described in claim 1, characterized in that, The method includes: Integrated lightweight state cache unit; During system standby, communication interruption, or signal switching, the context state of the traffic prediction task is automatically latched according to the lightweight state cache unit.
7. The traffic flow prediction method based on intelligent video analysis as described in claim 6, characterized in that, The method includes: A traffic anomaly event modal database is constructed, which stores traffic accident patterns, road construction and obstruction patterns, and severe weather traffic patterns. Based on the traffic anomaly event modal database and combined with the context state of the traffic prediction task latched by the lightweight state cache unit, the signal collaborative optimization scheme is dynamically corrected.
8. A traffic flow prediction system based on intelligent video analytics, characterized in that, The system is used to implement the traffic flow prediction method based on intelligent video analysis as described in any one of claims 1-7, and the system comprises: The operating environment parameter acquisition module is used to acquire traffic infrastructure parameters and operating environment parameters at the intersection. The multi-objective optimization model configuration module is used to configure a multi-objective optimization model that meets the needs of urban traffic management based on the traffic infrastructure parameters and operating environment parameters, and to dynamically adjust the signal phase switching timing, video feature extraction frequency and edge computing resource allocation strategy. The optimization scheme generation module is used to simulate the road network operation efficiency under different traffic flow states, and combined with the multi-objective optimization model, generate a signal collaborative optimization scheme under dynamic traffic flow prediction. The signal control command issuing module is used to deploy edge computing nodes and configure the issuance of signal control commands according to the signal collaborative optimization scheme.