Traffic control strategy self-adaptive method and system based on simulation feedback

Abstract

The application provides a traffic control strategy adaptive method and system based on simulation feedback, belonging to the technical field of traffic prediction and control. First, traffic control scene data containing traffic flow data and road condition information is obtained, then a simulation evaluation environment is constructed, scene parameters are configured based on the above traffic control scene data, the traffic running state under different traffic control strategies is simulated, then a pre-trained reinforcement learning model is called to simulate and evaluate each strategy in the traffic control strategy set, generating a strategy effect feedback set containing traffic operation efficiency indicators and traffic order stability indicators, according to the strategy effect feedback set, the parameters of the strategy are adjusted considering the index correlation, obtaining the adjusted strategy, finally the adjusted strategy is output to the traffic control system, realizing strategy updating, thereby effectively improving traffic operation efficiency, guaranteeing traffic order stability, and realizing adaptive optimization of traffic control strategy.

Description

Technical Field

[0001] The present invention relates to the technical field of traffic control, and in particular to a traffic control strategy adaptive method and system based on simulation feedback. Background Art

[0002] In the current field of traffic control, traditional methods for formulating and adjusting traffic control strategies face numerous limitations. For one thing, many cities still rely on experienced traffic management personnel to formulate control strategies based on past experience and real-time observations. This approach relies heavily on individual experience, lacks scientific and systematic principles, and is difficult to adapt to complex and changing traffic conditions. The experience levels of different managers vary, resulting in inconsistent strategies and an inability to guarantee effective traffic control in different scenarios.

[0003] On the other hand, some approaches that use fixed models to formulate traffic control strategies, while scientific to a certain extent, are often based on specific traffic scenarios and assumptions. However, real-world traffic scenarios are dynamic and subject to a variety of factors, such as time, weather, and emergencies. Fixed models are unable to adapt to these changes in a timely manner, resulting in poor strategy effectiveness in new traffic scenarios and an inability to effectively improve traffic efficiency and maintain traffic order stability. Summary of the Invention

[0004] In view of the above-mentioned problems, in combination with the first aspect of the present invention, an embodiment of the present invention provides a traffic control strategy adaptive method based on simulation feedback, the method comprising:

[0005] Acquire traffic control scenario data, the traffic control scenario data including traffic flow data and road condition information, the traffic flow data representing the flow state of vehicles within the control area, and the road condition information representing the physical properties and sign configuration state of roads within the control area;

[0006] Constructing a simulation evaluation environment, wherein the simulation evaluation environment is used to simulate traffic operation status under different traffic control strategies, and the simulation evaluation environment configures scenario parameters based on traffic flow data and road condition information in the traffic control scenario data;

[0007] Calling a pre-trained reinforcement learning model to perform simulation evaluation processing on each traffic control strategy in the traffic control strategy set in the simulation evaluation environment, and generating a strategy effect feedback set corresponding to each traffic control strategy, wherein the strategy effect feedback set includes a traffic operation efficiency index and a traffic order stability index;

[0008] performing parameter adjustment processing on each traffic control strategy in the traffic control strategy set according to the strategy effect feedback set to obtain an adjusted traffic control strategy, wherein the parameter adjustment processing is performed based on a correlation between a traffic operation efficiency index and a traffic order stability index;

[0009] The adjusted traffic control strategy is output to the traffic control system to update the traffic control strategy currently executed by the traffic control system.

[0010] On the other hand, an embodiment of the present invention also provides a traffic control strategy adaptive system based on simulation feedback, including a processor and a machine-readable storage medium, wherein the machine-readable storage medium is connected to the processor, the machine-readable storage medium is used to store programs, instructions or codes, and the processor is used to execute the programs, instructions or codes in the machine-readable storage medium to implement the above method.

[0011] Based on the above aspects, the embodiments of the present invention can comprehensively and accurately understand the traffic conditions in the control area by acquiring traffic control scenario data including traffic flow data and road condition information, and construct a simulation evaluation environment based on traffic control scenario data, which can accurately simulate the traffic operation status under different traffic control strategies, effectively avoiding the risks and costs brought by directly testing strategies in actual traffic environments, calling pre-trained reinforcement learning models to simulate and evaluate traffic control strategies, and generating a strategy effect feedback set including traffic operation efficiency indicators and traffic order stability indicators. It can scientifically and objectively evaluate the strategy effect, adjust the strategy parameters according to the strategy effect feedback set, and fully consider the correlation between traffic operation efficiency indicators and traffic order stability indicators, so that the adjusted strategy can better adapt to actual traffic scenarios, effectively improve traffic operation efficiency, ensure traffic order stability, and realize adaptive optimization of traffic control strategies. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 It is a schematic diagram of the execution flow of the traffic control strategy adaptive method based on simulation feedback provided by an embodiment of the present invention.

[0013] Figure 2 Schematic diagram of exemplary hardware and software components of a traffic control strategy adaptive system based on simulation feedback provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0014] The present invention will be described in detail below with reference to the accompanying drawings. Figure 1 This is a flow chart of a traffic control strategy adaptive method based on simulation feedback provided by an embodiment of the present invention. The traffic control strategy adaptive method based on simulation feedback is introduced in detail below.

[0015] This embodiment takes traffic control in the city center as an application scenario and elaborates on the specific implementation process.

[0016] Step S110: Obtain traffic control scenario data, wherein the traffic control scenario data includes traffic flow data and road condition information. The traffic flow data represents the flow state of vehicles in the control area, and the road condition information represents the physical properties and sign configuration status of the roads in the control area.

[0017] In traffic control scenarios in urban central areas, comprehensive and accurate understanding of the area's traffic conditions requires the collection of traffic control scenario data from multiple channels. Traffic flow data can be acquired through monitoring equipment deployed at various intersections and road sections within the control area. These monitoring devices can capture information about passing vehicles in real time, thereby statistically analyzing vehicle flow patterns. Road condition information must be integrated with a variety of sources, including urban road planning archives, road maintenance records from traffic management departments, and field-collected road sign information. For example, physical road attributes include length, width, and pavement material, while sign configuration status includes the location of traffic lights, the type of traffic signs, and their placement. By collecting this data, comprehensive traffic control scenario data can be constructed.

[0018] Step S111: Divide the control area into sub-areas to obtain multiple control sub-areas, each of which corresponds to an independent geographical boundary range.

[0019] After obtaining the traffic control scenario data, the first step is to divide the control area of ​​the city center into sub-regions. The basis for division can be the natural separation of roads, the distribution characteristics of traffic flow, and the division of administrative regions. For example, the city center area is divided into multiple control sub-regions based on the main rivers, railways and large arterial roads in the city. Each control sub-region has a clear and independent geographical boundary range. The above division helps to manage and analyze the traffic conditions of each area more finely. For example, a control sub-region may be bounded by an east-west main road and a north-south main road, forming a rectangular geographical range, which contains several secondary roads and branches.

[0020] Step S112: Collect traffic flow data in each controlled sub-area, where the traffic flow data includes the number of vehicles entering the controlled sub-area and the number of vehicles leaving the controlled sub-area per unit time.

[0021] After completing the sub-area division, traffic flow data is collected for each controlled sub-area. This collection can be achieved by installing vehicle detection equipment, such as induction coils and video detectors, at the entrances and exits of each controlled sub-area. These devices can accurately count the number of vehicles entering and leaving the sub-area per unit time. For example, at each entrance to a controlled sub-area, vehicle detection equipment will record the number of vehicles entering the area once an hour, and at each exit, the number of vehicles leaving. Through this collection method, traffic flow data for each controlled sub-area at different time periods can be obtained. This traffic flow data reflects the entry and exit of vehicles in the sub-area.

[0022] Step S113: Collecting traffic condition information in each control sub-area, wherein the traffic condition information includes the number of lanes, lane division method, road marking type, and speed limit sign configuration status.

[0023] To collect road condition information within each controlled sub-area, a dedicated collection team conducts on-site surveys, supplementing and verifying existing road information databases. This collection process includes detailed documentation of the number of lanes on a road, such as whether it has four or six lanes in either direction; the lane divisions, such as whether they are divided into straight-ahead, left-turn, and right-turn lanes, or whether they include variable lanes; the types of road markings, such as solid white, dashed white, and solid yellow, as different markings serve different traffic guidance purposes; and the configuration of speed limit signs, including the speed limit value and location. For example, within one controlled sub-area, a major arterial road has six lanes in either direction, with the two center lanes being variable lanes. The road's edge lines are solid white, and the lane dividers are dashed white. Speed ​​limit signs are also placed at the road's starting point and key intersections. This meticulous collection ensures the accuracy and completeness of road condition information.

[0024] Step S114: Perform timestamp alignment processing on the traffic flow data and road condition information within the same control sub-area, and integrate the traffic flow data and road condition information of each control sub-area after timestamp alignment to generate traffic control scene data covering the entire control area. The data of each control sub-area in the traffic control scene data are associated through the geographical boundary range.

[0025] After collecting traffic flow data and road condition information for each controlled sub-area, timestamp alignment is required. Because traffic flow data is collected at different time points, and road condition information, while relatively stable, can change due to road construction, maintenance, and other factors, the timestamps of both need to be aligned to the same benchmark for subsequent integrated analysis. For example, both traffic flow data and road condition information can be unified to hourly timestamps to ensure that both are processed within the same timeframe. After timestamp alignment, the data from each controlled sub-area is integrated. Since each controlled sub-area has clear geographic boundaries, these boundaries can be used to link them together, forming traffic control scenario data covering the entire city center controlled area. This allows managers to gain a holistic understanding of the traffic conditions across the entire area.

[0026] Step S120: Constructing a simulation evaluation environment, wherein the simulation evaluation environment is used to simulate traffic operation status under different traffic control strategies, and the simulation evaluation environment configures scenario parameters based on traffic flow data and road condition information in the traffic control scenario data.

[0027] After acquiring the traffic control scenario data, we can begin building a simulation evaluation environment. This is a virtual platform capable of simulating traffic operation status. By inputting traffic flow data and road condition information from the traffic control scenario data, it configures the corresponding scenario parameters, thereby simulating traffic operation status under different traffic control strategies. Building this environment requires the use of various technologies, such as computer graphics and traffic flow theory, to ensure the authenticity and accuracy of the simulation. For example, the initial number and direction of vehicles on the virtual road are set based on traffic flow data, and a virtual road network and traffic signs are constructed based on road condition information, allowing the simulation environment to restore the real traffic scene as closely as possible.

[0028] Step S121: determining the road network structure parameters of the simulation evaluation environment based on the road condition information in the traffic control scenario data, wherein the road network structure parameters include the connection relationship of the roads, the number of lanes, and the lane division method.

[0029] When constructing a simulation evaluation environment, the first step is to determine the road network structure parameters based on the road condition information from the traffic control scenario data. This requires in-depth analysis of the road condition information to extract key information related to the road network structure. Road connectivity refers to the connection between different roads, such as intersections and T-junctions; the number of lanes refers to the total number of lanes on each road; and the lane division refers to the functional allocation of lanes, such as straight ahead, left turn, and right turn. For example, based on the road condition information, a main road intersects with a secondary road. The main road has four lanes: two for straight ahead, one for left turn, and one for right turn; the secondary road has two lanes: one for straight ahead and one for right turn. Using this information, the road network structure parameters for the area in the simulation evaluation environment can be determined.

[0030] Step S1211: Analyze the road condition information in the traffic control scene data and extract the physical attribute data of the roads in each control sub-area. The physical attribute data includes the starting point coordinates, end point coordinates, road width and road type of the road.

[0031] Parsing the road condition information in the traffic control scenario data is the first step in determining the road network structure parameters. During the parsing process, data processing tools are needed to filter and extract the collected road condition information to obtain the physical attribute data of the roads in each control sub-area. The starting and ending coordinates of the road are used to determine the position and direction of the road in the virtual space; the road width affects the number and width of the virtual lanes; the road type determines the road's traffic capacity and design standards, such as expressways, main roads, secondary roads, and branch roads. For example, when parsing the road condition information of a certain control sub-area, the starting and ending coordinates of a secondary road are extracted to clarify its extension direction in the virtual space, and at the same time, the width of the road and the type of secondary road it belongs to are known.

[0032] Step S1212: Determine the spatial position relationship of each road in the control area based on the starting point coordinates and the end point coordinates of the road, and construct a road topology relationship diagram, in which the nodes in the road topology relationship diagram represent road intersections and the edges represent road segments connecting the intersections.

[0033] After obtaining the starting and ending coordinates of the roads, coordinate analysis is used to determine the spatial positional relationships of each road within the control area. This can be achieved with the help of geographic information system software, by marking the starting and ending coordinates of the roads on a virtual map, and then analyzing their relative positions and connections. Based on these spatial positional relationships, a road topology diagram is constructed. In this diagram, nodes represent road intersections, where different roads meet; edges represent road segments connecting these intersections. For example, two roads intersect to form an intersection, which is a node, and the sections of the two roads from their respective starting points to the intersection and from the intersection to their respective endings constitute different edges. By constructing a road topology diagram, the connectivity of the roads within the entire control area can be clearly displayed.

[0034] Step S1213: determining the connection relationship of the roads based on the road topology relationship graph, wherein the connection relationship indicates whether there is a directly connected road segment between any two intersections and the number of road segments.

[0035] Based on the constructed road topology diagram, the connectivity of the roads is further determined. This requires traversing all nodes and edges in the road topology diagram and analyzing whether any two intersections (nodes) have directly connected road segments (edges), as well as the number of connected road segments. For example, in the road topology diagram, if there are two directly connected road segments between nodes A and B—one is a two-way main road, and the other is a one-way branch road—then the connectivity between nodes A and B is the presence of two directly connected road segments. This analysis clearly defines the connectivity between each intersection in the road network.

[0036] Step S1214: According to the lane quantity data in the road condition information, a lane quantity attribute is configured for each road segment in the road topology diagram.

[0037] Based on the lane number data extracted from traffic information, a corresponding lane number attribute is assigned to each road segment in the road topology diagram. For example, if a road segment is shown as having three lanes in the traffic information, the corresponding road segment in the road topology diagram is assigned a three-lane attribute. This ensures that the number of lanes on the virtual road segment in the simulation evaluation environment matches that of the real road, ensuring simulation authenticity. Furthermore, the lane number configuration also affects the driving behavior of virtual vehicles on the road, such as the possibility of overtaking and lane changing.

[0038] Step S1215: Analyze the lane division data in the road condition information to determine the functional division type of the lanes in each road segment, where the functional division type includes a through lane, a left-turn lane, a right-turn lane, and a mixed lane.

[0039] Parse the lane division data in the traffic information to clarify the functional division type of lanes within each road segment. This requires carefully analyzing the descriptions and symbols of lane divisions in the traffic information to determine the specific function of each lane. For example, a road segment has four lanes. According to the traffic information, the first lane from left to right is a left-turn lane, the second and third lanes are through lanes, and the fourth lane is a right-turn lane. Therefore, the corresponding functional division type is assigned to each lane in this road segment. A mixed lane is a lane that can simultaneously meet the needs of multiple driving directions. For example, at intersections with low traffic volume, a mixed lane may be set up to allow vehicles to go through and turn right. By determining the functional division type of the lane, the vehicle's driving route and turning behavior can be accurately simulated in the simulation evaluation environment.

[0040] Step S1216: Integrate the road connection relationship, the number of lanes, and the lane division method into road network structure parameters. The road network structure parameters are used to define the basic framework of the virtual road network in the simulation evaluation environment.

[0041] The previously determined road connectivity, lane number, and lane division are integrated to form complete road network structural parameters. These parameters together form the basic framework of the virtual road network within the simulation evaluation environment and are the core basis for constructing the virtual road network. For example, information such as the connectivity of all road segments within a region, the number of lanes per segment, and the functional division type of each lane are integrated to form the road network structural parameters for that region. These parameters are then input into the simulation evaluation environment's construction program, which automatically generates the corresponding virtual road network based on these parameters.

[0042] Step S122: Determine initial traffic flow parameters of the simulation evaluation environment based on the traffic flow data in the traffic control scenario data, wherein the initial traffic flow parameters include the initial vehicle density and the initial vehicle driving speed distribution of each control sub-area.

[0043] After determining the road network structure parameters, the initial traffic flow parameters are determined based on the traffic flow data in the traffic control scenario data. The initial vehicle density refers to the number of vehicles per unit area, which reflects the degree of traffic congestion within the controlled sub-area at a given moment. The initial vehicle speed distribution refers to the distribution of vehicle speeds within the sub-area. Different types of vehicles and different road conditions will lead to differences in speeds. For example, according to traffic flow data statistics, the initial vehicle density in a controlled sub-area is relatively high during the morning rush hour, and the initial vehicle speed distribution shows a low-speed concentration, with most vehicles traveling at a low speed. During off-peak hours, the initial vehicle density is low, the initial vehicle speed distribution is relatively dispersed, and the overall speed is higher. Using this data, it is possible to set an initial traffic flow state that conforms to the actual situation for the simulation evaluation environment.

[0044] Step S123: Input the road network structure parameters and initial traffic flow parameters into the scenario modeling module to generate a basic scenario model of the simulation evaluation environment, wherein the basic scenario model includes a virtual road network and an initial vehicle distribution state consistent with the geographical structure of the control area.

[0045] The determined road network structure parameters and initial traffic flow parameters are input into the scenario modeling module. The scenario modeling module is a software module specifically designed for constructing simulation scenarios. It automatically generates a basic scenario model based on the input parameters. The basic scenario model is a core component of the simulation evaluation environment. It contains a virtual road network that is consistent with the geographical structure of the control area. The virtual road network accurately reproduces the connectivity, number of lanes, and lane divisions of the real roads according to the road network structure parameters. The basic scenario model also includes the initial vehicle distribution state. The distribution of these vehicles is set according to the initial vehicle density and driving speed distribution in the initial traffic flow parameters, so that the position and driving state of the virtual vehicles at the initial moment are consistent with the real traffic flow conditions. For example, in the basic scenario model, the layout of the main roads, secondary roads, and branches of the virtual road network is consistent with the real city center area. The virtual vehicles are distributed on each road segment according to the initial density and are ready to start at the corresponding initial driving speed.

[0046] Step S124: According to the road marking type and speed limit sign configuration status in the traffic control scene data, the attributes of the road elements in the basic scene model are configured so that the road markings and speed limit signs in the virtual road network are consistent with the road condition information of the actual control area.

[0047] After the basic scenario model is generated, the attributes of the road elements in the basic scenario model need to be configured according to the road marking type and speed limit sign configuration status in the traffic control scenario data. Road elements include road markings and speed limit signs, which are configured to be consistent with the road conditions in the actual control area. For example, for road markings, according to the type information in the data, the corresponding white solid lines, white dashed lines, yellow solid lines, etc. are drawn on the virtual road network, and the position and length of the markings are ensured to be consistent with the real road; for speed limit signs, speed limit signs are set at the corresponding positions on the virtual road network according to the configuration status, and the speed limit value on the sign is the same as the speed limit requirement of the actual road. In this way, the road elements in the virtual road network can accurately reflect the traffic rules and guidance information of the real road.

[0048] Step S125: Set boundary conditions for the configured basic scenario model to determine the entry and exit boundaries of the vehicle in the simulation evaluation environment. The entry and exit boundaries correspond to the actual entry and exit positions of the control area, generating a simulation evaluation environment that can be used to simulate traffic operation status.

[0049] The boundary conditions of the configured basic scenario model are set, mainly to determine the vehicle's entry boundary and exit boundary. The entry boundary refers to the location where the vehicle can enter the simulation evaluation environment, and the exit boundary refers to the location where the vehicle can leave the environment. The setting of these boundaries must correspond to the actual entrance and exit locations of the control area. For example, in the actual traffic in the city center area, vehicles can enter the area through certain specific road entrances and leave through specific exits. In the simulation evaluation environment, it is necessary to set the entry boundary and exit boundary at the corresponding location in the virtual road network. After the setting is completed, a simulation evaluation environment that can be used to simulate traffic operation status is generated. In this environment, virtual vehicles can enter and leave according to the set boundary conditions, simulating the real traffic flow process.

[0050] Step S130: Call the pre-trained reinforcement learning model to perform simulation evaluation processing on each traffic control strategy in the traffic control strategy set in the simulation evaluation environment, and generate a strategy effect feedback set corresponding to each traffic control strategy, wherein the strategy effect feedback set includes a traffic operation efficiency index and a traffic order stability index.

[0051] After establishing the simulation evaluation environment, a pre-trained reinforcement learning model is used to simulate and evaluate each traffic control strategy in the set. By interacting with the simulation evaluation environment, the reinforcement learning model learns the traffic operation effects under different traffic control strategies and then evaluates each strategy. The simulation evaluation process involves applying each traffic control strategy to the simulation evaluation environment, observing changes in traffic operation status, and generating a corresponding set of strategy effect feedback based on these changes. Traffic operation efficiency indicators are used to measure the operating efficiency of the traffic system, while traffic order stability indicators are used to assess the stability of traffic order. Through these evaluations, the most optimal traffic control strategy can be selected.

[0052] Step S131: Select a traffic control strategy to be evaluated from the traffic control strategy set, where the traffic control strategy includes lane control parameters and speed limit control parameters. The lane control parameters represent the lane usage allocation method, and the speed limit control parameters represent the driving speed limit values ​​of different road sections.

[0053] A traffic control strategy to be evaluated is selected from a set of traffic control strategies. The set of traffic control strategies includes a plurality of different strategies, each of which has specific lane control parameters and speed limit control parameters. The lane control parameters determine the allocation of lanes, for example, designating a lane as a bus lane at a certain time period, or dynamically adjusting the driving direction of a lane according to traffic flow, etc. The speed limit control parameters specify the speed limit values of different road segments, such as setting a lower speed limit value near a school, or setting a relatively higher speed limit value on a main road, etc. When selecting the strategy to be evaluated, it can be selected in a random manner, or according to the priority of the strategy, to ensure that each strategy is evaluated.

[0054] Step S132: input the traffic control strategy to be evaluated into the policy execution module of the reinforcement learning model, and configure the lane control parameters and speed limit control parameters to the corresponding road segments of the simulation evaluation environment by the policy execution module.

[0055] In this embodiment, after the policy execution module of the reinforcement learning model receives the traffic control strategy to be evaluated, it can first parse the lane control parameters and speed limit control parameters contained in the strategy. The lane control parameters may involve the number allocation of straight lanes, left-turn lanes and right-turn lanes in a road segment, and whether to set a temporary dedicated lane, etc. The speed limit control parameters specify the speed limit requirements of different road segments at different times. The policy execution module finds the corresponding road segments in the virtual road network of the simulation evaluation environment according to these parameters, configures the lane control parameters to the lane function settings of the road segments, for example, adjusts the number of straight lanes of a road segment, or temporarily sets a lane as a bus lane, etc. At the same time, the speed limit control parameters are mapped to the speed limit identification attributes of the road segments, and the speed limit information of the road segments in the virtual road network is updated. In this way, the simulation evaluation environment can simulate the operation of traffic flow according to the traffic control strategy to be evaluated.

[0056] Step S133: start the simulation running program of the simulation evaluation environment, and make the simulation evaluation environment simulate the traffic flow running process under the condition of configuring the traffic control strategy to be evaluated. The duration of the simulation running process is consistent with the collection period of the traffic control scene data.

[0057] After completing the configuration of the traffic control strategy to be evaluated, the simulation evaluation environment's simulation program is launched. The simulation program will begin simulating traffic flow based on the configured virtual road network, the initial vehicle distribution, and the traffic control strategy to be evaluated. During the simulation, virtual vehicles will navigate the virtual road network according to traffic rules and their own driving logic, including starting, accelerating, decelerating, turning, and changing lanes. The duration of the simulation run is set to the same as the collection period for the traffic control scenario data. This ensures that the simulation results are comparable to the actual collected data in terms of time, facilitating accurate evaluation of the effectiveness of the strategy. For example, if the collection period for the traffic control scenario data is a full morning rush hour period, the simulation run will also last for the duration of the morning rush hour period to fully simulate the traffic flow during that period under the strategy to be evaluated.

[0058] Step S134: During the simulation operation, traffic operation status data in the simulation evaluation environment is collected at preset time intervals. The traffic operation status data includes vehicle speed, vehicle spacing, and vehicle lane change frequency in each road section.

[0059] During the simulation run in the evaluation environment, the system collects traffic status data at preset intervals. The preset intervals can be set based on the changing characteristics of traffic flow. If traffic flow fluctuates rapidly, the interval can be shorter; if traffic flow is relatively stable, the interval can be extended appropriately. The collected traffic status data covers vehicle speeds, vehicle spacing, and lane-changing frequency for each road section. Vehicle speed refers to the speed of all vehicles on a road section at a given moment; vehicle spacing refers to the distance between adjacent vehicles in the same lane; and lane-changing frequency refers to the ratio of the number of lane changes on a road section per unit time to the total number of vehicles on the road section. By collecting this data at intervals, a traffic status dataset containing a time series can be formed, reflecting how traffic flow changes over time under the traffic control strategy being evaluated.

[0060] Step S1341: After the simulation running program of the simulation evaluation environment is started, a time collection window is set, where the start time of the time collection window is the start time of the simulation running, and the end time is the end time of the simulation running.

[0061] After starting the simulation run program in the simulation evaluation environment, first set the time collection window. The start time of the time collection window is the time when the simulation run begins, and the end time is the time when the simulation run ends. Its span is consistent with the duration of the simulation run. The purpose of setting the time collection window is to clarify the time range of data collection and ensure that the collected data can fully cover the entire simulation run process. For example, if the simulation run starts at the beginning of the morning rush hour and ends at the end of the morning rush hour, then the time collection window is set from the beginning of the morning rush hour to the end of the morning rush hour, and all data collection operations will be performed within this window.

[0062] Step S1342: Divide the time collection window into multiple collection periods according to a preset time interval, and each collection period corresponds to a timestamp.

[0063] The set time collection window is divided into multiple collection periods at preset time intervals. Each collection period is assigned a unique timestamp. The timestamp identifies the collection period's position within the entire time collection window, facilitating subsequent differentiation and analysis of data from different time periods. For example, if the time collection window is two hours during the morning rush hour and the preset interval is ten minutes, the time collection window will be divided into multiple collection periods, each with a timestamp that increments by ten minutes from the start of the morning rush hour.

[0064] Step S1343: In each collection period, all road sections in the simulation evaluation environment are traversed, and a vehicle status collection operation is performed on each road section.

[0065] During each collection period, the system traverses all road segments in the simulation evaluation environment to ensure that no segment's traffic status data is missed. The traversal can be performed sequentially, either by segment number or by the road network topology. Vehicle status collection is performed on each segment to fully capture traffic information for that segment during the current collection period.

[0066] Step S1343-1: Obtain the real-time driving speeds of all vehicles in the road section, and calculate the average driving speed of the vehicles in the road section as the vehicle driving speed of the road section.

[0067] When collecting vehicle status for a particular road section, the real-time speeds of all virtual vehicles within the section during the current collection period are first obtained. These real-time speeds form a multi-dimensional set, encompassing the driving status of every vehicle within the section. The average of these real-time speeds is then calculated and used as the vehicle speed for the section during the current collection period. This average reflects the overall speed level for the section during the current period.

[0068] Step S1343-2: Obtain the distance values ​​between adjacent vehicles in the road section, and calculate the average value of all adjacent vehicle distance values ​​as the vehicle distance of the road section.

[0069] Next, the distances between adjacent vehicles in the same lane within the road section are obtained. These distance values ​​also form a multi-dimensional set. The average of these adjacent vehicle distances is calculated and used as the inter-vehicle distance for the road section during the current collection period. The inter-vehicle distance can, to a certain extent, reflect the density of traffic flow and driving safety.

[0070] Step S1343-3: Count the number of vehicles that change lanes within the road section during the collection period, and use the ratio of the number of lane-changing vehicles to the total number of vehicles in the road section as the lane-changing frequency of the road section.

[0071] Count the number of vehicles that changed lanes on the road section during the current collection period, and also record the total number of vehicles on the road section during that period. Divide the number of vehicles that changed lanes by the total number of vehicles, and the resulting ratio is used as the lane-changing frequency for that road section during the current collection period. The lane-changing frequency reflects the flexibility and stability of traffic flow; a high frequency may indicate instability in the traffic flow.

[0072] Step S1344: Associating the vehicle speed, vehicle spacing, and vehicle lane-changing frequency of each road section during each collection period with the corresponding timestamp and road section identifier to generate traffic operation status data including time and space dimensions.

[0073] After collecting vehicle status data for all road sections during each collection period, the vehicle speed, inter-vehicle spacing, and lane change frequency for each section are associated with the corresponding timestamp and section ID. The section ID uniquely identifies each road section, while the timestamp identifies the collection period. This association generates traffic status data that encompasses both temporal and spatial dimensions, forming a multidimensional dataset. Technicians can query traffic status at different times using timestamps or by using section IDs.

[0074] Step S135: The collected traffic operation status data is subjected to index extraction and processing to calculate a traffic operation efficiency index and a traffic order stability index. The traffic operation efficiency index is calculated based on vehicle speed and road section traffic volume, and the traffic order stability index is calculated based on the standard deviation of vehicle spacing and the fluctuation amplitude of lane change frequency.

[0075] After collecting traffic status data, it is necessary to extract and process indicators to derive traffic efficiency and traffic order stability indicators, which can measure the effectiveness of traffic control strategies. The traffic efficiency indicator primarily reflects the operational efficiency of the traffic system and is calculated comprehensively based on vehicle speed and road section traffic volume. The traffic order stability indicator reflects the stability of traffic flow and is calculated based on the standard deviation of vehicle spacing and the fluctuation range of lane change frequency. These two indicators evaluate the effectiveness of traffic control strategies from different perspectives.

[0076] Step S1351: extracting the vehicle speed and vehicle quantity data of each road section during all collection periods from the traffic operation status data.

[0077] From the generated traffic status data, which includes both temporal and spatial dimensions, we extract the vehicle speeds and the number of vehicles passing through each road section during each collection period. This data is the basis for calculating traffic efficiency indicators, and its accuracy and completeness must be ensured. For example, for a given road section, we extract the set of vehicle speeds and the corresponding number of vehicles for all collection periods.

[0078] Step S1352: Calculate the road section traffic volume of each road section in each collection period, where the road section traffic volume is the number of vehicles passing through the end point of the road section during the collection period.

[0079] For each road segment, during each collection period, the number of vehicles passing through the segment's endpoint is counted and used as the segment's traffic volume for that period. The segment's traffic volume reflects the traffic volume during that period and is an important parameter for measuring a segment's carrying capacity and operating efficiency. For example, if a certain number of vehicles pass through the segment's endpoint during a certain collection period, this number represents the segment's traffic volume for that period.

[0080] Step S1353: averaging the vehicle speeds of the same road section during all collection periods to obtain the average speed of the road section.

[0081] For a given road section, the average speed of vehicles across all collected time periods is calculated. This average speed reflects the overall speed level of the section throughout the simulation, unaffected by speed fluctuations during individual time periods, and better reflects the long-term operating efficiency of the section.

[0082] Step S1354: Sum up the traffic volume of the same road section in all collection periods to obtain the total traffic volume of the road section.

[0083] The total traffic volume of a road section is obtained by adding up the traffic volume of all sampling periods. The total traffic volume of a road section reflects the total traffic flow of the road section during the entire simulation process and reflects the traffic carrying capacity of the road section.

[0084] Step S1355: The product of the average driving speed of the road section and the total traffic volume of the road section is used as the traffic operation efficiency component of the road section, and the weighted average of the traffic operation efficiency components of all road sections is taken to obtain the traffic operation efficiency index of the entire control area. The weight of the weighted average is proportional to the number of lanes in the road section.

[0085] Multiply the average speed of each road section by the total traffic volume of the section to obtain the traffic operation efficiency component of the section. The section traffic operation efficiency component comprehensively considers the driving speed and traffic volume of the section and can better reflect the operation efficiency of a single road section. Then, take the weighted average of the traffic operation efficiency components of all sections, where the weight setting is proportional to the number of lanes of the section. The more lanes a section has, the greater its proportion in the traffic operation of the entire control area, and the higher its weight. Through the above weighted calculation, the traffic operation efficiency index of the entire control area is obtained. This traffic operation efficiency index can comprehensively reflect the traffic operation efficiency of the entire control area under the traffic control strategy to be evaluated.

[0086] Step S1356: Extract the vehicle spacing and vehicle lane-changing frequency data of each road section during all collection periods from the traffic operation status data.

[0087] From traffic status data, extract the set of vehicle spacing and lane-changing frequency data for each road section during each collection period. This data is key to calculating traffic order stability indicators and requires accurate extraction and organization. For example, for a particular road section, extract the set of vehicle spacing and the corresponding lane-changing frequency for all collection periods.

[0088] Step S1357: Calculate the standard deviation of the vehicle distances of each road section in each collection period, where the standard deviation represents the degree of dispersion of the vehicle distances within the road section.

[0089] For each road section, during each collection period, the inter-vehicle standard deviation is calculated based on the extracted inter-vehicle distance set. This standard deviation reflects the degree of dispersion of inter-vehicle distances on the road section during that period. A larger standard deviation indicates a more uneven distribution of inter-vehicle distances and less stable traffic order. Conversely, a smaller standard deviation indicates a more even distribution of inter-vehicle distances and more stable traffic order.

[0090] Step S1358: Calculate the difference between the maximum and minimum lane-changing frequencies of vehicles on each road section during all acquisition periods to obtain the lane-changing frequency fluctuation amplitude.

[0091] For each road section, the maximum and minimum values are found in the lane-changing frequency data of all collection periods, and then the difference between the two is calculated to obtain the lane-changing frequency fluctuation amplitude of the road section. The lane-changing frequency fluctuation amplitude reflects the degree of change of the vehicle lane-changing frequency during the entire simulation process. The greater the fluctuation amplitude, the more unstable the lane-changing behavior, and the greater the disturbance to the traffic order.

[0092] Step S1359: Average the inter-vehicle distance standard deviations of the same road section in all collection periods to obtain the average inter-vehicle distance standard deviation of the road section.

[0093] The inter-vehicle distance standard deviations of the same road section in all collection periods are averaged to obtain the average inter-vehicle distance standard deviation of the road section. The average inter-vehicle distance standard deviation of the road section integrates the inter-vehicle distance dispersion of each period and can reflect the stability of the inter-vehicle distance of the road section during the entire simulation process.

[0094] Step S13510: Add the average inter-vehicle distance standard deviation and the lane-changing frequency fluctuation amplitude to obtain the traffic order stability component of the road section. Take a weighted average of the traffic order stability components of all road sections to obtain the traffic order stability index of the entire control area, and the weight of the weighted average is proportional to the vehicle density of the road section.

[0095] The average inter-vehicle distance standard deviation of each road section is added to the lane-changing frequency fluctuation amplitude of the road section to obtain the traffic order stability component of the road section. The road section traffic order stability component considers the stability of the inter-vehicle distance and the fluctuation of the lane-changing frequency, and can better reflect the traffic order stability of a single road section. Then, a weighted average of the traffic order stability components of all road sections is taken, and the weight is proportional to the vehicle density of the road section. The greater the vehicle density of the road section, the greater the influence of its traffic order on the entire control area, and the higher the weight. Through the above weighted calculation, the traffic order stability index of the entire control area is obtained, which can comprehensively reflect the stability of the traffic order of the entire control area under the to-be-evaluated traffic control strategy.

[0096] Step S136: Associate the traffic operation efficiency index and the traffic order stability index to the corresponding to-be-evaluated traffic control strategy to generate a strategy effect feedback set corresponding to the traffic control strategy.

[0097] The calculated traffic operation efficiency index and traffic order stability index are associated with the corresponding to-be-evaluated traffic control strategy to form a strategy effect feedback set corresponding to the strategy. The strategy effect feedback set completely records the operation effect of the traffic control strategy in the simulation evaluation environment and is a direct basis for evaluating and adjusting the strategy. For example, a to-be-evaluated traffic control strategy is bound with the traffic operation efficiency index and the traffic order stability index obtained through simulation thereof to be stored as a strategy effect feedback set.

[0098] Step S137: The above steps are repeatedly executed until all traffic control strategies in the traffic control strategy set are simulated and evaluated to obtain a strategy effect feedback set corresponding to each traffic control strategy.

[0099] According to the process of steps S131 to S136, each traffic control strategy in the traffic control strategy set is sequentially simulated and evaluated. Each time, an unevaluated strategy is selected, configured into the simulation evaluation environment, the simulation running program is started, the traffic operation state data is collected, the strategy effect index is calculated, and the strategy effect feedback set is generated. The process is repeated until all traffic control strategies are evaluated, and finally a strategy effect feedback set corresponding to each strategy is obtained.

[0100] Step S140: According to the strategy effect feedback set, parameter adjustment processing is performed on each traffic control strategy in the traffic control strategy set to obtain an adjusted traffic control strategy, and the parameter adjustment processing is performed based on the association relationship between the traffic operation efficiency index and the traffic order stability index.

[0101] After obtaining the strategy effect feedback sets corresponding to all traffic control strategies, parameter adjustment processing is performed on each traffic control strategy according to the feedback sets. The basis for parameter adjustment is the association relationship between the traffic operation efficiency index and the traffic order stability index. By analyzing the relationship between the two, the parameters to be adjusted and the adjustment direction and amplitude are determined to enable the traffic control strategy to maintain good traffic order stability while ensuring traffic operation efficiency, and finally an adjusted traffic control strategy is obtained.

[0102] Step S141: A traffic control strategy and its corresponding strategy effect feedback set are selected from the traffic control strategy set, and the traffic operation efficiency index and the traffic order stability index in the strategy effect feedback set are extracted.

[0103] Select a traffic control strategy that requires parameter adjustment from the set of traffic control strategies and extract its corresponding traffic operation efficiency index and traffic order stability index from the strategy effect feedback set. These two indicators are core data for evaluating the effectiveness of the strategy. By analyzing them, we can understand the strategy's performance in terms of traffic operation efficiency and traffic order stability, providing guidance for parameter adjustment. For example, select a traffic control strategy and extract its corresponding traffic operation efficiency index and traffic order stability index.

[0104] Step S142: Analyze the correlation between the traffic operation efficiency index and the traffic order stability index, and determine the collaborative optimization direction between the two. The collaborative optimization direction represents the expected change trend of the traffic operation efficiency index and the traffic order stability index when adjusting the parameters of the traffic control strategy.

[0105] The extracted traffic operation efficiency index and traffic order stability index are analyzed for correlation. The direction of their collaborative optimization can be determined by comparing their numerical change trends or analyzing their mutual influence under different traffic conditions. The collaborative optimization direction clarifies the expected change trend of the traffic operation efficiency index and the traffic order stability index when adjusting the parameters of the traffic control strategy. For example, while improving traffic operation efficiency, maintain or improve traffic order stability, or while ensuring traffic order stability, improve traffic operation efficiency as much as possible. For example, through analysis, it is found that when a certain parameter is adjusted, the traffic operation efficiency index increases while the traffic order stability index can also remain stable. This is an ideal collaborative optimization direction.

[0106] Step S143: Identify target parameters that need to be adjusted in the current traffic control strategy based on the collaborative optimization direction, where the target parameters are at least one of lane control parameters or speed limit control parameters.

[0107] Based on the determined collaborative optimization direction, the system analyzes which parameters in the current traffic control strategy can adjust to drive the traffic efficiency and traffic order stability indicators in the desired direction, thereby identifying the target parameters that need to be adjusted. The target parameters may be lane control parameters, speed limit control parameters, or both. For example, if the collaborative optimization direction is to improve traffic efficiency while maintaining traffic order stability, and analysis reveals that adjusting the number of through lanes on a certain road section (lane control parameters) and appropriately increasing the speed limit on that section (speed limit control parameters) can achieve this goal, then these two parameters are determined as target parameters.

[0108] Step S144: Calculate the adjustment range of the target parameter according to the degree of deviation between the current values ​​of the traffic operation efficiency index and the traffic order stability index and the preset reference threshold value, and the adjustment range is positively correlated with the degree of deviation.

[0109] Pre-set reference thresholds are established based on historical traffic data, traffic management experience, and traffic system design standards. They serve as benchmarks for measuring traffic operation efficiency and traffic order stability. These thresholds include a traffic operation efficiency reference threshold and a traffic order stability reference threshold, corresponding to the evaluation criteria for the traffic operation efficiency and traffic order stability indicators, respectively.

[0110] Step S1441: Obtain a preset reference threshold set, where the preset reference threshold set includes a traffic operation efficiency reference threshold and a traffic order stability reference threshold.

[0111] By calling the preset reference threshold database stored in the system, a set of preset reference thresholds is obtained. The traffic operation efficiency reference threshold in this set is determined based on the statistical values ​​of traffic operation efficiency indicators under the same traffic control scenario during the same historical period, combined with the efficiency targets set by the traffic management department; the traffic order stability reference threshold is formulated based on historical traffic order data and safe operation standards. For example, the traffic operation efficiency reference threshold may be determined by analyzing the traffic operation efficiency indicators of the control area during multiple morning rush hours in the past and taking the middle value of a reasonable range; the traffic order stability reference threshold may be set according to the design standards and safe driving requirements of the roads in the area to ensure that the spacing and lane changing behaviors of vehicles during driving are within a safe and stable range.

[0112] Step S1442: Calculate the difference between the current value of the traffic operation efficiency index and the traffic operation efficiency reference threshold to obtain the efficiency deviation value. If the efficiency deviation value is positive, it means that the current traffic operation efficiency is better than the reference level. If it is negative, it means that the current traffic operation efficiency is lower than the reference level.

[0113] The difference between the current value of the traffic operation efficiency index calculated in step S1355 and the traffic operation efficiency reference threshold obtained in step S1441 is calculated to obtain the efficiency deviation value. The calculation method is to subtract the traffic operation efficiency reference threshold from the current value of the traffic operation efficiency index. For example, if the current traffic operation efficiency index is a multidimensional value set, and the traffic operation efficiency reference threshold is a corresponding multidimensional reference value set, then the efficiency deviation value is a new multidimensional set formed by subtracting the corresponding elements in the two multidimensional sets. When an element in the efficiency deviation value is positive, it means that the current traffic operation efficiency in this dimension is better than the reference level; when the element is negative, it means that the current traffic operation efficiency in this dimension is lower than the reference level.

[0114] Step S1443: Calculate the difference between the current value of the traffic order stability index and the traffic order stability reference threshold to obtain the order deviation value. If the order deviation value is positive, it means that the current traffic order stability is better than the reference level. If it is negative, it means that the current traffic order stability is lower than the reference level.

[0115] Similarly, the current value of the traffic order stability index calculated in step S13510 is subtracted from the traffic order stability reference threshold obtained in step S1441 to obtain the order deviation value. The calculation method is to subtract the traffic order stability reference threshold from the current value of the traffic order stability index. Similarly, the traffic order stability index and the reference threshold are both multi-dimensional value sets, and the order deviation value is a multi-dimensional set formed by subtracting the corresponding elements. When an element in the order deviation value is positive, it indicates that the current traffic order stability in this dimension is better than the reference level; when it is negative, it indicates that the current traffic order stability in this dimension is lower than the reference level.

[0116] Step S1444: Determine the deviation level according to the sign combination of the efficiency deviation value and the order deviation value, and query a preset adjustment range mapping table based on the deviation level to obtain a basic adjustment range corresponding to the deviation level.

[0117] Analyze the sign combination of each element in the efficiency deviation value and the order deviation value to determine the overall deviation degree level. The deviation degree level is divided into multiple levels. For example, when the efficiency deviation value and the order deviation value are both negative and the absolute value is large, it belongs to a higher deviation degree level; when both are positive and the absolute value is small, it belongs to a lower deviation degree level. The preset adjustment range mapping table stores the correspondence between different deviation degree levels and the corresponding basic adjustment ranges. The basic adjustment range is a multi-dimensional numerical set that corresponds to the adjustment benchmark amount of each parameter in the target parameter. For example, the higher the deviation degree level, the larger the corresponding basic adjustment range, so as to adjust the target parameters to a greater extent to improve traffic conditions. By querying the mapping table, the corresponding basic adjustment range is obtained according to the determined deviation degree level.

[0118] Step S1445: Perform weighted correction on the basic adjustment range according to the ratio of the absolute values ​​of the efficiency deviation value and the order deviation value. If the absolute value of the efficiency deviation value is greater than the absolute value of the order deviation value, the adjustment weight of the parameters related to traffic operation efficiency in the target parameters is increased; otherwise, the adjustment weight of the parameters related to traffic order stability is increased.

[0119] Calculate the absolute values ​​of the efficiency deviation and the order deviation, and then determine the ratio of their absolute values. For example, calculate the ratio of the sum of the absolute values ​​of the efficiency deviation elements to the sum of the absolute values ​​of the order deviation elements. Based on this ratio, weighted adjustments are made to the basic adjustment range. If the absolute sum of the efficiency deviations is greater than the absolute sum of the order deviations, this indicates a more pronounced problem with traffic efficiency. Therefore, the adjustment weights for target parameters related to traffic efficiency (such as the portion of speed limit control parameters that affects driving speed) are increased by multiplying the basic adjustment range for these parameters by a correction factor greater than 1. Conversely, if the absolute sum of the order deviations is greater, increase the adjustment weights for parameters related to traffic order stability (such as the portion of lane control parameters that affects vehicle spacing and lane changes). This weighted adjustment ensures that the adjustment range more closely reflects the primary issues facing the current traffic situation.

[0120] Step S1446: Determine the corrected adjustment range as the final adjustment range of the target parameter, and the final adjustment range is used to guide the modification process of the target parameter.

[0121] The final adjustment of the target parameter is the adjustment amplitude after weighted correction. This adjustment amplitude is a multi-dimensional set of numerical values ​​that corresponds one-to-one with the dimensions of the target parameter. The value of each element represents the adjustment amount for the corresponding target parameter. The positive or negative sign of the adjustment indicates the direction of the adjustment, and the absolute value indicates the magnitude of the adjustment. For example, if the target parameters are the number of through lanes and the speed limit on a certain road section, a positive element corresponding to the number of through lanes in the final adjustment amplitude indicates an increase in the number of through lanes, and a positive element corresponding to the speed limit indicates an increase in the speed limit. The specific numerical values ​​are determined based on the corrected results.

[0122] Step S145: modify the target parameters according to the calculated adjustment range to generate an intermediate traffic control strategy after parameter adjustment.

[0123] According to the final adjustment range determined in step S1446, the target parameters identified in step S143 are modified. During the modification process, it is necessary to ensure that the modification of the target parameters complies with the road design specifications and traffic management requirements. For example, for the number of through lanes in the lane control parameters, increase or decrease according to the corresponding value in the final adjustment range. If the adjustment range is an increase, the corresponding number of through lanes is added on the basis of the original lane division, and the number of other types of lanes is adjusted to keep the total number of lanes unchanged; for the speed limit control parameters, increase or decrease the speed limit value according to the corresponding value in the final adjustment range to ensure that the new speed limit value is within the safety range allowed by the road. After the modification is completed, the intermediate traffic control strategy after the parameter adjustment is generated.

[0124] Step S146: Input the intermediate traffic control strategy into the simulation evaluation environment for secondary simulation evaluation, and collect the new traffic operation efficiency index and new traffic order stability index obtained from the secondary simulation evaluation.

[0125] The generated intermediate traffic control strategy is input into the simulation evaluation environment, and a secondary simulation evaluation is performed according to the process from steps S133 to S135. Specifically, the simulation program is started, traffic status data is collected at preset intervals, and new traffic efficiency and traffic order stability indicators are calculated. The purpose of the secondary simulation evaluation is to verify whether the adjusted strategy achieves the expected results. The simulation run time, data collection interval, and other settings remain consistent with the initial evaluation to ensure comparability of the evaluation results.

[0126] Step S147: Determine whether the new traffic operation efficiency index and the new traffic order stability index are both better than the indexes before adjustment. If so, determine the intermediate traffic control strategy as the adjusted traffic control strategy.

[0127] The new traffic efficiency and traffic order stability indicators obtained from the secondary simulation are compared with their pre-adjustment counterparts. Each element of the multi-dimensional indicators is evaluated separately. If all elements of the new indicators are superior to the pre-adjustment indicators (i.e., all elements of the new traffic efficiency indicators are greater than or equal to those of the pre-adjustment indicators, and all elements of the new traffic order stability indicators are greater than or equal to those of the pre-adjustment indicators), then the intermediate traffic control strategy has improved or maintained stability in both traffic efficiency and traffic order stability. In this case, the intermediate traffic control strategy is determined as the adjusted traffic control strategy.

[0128] Step S148: If not, recalculate the adjustment range and repeat the parameter modification and secondary simulation evaluation steps until an adjusted traffic control strategy that meets the conditions is obtained.

[0129] If at least one element in the new traffic efficiency index and the new traffic order stability index is not better than the index before adjustment, it means that the current adjustment range may be inappropriate and needs to be recalculated. When recalculating, return to step S144, reanalyze the degree of deviation based on the new evaluation results, adjust the weighted correction coefficient, and obtain a new final adjustment range. Then, according to the process of steps S145 to S147, re-modify the parameters and conduct a second simulation evaluation. Repeat this process until the new traffic efficiency index and the new traffic order stability index are both better than the index before adjustment. At this time, the corresponding intermediate traffic control strategy is the adjusted traffic control strategy.

[0130] Step S150: Output the adjusted traffic control strategy to the traffic control system to update the traffic control strategy currently executed by the traffic control system.

[0131] After obtaining the adjusted traffic control strategy, it is output to the actual traffic control system. Upon receiving the strategy, the traffic control system updates its currently executed traffic control strategy to manage traffic in the control area according to the adjusted strategy. For example, the traffic control system resets lane indicators on the road based on the adjusted lane control parameters and updates speed limit signs on the road based on the adjusted speed limit control parameters. The new control strategy is then implemented through the traffic light control system, traffic guidance system, and other systems to optimize traffic flow and improve traffic operation efficiency and traffic order stability.

[0132] In addition, in order to ensure that the reinforcement learning model can continuously adapt to changes in traffic conditions, this method also includes the steps of continuously training and optimizing the reinforcement learning model.

[0133] Step S210: Collecting real traffic operation data after the actual traffic control system implements the adjusted traffic control strategy.

[0134] After the traffic control system has updated its strategy and been running for a period of time, real-world traffic data from that period is collected. This data includes vehicle speeds, vehicle spacing, lane change frequency, and road volume for each road section, as collected by monitoring equipment. This data is consistent with the types of data collected in the simulation evaluation environment. When collecting real-world data, privacy protection technologies must be employed to anonymize sensitive data such as vehicle identification and driver information. For example, license plate information must be removed and driver profile data must be desensitized to ensure that data collection and use comply with relevant laws and regulations on data security and privacy protection.

[0135] Step S220: comparing the actual traffic operation data with the simulated traffic operation data under the corresponding strategy in the simulation evaluation environment, and calculating the deviation value between the two.

[0136] The collected real-world traffic data is compared with simulated traffic data obtained from a secondary simulation of the adjusted traffic control strategy in a simulation evaluation environment. During this comparison, each data point for the same road section and time interval is compared individually, and the deviation between the two is calculated. For example, for the vehicle speed on a certain road section during a certain collection period, the difference between the real data and the simulated data is the speed deviation for that section during that period. The deviation values ​​for all road sections and all collection periods are combined to form a comprehensive deviation value set.

[0137] Step S230: adjusting the scenario parameter configuration of the simulation evaluation environment based on the deviation value, so that the simulation result of the simulation evaluation environment is closer to the actual traffic condition.

[0138] According to the calculated deviation value set, analyze which scene parameter settings in the simulation evaluation environment deviate from the actual situation, and then adjust these scene parameters. For example, if the vehicle driving speed of a certain road segment in the real data is generally lower than the simulation data, it may be that the virtual road resistance coefficient of the road segment is set too low, at which time the virtual road resistance coefficient of the road segment needs to be appropriately increased, so that the driving speed of the simulation vehicle on the road segment is closer to the real situation; if the simulation data and the real data of the vehicle lane changing frequency deviate greatly, it may be that the parameter settings of the vehicle lane changing decision model in the simulation evaluation environment are unreasonable, and the lane changing threshold and other parameters need to be adjusted. By continuously adjusting the scene parameters, the authenticity and accuracy of the simulation evaluation environment are improved.

[0139] Step S240: retrain the reinforcement learning model using the adjusted simulation evaluation environment and the new traffic control strategy set.

[0140] When the scene parameter adjustment of the simulation evaluation environment is completed, a new traffic control strategy set is constructed, which can be a new strategy generated by mutation and combination based on the original strategy. The adjusted simulation evaluation environment and the new traffic control strategy set are used as training data to retrain the reinforcement learning model. During training, the reinforcement learning model interacts with the adjusted simulation evaluation environment, continuously learns the traffic operation law under the new strategy, and updates the decision parameters of the model. During training, appropriate training rounds and learning rates are set to ensure that the model converges to a better state, improving the adaptability to complex traffic conditions and the accuracy of strategy evaluation.

[0141] Step S250: replace the original reinforcement learning model with the retrained reinforcement learning model for subsequent traffic control strategy evaluation and parameter adjustment.

[0142] When the reinforcement learning model is retrained and reaches the preset performance indicators, it is replaced with the original reinforcement learning model for subsequent simulation evaluation and parameter adjustment of the traffic control strategy set. By periodically retraining the reinforcement learning model, it can continuously adapt to changes in the traffic control scene, such as road construction, traffic growth, and implementation of new traffic rules, ensuring that the traffic control strategy based on the model can always effectively optimize traffic operation conditions.

[0143] Through the above steps, the method realizes adaptive adjustment of traffic control strategies based on simulation feedback, can continuously optimize traffic control strategies according to the actual situation of traffic operation and simulation evaluation results, improve the operation efficiency and stability of the traffic system, and through continuous training of the reinforcement learning model, enhance the adaptability and robustness of the method, so that the traffic control strategy can better cope with complex and variable traffic conditions.

[0144] Figure 2A schematic diagram illustrates exemplary hardware and software components of a traffic control policy adaptive system 100 based on simulation feedback, which can implement the concepts of the present application, as provided in some embodiments of the present application. For example, the processor 120 can be used in the traffic control policy adaptive system 100 based on simulation feedback and perform the functions described in the present application.

[0145] The traffic control policy adaptation system 100 based on simulation feedback can be a general-purpose server or a special-purpose server, both of which can be used to implement the traffic control policy adaptation method based on simulation feedback of the present application. Although only one server is shown in this application, for convenience, the functions described in this application can be implemented in a distributed manner on multiple similar platforms to balance the processing load.

[0146] For example, the traffic control strategy adaptive system 100 based on simulation feedback may include a network port 110 connected to a network, one or more processors 120 for executing program instructions, a communication bus 130, and storage media 140 in different forms, such as a disk, ROM, or RAM, or any combination thereof. Exemplarily, the traffic control strategy adaptive system 100 based on simulation feedback may also include program instructions stored in ROM, RAM, or other types of non-transitory storage media, or any combination thereof. The method of the present application can be implemented according to these program instructions. The traffic control strategy adaptive system 100 based on simulation feedback also includes an I / O interface 150 between the computer and other input and output devices.

[0147] For ease of explanation, only one processor is described in the traffic control strategy adaptive system based on simulation feedback 100. However, it should be noted that the traffic control strategy adaptive system based on simulation feedback 100 in the present application may also include multiple processors, so the steps performed by one processor described in the present application may also be performed jointly or individually by multiple processors. For example, if the processor of the traffic control strategy adaptive system based on simulation feedback 100 executes step A and step B, it should be understood that step A and step B may also be performed jointly by two different processors or individually in one processor. For example, the first processor executes step A, the second processor executes step B, or the first processor and the second processor execute steps A and B together.

[0148] In addition, an embodiment of the present invention further provides a readable storage medium, in which computer-executable instructions are preset. When a processor executes the computer-executable instructions, the above-mentioned traffic control strategy adaptive method based on simulation feedback is implemented.

[0149] It should be noted that in order to simplify the description of the present invention and thus help understand one or more embodiments of the invention, in the foregoing description of the embodiments of the present invention, multiple features are sometimes combined into one embodiment, figure or description thereof.

Claims

1. A traffic control strategy adaptive method based on simulation feedback, characterized in that: The method comprises: Acquire traffic control scenario data, the traffic control scenario data including traffic flow data and road condition information, the traffic flow data representing the flow state of vehicles within the control area, and the road condition information representing the physical properties and sign configuration state of roads within the control area; Constructing a simulation evaluation environment, wherein the simulation evaluation environment is used to simulate traffic operation status under different traffic control strategies, and the simulation evaluation environment configures scenario parameters based on traffic flow data and road condition information in the traffic control scenario data; Calling a pre-trained reinforcement learning model to perform simulation evaluation processing on each traffic control strategy in the traffic control strategy set in the simulation evaluation environment, and generating a strategy effect feedback set corresponding to each traffic control strategy, wherein the strategy effect feedback set includes a traffic operation efficiency index and a traffic order stability index; performing parameter adjustment processing on each traffic control strategy in the traffic control strategy set according to the strategy effect feedback set to obtain an adjusted traffic control strategy, wherein the parameter adjustment processing is performed based on a correlation between a traffic operation efficiency index and a traffic order stability index; Outputting the adjusted traffic control strategy to the traffic control system to update the traffic control strategy currently executed by the traffic control system; The construction of the simulation evaluation environment includes: Determining the road network structure parameters of the simulation evaluation environment based on the road condition information in the traffic control scenario data, wherein the road network structure parameters include the connection relationship of the roads, the number of lanes, and the lane division method; Determining initial traffic flow parameters of the simulation evaluation environment based on traffic flow data in the traffic control scenario data, wherein the initial traffic flow parameters include an initial vehicle density and an initial vehicle speed distribution in each control sub-area; Inputting the road network structure parameters and initial traffic flow parameters into a scenario modeling module to generate a basic scenario model of a simulation evaluation environment, wherein the basic scenario model includes a virtual road network and an initial vehicle distribution state consistent with the geographical structure of the control area; According to the road marking type and speed limit sign configuration status in the traffic control scene data, the attributes of the road elements in the basic scene model are configured so that the road markings and speed limit signs in the virtual road network are consistent with the road condition information of the actual control area; Boundary conditions are set for the configured basic scenario model to determine the vehicle entry and exit boundaries in the simulation evaluation environment. The entry and exit boundaries correspond to the actual entry and exit locations of the control area, generating a simulation evaluation environment that can be used to simulate traffic operation status.

2. The traffic control strategy adaptive method based on simulation feedback according to claim 1 is characterized in that: The obtaining of traffic control scene data includes: Divide the control area into sub-areas to obtain multiple control sub-areas, each of which corresponds to an independent geographical boundary range; Collecting traffic flow data within each controlled sub-area, wherein the traffic flow data includes the number of vehicles entering and leaving the controlled sub-area per unit time; Collecting traffic information within each controlled sub-area, including the number of lanes, lane division method, road marking type, and speed limit sign configuration status; The traffic flow data and road condition information within the same control sub-area are timestamp aligned, and the traffic flow data and road condition information of each control sub-area after timestamp alignment are integrated to generate traffic control scenario data covering the entire control area. The data of each control sub-area in the traffic control scenario data are associated through geographical boundary ranges.

3. The traffic control strategy adaptive method based on simulation feedback according to claim 1 is characterized in that: Determining the road network structure parameters of the simulation evaluation environment based on the road condition information in the traffic control scenario data includes: Analyze the road condition information in the traffic control scene data and extract the physical attribute data of the roads in each control sub-area. The physical attribute data includes the starting point coordinates, end point coordinates, road width and road type of the road; Determine the spatial position relationship of each road within the control area based on the starting and ending coordinates of the road, and construct a road topology diagram, where nodes in the road topology diagram represent road intersections and edges represent road segments connecting intersections; Determine the connection relationship of the roads based on the road topology relationship graph, wherein the connection relationship indicates whether there are directly connected road segments between any two intersections and the number of road segments; According to the lane quantity data in the road condition information, the lane quantity attribute is configured for each road segment in the road topology diagram; Analyze lane division data in the traffic information to determine the functional division type of lanes in each road segment, where the functional division types include through lanes, left-turn lanes, right-turn lanes, and mixed lanes; The road connection relationship, lane number and lane division method are integrated into road network structure parameters, which are used to define the basic framework of the virtual road network in the simulation evaluation environment.

4. The traffic control strategy adaptive method based on simulation feedback according to claim 1 is characterized in that: The calling of the pre-trained reinforcement learning model performs simulation evaluation processing on each traffic control strategy in the traffic control strategy set in the simulation evaluation environment to generate a strategy effect feedback set corresponding to each traffic control strategy, including: Selecting a traffic control strategy to be evaluated from a set of traffic control strategies, wherein the traffic control strategy includes a lane control parameter and a speed limit control parameter, wherein the lane control parameter represents a lane usage allocation method, and the speed limit control parameter represents a driving speed limit value for different road sections; Inputting the traffic control strategy to be evaluated into the strategy execution module of the reinforcement learning model, and the strategy execution module configuring the lane control parameters and the speed limit control parameters to the corresponding road section of the simulation evaluation environment; Start the simulation running program of the simulation evaluation environment, so that the simulation evaluation environment simulates the traffic flow operation process under the condition that the traffic control strategy to be evaluated is configured. The duration of the simulation operation process is consistent with the collection period of the traffic control scenario data; During the simulation operation, traffic operation status data in the simulation evaluation environment is collected at preset time intervals. The traffic operation status data includes vehicle speed, vehicle spacing and vehicle lane change frequency in each road section; Performing index extraction and processing on the collected traffic operation status data to calculate the traffic operation efficiency index and the traffic order stability index. The traffic operation efficiency index is calculated based on vehicle speed and road section traffic volume, and the traffic order stability index is calculated based on the standard deviation of vehicle spacing and the fluctuation amplitude of lane change frequency. Associating the traffic operation efficiency index and the traffic order stability index with the corresponding traffic control strategy to be evaluated, and generating a strategy effect feedback set corresponding to the traffic control strategy; Repeat the above steps until all traffic control strategies in the traffic control strategy set have completed simulation evaluation, and obtain a strategy effect feedback set corresponding to each traffic control strategy.

5. The traffic control strategy adaptive method based on simulation feedback according to claim 4 is characterized in that: The index extraction and processing of the collected traffic operation status data to calculate the traffic operation efficiency index and the traffic order stability index include: Extract the vehicle speed and vehicle quantity data of each road section during all collection periods from the traffic operation status data; Calculate the road section traffic volume of each road section in each collection period, where the road section traffic volume is the number of vehicles passing through the end point of the road section during the collection period; The average speed of vehicles on the same road section during all sampling periods is taken to obtain the average speed of the road section; The total traffic volume of the same road section during all collection periods is summed up to obtain the total traffic volume of the road section; The product of the average speed of a road section and the total traffic volume of the road section is used as the traffic operation efficiency component of the road section. The weighted average of the traffic operation efficiency components of all road sections is taken to obtain the traffic operation efficiency index of the entire control area. The weight of the weighted average is proportional to the number of lanes in the road section. Extract the vehicle spacing and lane-changing frequency data of each road section during all collection periods from the traffic operation status data; Calculating the standard deviation of vehicle spacing for each road section during each collection period, wherein the standard deviation of vehicle spacing represents the degree of dispersion of vehicle spacing within the road section; Calculate the difference between the maximum and minimum lane-changing frequencies of vehicles on each road section during all sampling periods to obtain the lane-changing frequency fluctuation amplitude; The average standard deviation of vehicle spacing on the same road section during all sampling periods is taken to obtain the average standard deviation of vehicle spacing on the road section; The sum of the standard deviation of the average spacing between road sections and the fluctuation amplitude of the lane-changing frequency is taken as the traffic order stability component of the road section. The weighted average of the traffic order stability components of all road sections is taken to obtain the traffic order stability index of the entire control area. The weight of the weighted average is proportional to the vehicle density of the road section.

6. The traffic control strategy adaptive method based on simulation feedback according to claim 4 is characterized in that: During the simulation operation, the traffic operation status data in the simulation evaluation environment is collected at preset time intervals, including: After the simulation running program of the simulation evaluation environment is started, a time collection window is set, wherein the starting time of the time collection window is the simulation running start time, and the ending time is the simulation running end time; The time collection window is divided into multiple collection periods according to the preset time interval, and each collection period corresponds to a timestamp; During each collection period, all road sections in the simulation evaluation environment are traversed, and the vehicle status collection operation is performed on each road section; The vehicle status collection operation includes: Obtain the real-time driving speed of all vehicles in the road section, and calculate the average driving speed of the vehicles in the road section as the vehicle driving speed of the road section; Obtain the distance values ​​between adjacent vehicles on the road section, and calculate the average value of all adjacent vehicle distance values ​​as the vehicle spacing of the road section; The number of vehicles that changed lanes within the road section during the collection period was counted, and the ratio of the number of lane-changing vehicles to the total number of vehicles in the road section was used as the lane-changing frequency of the road section. The vehicle speed, vehicle spacing, and lane-changing frequency of each road section during each collection period are associated with the corresponding timestamp and road section identifier to generate traffic operation status data containing both time and space dimensions.

7. The traffic control strategy adaptive method based on simulation feedback according to claim 1 is characterized in that: The step of performing parameter adjustment processing on each traffic control strategy in the traffic control strategy set according to the strategy effect feedback set to obtain an adjusted traffic control strategy includes: Select a traffic control strategy and its corresponding strategy effect feedback set from the traffic control strategy set, and extract the traffic operation efficiency index and traffic order stability index in the strategy effect feedback set; Analyze the correlation between traffic operation efficiency indicators and traffic order stability indicators, and determine the direction of collaborative optimization between the two. The collaborative optimization direction represents the expected change trend of traffic operation efficiency indicators and traffic order stability indicators when adjusting the parameters of the traffic control strategy; Identifying a target parameter that needs to be adjusted in the current traffic control strategy based on the collaborative optimization direction, where the target parameter is at least one of a lane control parameter or a speed limit control parameter; Calculating the adjustment range of the target parameter based on the degree of deviation between the current values ​​of the traffic operation efficiency index and the traffic order stability index and the preset reference threshold, wherein the adjustment range is positively correlated with the degree of deviation; Modify the target parameters according to the calculated adjustment range to generate an intermediate traffic control strategy after parameter adjustment; Input the intermediate traffic control strategy into the simulation evaluation environment for secondary simulation evaluation, and collect the new traffic operation efficiency index and new traffic order stability index obtained from the secondary simulation evaluation; Determine whether the new traffic operation efficiency index and the new traffic order stability index are both better than the indexes before adjustment. If so, determine the intermediate traffic control strategy as the adjusted traffic control strategy; If not, the adjustment range is recalculated and the parameter modification and secondary simulation evaluation steps are repeated until an adjusted traffic control strategy that meets the conditions is obtained.

8. The traffic control strategy adaptive method based on simulation feedback according to claim 7 is characterized in that: The calculation of the adjustment range of the target parameter according to the degree of deviation between the current values ​​of the traffic operation efficiency index and the traffic order stability index and the preset reference threshold value includes: Obtaining a preset reference threshold set, wherein the preset reference threshold set includes a traffic operation efficiency reference threshold and a traffic order stability reference threshold; Calculate the difference between the current value of the traffic operation efficiency index and the traffic operation efficiency reference threshold to obtain the efficiency deviation value. If the efficiency deviation value is positive, it means that the current traffic operation efficiency is better than the reference level. If it is negative, it means that the current traffic operation efficiency is lower than the reference level. Calculate the difference between the current value of the traffic order stability index and the traffic order stability reference threshold to obtain the order deviation value. If the order deviation value is positive, it means that the current traffic order stability is better than the reference level. If it is negative, it means that the current traffic order stability is lower than the reference level. Determine the deviation level according to the sign combination of the efficiency deviation value and the order deviation value, and query the preset adjustment range mapping table based on the deviation level to obtain the basic adjustment range corresponding to the deviation level; The basic adjustment range is weighted and corrected according to the ratio of the absolute values ​​of the efficiency deviation value and the order deviation value. If the absolute value of the efficiency deviation value is greater than the absolute value of the order deviation value, the adjustment weight of the parameters related to traffic operation efficiency in the target parameters is increased; otherwise, the adjustment weight of the parameters related to traffic order stability is increased; The corrected adjustment range is determined as the final adjustment range of the target parameter, and the final adjustment range is used to guide the modification process of the target parameter.

9. A traffic control strategy adaptive system based on simulation feedback, characterized in that: It includes a processor and a memory, the memory is connected to the processor, the memory is used to store programs, instructions or codes, and the processor is used to execute the programs, instructions or codes in the memory to implement the traffic control strategy adaptive method based on simulation feedback as described in any one of claims 1 to 8.

Images