Road shunting suggestion generation method and system based on road network safety research and judgment

Abstract

The invention belongs to the field of traffic control, and relates to a road network safety research and judgment-based road diversion suggestion generation method and system, and the method comprises the steps: collecting traffic data through various types of sensors, and carrying out the processing of the traffic data, and obtaining scene-based traffic data; the scenarized traffic data comprises shunting area data and converging area data; road network safety research and judgment are carried out on the scenarized traffic data, and a road network safety research and judgment result is obtained; the road network safety research and judgment result comprises a diversion area research and judgment result and a confluence area research and judgment result; performing scene-based shunting processing based on the road network safety research and judgment result to obtain a shunting induction strategy; the shunting induction strategy comprises a shunting area induction strategy and a converging area induction strategy; executing a shunting induction strategy through shunting equipment, and continuously optimizing a road network safety research and judgment result based on a shunting effect; the scene adaptability is improved, and the relevance between shunting suggestions and actual safety requirements is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of traffic control, and specifically discloses a road diversion suggestion generation method and system based on road network safety research and judgment. BACKGROUND

[0002] The current road network diversion technology has three major scene-based pain points: first, the low-visibility diversion area is missing. In scenes such as heavy fog, rain and snow, and night, the lane line and contour are blurred. Traditional diversion relies only on manual instructions, and there is no special intelligent induction device to strengthen the lane contour. This leads to a lack of timely preparation for lane changes by vehicles, and an increased risk of driving off the road boundary. Second, there is a lack of accident prediction in the merging area. The speed / flow difference between the main line and ramp vehicles in the merging area can easily cause traffic flow disorder. It is difficult to accurately predict the possibility of merging accidents, and it is only possible to rely on post-event disposal, which cannot control the vehicle speed and entry time in advance. Finally, the induction is not coordinated. The induction mode is not linked with real-time / expected traffic flow, which leads to a disconnection between the diversion suggestion and the actual safety needs.

[0003] Therefore, the present application provides a road diversion suggestion generation method and system based on road network safety research and judgment, which improves the scene adaptability and the relevance of the diversion suggestion and the actual safety needs. SUMMARY

[0004] The present application aims to provide a road diversion suggestion generation method based on road network safety research and judgment, which solves the problem of how to optimize the induction strategy in the low-visibility scene of the diversion area, based on real-time traffic flow data, to induce vehicles and avoid safety accidents. The specific scheme is as follows: A road diversion suggestion generation method based on road network safety research and judgment, comprising: collecting traffic data through multiple types of sensors and processing the traffic data to obtain scene-based traffic data; the scene-based traffic data includes diversion area data and merging area data; performing road network safety research and judgment on the scene-based traffic data to obtain road network safety research and judgment results; the road network safety research and judgment results include diversion area research and judgment results and merging area research and judgment results; performing scene-based diversion processing based on the road network safety research and judgment results to obtain a diversion induction strategy; the diversion induction strategy includes a diversion area induction strategy and a merging area induction strategy; and executing the diversion induction strategy through a diversion device and continuously optimizing the road network safety research and judgment results based on the diversion effect.

[0005] Further, the diverging area data includes basic traffic data, environmental data, and device state data: the basic traffic data includes traffic volume, vehicle speed, and vehicle type distribution; the environmental data includes visibility and weather type; the device state data includes the operating state of the induction light, variable information board, and broadcast; the merging data includes road condition data and time data: the road condition data includes merging lane traffic volume, gateway traffic volume, main line vehicle speed, gateway vehicle speed, and lane occupancy rate; the event data includes the occurrence frequency and duration of merging conflict events.

[0006] Further, the processing of the traffic data to obtain scenario-based traffic data includes: data cleaning of the traffic data to eliminate abnormal data to obtain preprocessed traffic data; adding a scenario label to the preprocessed traffic data to obtain scenario-based traffic data; the scenario label includes a diverging and merging label and a weather label; the diverging and merging label is related to the diverging and merging effect of the road; the weather label is related to the visibility.

[0007] Further, obtaining the diverging area research and judgment result includes: based on the early warning events in the diverging area data, performing multi-dimensional statistical analysis to obtain an event occurrence frequency heat map; the early warning events are related to the visibility and traffic volume; matching the current visibility and current traffic volume to a plurality of diverging safety levels to obtain a current diverging safety level; based on the current diverging safety level, constructing the diverging area research and judgment result; obtaining the merging area research and judgment result includes: using a random forest model to process the merging area data to obtain a merging accident prediction probability; matching the current merging accident prediction probability and the current lane occupancy rate to a plurality of merging safety levels to obtain a current merging safety level; based on the current merging safety level, constructing the merging area research and judgment result.

[0008] Further, the diverging device executes a diverging induction strategy, and the road network safety research and judgment result is continuously optimized based on the diverging effect, including: obtaining a diverging suggestion execution result; the diverging suggestion execution result includes the diverging device operating state, the actual diverging vehicle number, and the driver feedback rate; based on the diverging suggestion execution result, calculating safety indicators and efficiency indicators; based on the safety indicators and efficiency indicators, optimizing a plurality of diverging safety level intervals and / or a plurality of merging safety level intervals; based on the new plurality of diverging safety levels and / or the new plurality of merging safety levels, performing road network safety research and judgment.

[0009] The application also provides a road diversion suggestion generation system based on road network safety research and judgment, comprising a scene modeling module, a research and judgment module, a measurement generation module and an optimization module; the scene modeling module is used for collecting traffic data through various types of sensors and processing the traffic data to obtain scene-based traffic data; the scene-based traffic data comprises diversion area data and merging area data; the research and judgment module is used for conducting road network safety research and judgment on the scene-based traffic data to obtain road network safety research and judgment results; the road network safety research and judgment results comprise diversion area research and judgment results and merging area research and judgment results; the measurement generation module is used for conducting scene-based diversion processing based on the road network safety research and judgment results to obtain diversion induction strategies; the diversion induction strategies comprise diversion area induction strategies and merging area induction strategies; and the optimization module is used for executing the diversion induction strategies through diversion devices and continuously optimizing the road network safety research and judgment results based on the diversion effects.

[0010] Further, the diversion area data comprises basic traffic data, environmental data and device state data; the basic traffic data comprises traffic volume, vehicle speed and vehicle type distribution; the environmental data comprises visibility and weather type; the device state data comprises the operating states of induction lights, variable information boards and broadcasts; the merging data comprises road condition data and time data; the road condition data comprises merging lane traffic volume, gateway traffic volume, main line vehicle speed, gateway vehicle speed and lane occupancy rate; and the event data comprises the occurrence frequency and duration of lane merging conflict events.

[0011] Further, the scene modeling module comprises a data processing unit and a label adding unit; the data processing unit is used for performing data cleaning on the traffic data to remove abnormal data and obtain preprocessed traffic data; and the label adding unit is used for adding scene labels to the preprocessed traffic data to obtain scene-based traffic data; the scene labels comprise diversion and merging labels and weather labels; the diversion and merging labels are related to the diversion and merging effects of roads; and the weather labels are related to visibility.

[0012] Further, the research and judgment module comprises a diversion area research and judgment unit and a merging area research and judgment unit; the diversion area research and judgment unit is used for obtaining diversion area research and judgment results, comprising: based on pre-warning events in the diversion area data, performing multi-dimensional statistical analysis to obtain an event occurrence frequency heat map; the pre-warning events are related to visibility and traffic volume; matching the current visibility and the current traffic volume to a plurality of diversion safety levels to obtain a current diversion safety level; and based on the current diversion safety level, constructing the diversion area research and judgment results; and the merging area research and judgment unit is used for obtaining merging area research and judgment results, comprising: using a random forest model to process the merging area data to obtain a lane merging accident prediction probability; matching the current lane merging accident prediction probability and the current lane occupancy rate to a plurality of merging safety levels to obtain a current merging safety level; and based on the current merging safety level, constructing the merging area research and judgment results.

[0013] Further, the optimization module comprises an execution result obtaining unit, an index calculating unit, an optimization unit and a circulation unit; the execution result obtaining unit is configured to obtain a shunting suggestion execution result; the shunting suggestion execution result comprises a shunting device running state, an actual shunting vehicle number and a driver feedback rate; the index calculating unit is configured to calculate a safety index and an efficiency index based on the shunting suggestion execution result; the optimization unit is configured to optimize a plurality of shunting safety level intervals and / or a plurality of merging safety level intervals based on the safety index and the efficiency index; and the circulation unit is configured to make a road network safety research and judgment based on the new plurality of shunting safety levels and / or the new plurality of merging safety levels.

[0014] The present application has the following advantages and beneficial effects: The present application shortens the vehicle lane changing preparation time and reduces the probability of driving out of the road boundary in a low visibility scenario through intelligent shunting device control.

[0015] The present application can improve the accuracy of pre-judgment merging accident, shorten the accident response time, and improve the merging area traffic efficiency. BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1 FIG. 1 is an exemplary flowchart of a road shunting suggestion generation method based on road network safety research and judgment according to the present application; Figure 2 FIG. 2 is an interface display diagram of a GIS map according to the present application; Figure 3 FIG. 3 is a self-defined schematic diagram of a pre-warning logic in a shunting area induction strategy according to the present application; Figure 4 FIG. 4 is a self-defined schematic diagram of a pre-warning logic in a merging area induction strategy according to the present application. DETAILED DESCRIPTION

[0017] In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are some embodiments of the present application, but not all embodiments of the present application. The components of the embodiments of the present application described and shown in the drawings herein can be arranged and designed in various different configurations.

[0018] Figure 1 FIG. 1 is an exemplary flowchart of a road shunting suggestion generation method based on road network safety research and judgment according to the present application. As shown in FIG. 1, the road shunting suggestion generation method based on road network safety research and judgment provided by the present application comprises: Figure 1 ​Traffic data is collected by various types of sensors, and the traffic data is processed to obtain scenario-based traffic data. The scenario-based traffic data includes split zone data and merging zone data. The various types of sensors include video monitoring devices, radar monitoring devices, and weather monitoring devices, etc. The traffic data includes traffic flow data, event data, and weather data, etc. The scenario-based traffic data refers to the traffic data after split and merge labeling. The split zone data refers to the traffic-related data when the road is split. The merging data refers to the traffic-related data when the road is merged. The split zone data includes basic traffic data, environmental data, and device state data: The basic traffic data includes traffic volume, vehicle speed, and vehicle type distribution (including large, medium, and small vehicle types). The environmental data includes visibility and weather type; the weather type includes heavy fog, rain and snow, and / or night. The device state data includes the operating state of the induction lamp, the variable information board, and the broadcast. The induction lamp is used to indicate the lane contour. The broadcast operating state is used to indicate whether it is faulty, 1 indicating normal and 0 indicating fault; the broadcast is used to induce vehicles through voice prompts. The variable information board is used to display text prompts. The merging data includes road condition data and time data: the lane flow condition data on the right side of the main line is collected in real time by road line sensors and / or cameras, and the lane flow condition data includes merging lane flow, gate flow, main line speed, gate speed, and lane occupancy rate. The event data of lane merging conflict events (such as vehicles close to parallel) is collected by video and / or radar monitoring, and the event data includes the occurrence frequency and duration of the lane merging conflict event.

[0019] In some embodiments, the processing of the traffic data to obtain scenario-based traffic data includes data cleaning of the traffic data to remove abnormal data to obtain pre-processed traffic data. The data cleaning includes removing traffic volume data exceeding a maximum traffic volume threshold and vehicle speed data less than a minimum traffic speed of abnormal congestion. The pre-processed traffic data refers to data after removing abnormal data. Scene labels are added to the pre-processed traffic data to obtain scenario-based traffic data; the scene labels include split and merge labels (1 for split zone and 2 for merging zone) and weather labels (1 for low visibility and 0 for normal visibility); the split and merge labels are related to the split and merge of the road; the weather labels are related to the visibility. The scenario-based traffic data refers to the traffic data after labeling the split and merge scenes and the weather. For example, the scenario-based traffic data can be wherein, represents the scenario-based traffic data, represents the traffic data of the split zone; , represents the traffic volume, represents the vehicle speed, represents the vehicle type distribution, represents the visibility, representing a weather type, representing a running state of the broadcast; representing traffic data of a merging area ; representing a lane flow on the right side of the main line, representing a ramp flow, representing a main line speed, representing a ramp speed, representing a lane occupancy rate, representing a frequency of occurrence of a merging conflict event, representing a duration of a merging conflict event, representing a running state of the broadcast.

[0020] performing road network safety research and judgment on the scenario-based traffic data to obtain a road network safety research and judgment result; the road network safety research and judgment result includes a split area research and judgment result and a merging area research and judgment result. The road network safety research and judgment result refers to a result obtained after the road network safety is judged. The road network safety can be divided into multiple levels, and the research and judgment result refers to a range of road network safety levels into which a safety state of the current road network falls. The split area research and judgment result refers to a road network safety level into which a safety state of the split area falls. The merging area research and judgment result refers to a road network safety level into which a safety state of the merging area falls.

[0021] In some embodiments, obtaining the split area research and judgment result includes: based on a pre-warning event in the split area data, performing multi-dimensional statistical analysis to obtain an event occurrence frequency heat map. The pre-warning event is related to visibility and traffic flow. For example, an event in which the visibility is lower than a visibility threshold and the traffic flow is higher than a traffic flow threshold can be regarded as a pre-warning event. The event occurrence frequency heat map can be output by statistical analysis from three dimensions of time (morning peak or holiday), region, and event type (low visibility or congestion). For example, a split area low visibility event heat value in the morning peak from 7 to 9. The current visibility and the current traffic flow are matched to multiple split safety levels to obtain a current split safety level. The current visibility refers to the visibility at the current time. The current traffic flow refers to the traffic flow at the current time. The multiple split safety levels can include multiple safety levels. For example, the initial multiple split safety levels can include split high risk, split medium risk, and split low risk; the split low risk may refer to a visibility greater than or equal to 500 m and a traffic flow less than or equal to 1200 standard vehicle equivalent numbers per hour; the split medium risk refers to a visibility greater than or equal to 200 m and less than or equal to 500 m and a traffic flow greater than 1200 standard vehicle equivalent numbers per hour and less than or equal to 1800 standard vehicle equivalent numbers per hour; and the split high risk is referred to as visibility less than 200 m and traffic flow greater than 1800 standard vehicle equivalent numbers per hour. Based on the current split safety level, the split area research result is constructed. The split area research result includes the split safety level, the heat map and the expected traffic flow in the next 1 hour.

[0022] The merging area research result is obtained, including: using a random forest model to process the merging area data to obtain a merging accident prediction probability. The merging accident prediction probability is used to quantify the possibility of merging behavior accident. The traffic flow of the right lane of the main line , ramp traffic , main line speed , ramp speed and the occurrence frequency of merging conflict events are input into the trained random forest model, and the model outputs the merging accident prediction probability.

[0023] ; wherein, the merging accident prediction probability; i represents the decision tree variable; n represents the total number of decision trees; is the weight of the i-th decision tree; represents the output of the i-th decision tree; X is the input feature vector. The current merging accident prediction probability and the current lane occupancy rate are matched to a plurality of merging safety levels to obtain the current merging safety level. The current merging accident prediction probability represents the probability of the current merging accident. The current lane occupancy rate represents the lane occupancy ratio at the current time. The plurality of merging safety levels can include a plurality of merging safety levels related to traffic safety, which can include high-risk merging, medium-risk merging and low-risk merging. High-risk merging refers to merging accident prediction probability greater than or equal to 30% or lane occupancy rate greater than or equal to 80%; medium-risk merging refers to merging accident prediction probability greater than or equal to 15% and less than 30% or lane occupancy rate greater than or equal to 60% and less than 80%; low-risk merging refers to merging accident prediction probability less than 15% and lane occupancy rate less than 60%. The current merging safety level refers to the merging safety level at the current time. In some embodiments, the road network state is also visualized, and the split and merging area safety level is presented on the GIS map, and the traffic condition passing index is output, as shown in Figure 2 The icon boxed in the white box is a split and merging area safety warning, which indicates that there is a safety hazard in the split and merging area when the icon appears on the road section; wherein the average passing speed of the merging area is: ; wherein, the average passing speed of the merging area; represents the main line speed; represents the merging accident prediction probability. Based on the current merging safety level, the merging area research result is constructed. The merging area research result includes the merging accident prediction probability, the merging safety level and the merging area average speed.

[0024] Based on the road network safety research result, scene-based shunting processing is performed to obtain a shunting induction strategy. The shunting induction strategy includes a shunting area induction strategy and a merging area induction strategy. The shunting induction strategy refers to a strategy for shunting vehicles in the road network. The shunting area induction strategy refers to a strategy for shunting induction of vehicles on the shunting section, which can include induction light parameters, variable information board content and pre-warning logic. The merging area induction strategy refers to a strategy for shunting induction of vehicles on the merging section, which can include warning light frequency, speed limit instruction and merging prompt.

[0025] The shunting area induction strategy can include intelligent induction light configuration, induction mode configuration and pre-warning logic self-definition. The intelligent induction light configuration includes: when the visibility is lower than 500m, the lane-level road profile reinforcement indication is enabled, the induction light is lit at an interval of 2m per light, and the light color is white (profile indication) and yellow (lane change prompt); when the merging safety level is high risk of shunting, the induction light flashing frequency is increased to 2Hz, and the vehicle is guided to enter the shunting lane 1km in advance. The induction mode configuration includes: during holidays / bad weather, the variable information board is used to increase the text prompt "high risk of shunting area ahead, change lane 2km in advance", and the broadcast is cycled every 30s; during off-peak hours, it is simplified to "pay attention to safety, keep a distance" and the prompt is cycled. Figure 3 As shown in the figure, the pre-warning logic self-definition includes displaying green text "pay attention to safety, keep a distance" and voice prompt during normal driving; displaying red text "do not stay, drive away as soon as possible" and high-frequency voice prompt when the vehicle enters the anchor area; displaying red text "rear vehicle, drive away as soon as possible" and high-frequency voice prompt when the vehicle reverses in the anchor area; displaying yellow text "congestion ahead, slow down" and medium-frequency voice prompt when the ramp is congested.

[0026] The merging area induction strategy can include intelligent warning light control, speed and merging time control and pre-warning logic self-definition. The intelligent warning light control includes: when the merging is high risk, the flashing frequency is 3Hz (red), prompting the main line vehicle to slow down and the ramp vehicle to delay merging; when the merging is medium risk, the flashing frequency is 2Hz (yellow), prompting to keep a safe distance. The speed and merging time control includes: for the main line vehicle, the variable information board is used to issue a speed limit instruction , indicating the limit speed; for the ramp vehicle, the roadside broadcast is used to prompt "current merging risk is high, it is suggested to merge after 10s". As shown in the figure, Figure 4As shown, the early warning logic customization includes: main road vehicle, no vehicle in ramp, main road device green text warning "pay attention to safety, keep distance", ramp device yellow text warning "main road vehicle"; ramp vehicle, no vehicle in main road, main road device yellow text warning "ramp vehicle", ramp device green text warning "pay attention to safety, keep distance"; main road / ramp accident, main road device red text warning "accident ahead / ramp accident", ramp device red text warning "main road accident / accident ahead"; abnormal driving, main road device red text warning "rear vehicle attention, drive away as soon as possible", ramp device red text warning "rear vehicle attention, drive away as soon as possible"; congestion slow driving, main road device yellow text warning "congestion ahead, slow down", ramp device yellow text warning "congestion ahead, slow down".

[0027] The split diversion strategy is executed by the split diversion device, and the road network safety judgment result is continuously optimized based on the split diversion effect. The split diversion device refers to various traffic devices for split diversion. For example, the split diversion device can include intelligent diversion lights, intelligent warning lights, variable information boards, and roadside broadcasts. The split diversion effect refers to the change of traffic before and after the implementation of the split diversion strategy.

[0028] In some embodiments, the split diversion strategy is executed by the split diversion device, and the road network safety judgment result is continuously optimized based on the split diversion effect, including: taking the split diversion suggestion execution result; the split diversion suggestion execution result includes device operation state, actual split diversion vehicle number and driver feedback rate. The device operation state refers to the real-time working state of the device related to road split diversion in the process of executing the split diversion suggestion, which can include whether the device is normally running, fault type and position, etc. The actual split diversion vehicle number refers to the total number of vehicles actually driving according to the split diversion suggestion path, which can be obtained by statistical analysis through road monitoring devices during the execution of the split diversion suggestion. The driver feedback rate refers to the number of drivers who give effective feedback on the split diversion scheme, which accounts for the proportion of the total number of drivers passing through the split diversion section in this period, which can be collected through channels such as questionnaire survey, vehicle terminal feedback or hotline after the execution of the split diversion suggestion.

[0029] Based on the split diversion suggestion execution result, the safety index and the efficiency index are calculated. The safety index is: ; The efficiency index is: ; Wherein, represents the safety index; represents the accident rate before optimization; represents the accident rate after optimization; represents the efficiency index; represents the lane changing preparation time before optimization; The optimized lane-changing preparation time is represented.

[0030] Based on the safety index and the efficiency index, the multiple diverging safety level intervals and / or the multiple merging safety level intervals are optimized. For example, the weight of the difference between the main line vehicle speed and the ramp vehicle speed in the random forest model is adjusted based on the actual number of diverging vehicles and the safety index. For another example, the determination threshold of the diverging safety level interval is increased. The road network safety is judged based on the new multiple diverging safety levels and / or the new multiple merging safety levels.

[0031] The application also provides a road diverging suggestion generation system based on road network safety judgment, comprising a scenario modeling module, a judgment module, a measurement generation module and an optimization module. The scenario modeling module is used to collect traffic data through multiple types of sensors and process the traffic data to obtain scenario traffic data; the scenario traffic data includes diverging area data and merging area data; the judgment module is used to judge the road network safety based on the scenario traffic data to obtain road network safety judgment results; the road network safety judgment results include diverging area judgment results and merging area judgment results; the measurement generation module is used to perform diverging processing based on the road network safety judgment results to obtain diverging induction strategies; the diverging induction strategies include diverging area induction strategies and merging area induction strategies; the optimization module is used to execute the diverging induction strategies through diverging devices and continuously optimize the road network safety judgment results based on the diverging effects.

[0032] In some embodiments, the diverging area data includes basic traffic data, environmental data and device state data: the basic traffic data includes traffic volume, vehicle speed and vehicle type distribution; the environmental data includes visibility and weather type; the device state data includes the running state of the induction lamp, the variable information board and the broadcast; the merging data includes road condition data and time data: the road condition data includes merging lane traffic, gateway traffic, main line speed, gateway speed and lane occupancy rate; the event data includes the occurrence frequency and duration of lane conflict events.

[0033] In some embodiments, the scenario modeling module includes a data processing unit and a label adding unit; the data processing unit is used to clean the traffic data to remove abnormal data to obtain preprocessed traffic data; the label adding unit is used to add scenario labels to the preprocessed traffic data to obtain scenario traffic data; the scenario labels include diverging and merging labels and weather labels; the diverging and merging labels are related to the diverging and merging effects of the road; the weather labels are related to the visibility.

[0034] In some embodiments, the research module includes a split area research unit and a merging area research unit; the split area research unit is configured to obtain a split area research result, including: based on a warning event in the split area data, performing multi-dimensional statistical analysis to obtain an event occurrence frequency heat map; the warning event is related to visibility and traffic flow; matching the current visibility and the current traffic flow to a plurality of split safety levels to obtain a current split safety level; based on the current split safety level, constructing the split area research result; the merging area research unit is configured to obtain a merging area research result, including: using a random forest model to process the merging area data to obtain a parallel accident prediction probability; matching the current parallel accident prediction probability and the current lane occupancy rate to a plurality of merging safety levels to obtain a current merging safety level; based on the current merging safety level, constructing the merging area research result.

[0035] In some embodiments, the optimization module includes an execution result obtaining unit, an index calculation unit, an optimization unit, and a cycle unit; the execution result obtaining unit is configured to obtain a split suggestion execution result; the split suggestion execution result includes a split device operating state, an actual split vehicle number, and a driver feedback rate; the index calculation unit is configured to calculate a safety index and an efficiency index based on the split suggestion execution result; the optimization unit is configured to optimize a plurality of split safety level intervals and / or a plurality of merging safety level intervals based on the safety index and the efficiency index; and the cycle unit is configured to perform road network safety research based on new plurality of split safety levels and / or new plurality of merging safety levels.

[0036] The above is only a preferred embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application can have various modifications and changes. Any modification, equivalent replacement, improvement, etc. within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A method for generating road diversion suggestions based on road network safety assessment, characterized in that: include: Traffic data is collected through various types of sensors and processed to obtain scenario-based traffic data; Scenario-based traffic data includes diverging area data and merging area data; Conducting road network safety assessment on the scenario-based traffic data to obtain a road network safety assessment result; Road network safety assessment results include diverging area assessment results and merging area assessment results; Based on the results of road network safety assessment, traffic diversion is processed in different scenarios to obtain a diversion induction strategy; the diversion induction strategy includes a diversion area induction strategy and a merging area induction strategy; Diversion induction strategies are implemented through diversion equipment, and road network safety assessment results are continuously optimized based on the diversion effects.

2. The method for generating road diversion suggestions based on road network safety assessment according to claim 1 is characterized in that: Diversion area data includes basic traffic data, environmental data, and equipment status data: Basic traffic data includes traffic volume, speed, and vehicle type distribution; Environmental data includes visibility and weather type; Equipment status data includes the operating status of guidance lights, variable message boards and broadcasting; Merging data includes traffic condition data and time data: Traffic condition data includes merging lane flow, gate flow, mainline speed, gate speed, and lane occupancy rate; Event data includes the frequency and duration of merging conflicts.

3. The method for generating road diversion suggestions based on road network safety assessment according to claim 1 is characterized in that: The processing of traffic data to obtain scenario-based traffic data includes: Perform data cleaning on traffic data, remove abnormal data, and obtain pre-processed traffic data; Scenario tags are added to the pre-processed traffic data to obtain scenario-based traffic data; scenario tags include diverging and merging tags and weather tags; diverging and merging tags are related to the diverging and merging functions of the road; weather tags are related to visibility.

4. The method for generating road diversion suggestions based on road network safety assessment according to claim 1 is characterized in that: Obtain the results of the diversion area assessment, including: Based on the early warning events in the diversion area data, a multi-dimensional statistical analysis is performed to obtain a heat map of the frequency of events; the early warning events are related to visibility and traffic volume; Matching the current visibility and the current traffic volume to multiple diversion safety levels to obtain the current diversion safety level; Based on the current diversion safety level, construct the diversion area assessment results; Obtain the results of the merge area assessment, including: The random forest model is used to process the merging area data to obtain the predicted probability of merging accidents; Match the current merging accident prediction probability and the current lane occupancy rate to multiple merging safety levels to obtain the current merging safety level; Based on the current merging safety level, the merging area assessment results are constructed. The merging area assessment results include the predicted probability of merging accidents, the merging safety level, and the average speed in the merging area: ； in, represents the predicted probability of merging accidents; i represents the decision tree variable; n represents the total number of decision trees; is the weight of the decision tree for lesson i; represents the output of the i-th decision tree; X is the input feature vector; The average speed in the merging area is: ； in, Indicates the average speed of the merging area; Indicates the mainline speed; Represents the predicted probability of merging accidents.

5. The method for generating road diversion suggestions based on road network safety assessment according to claim 1 is characterized in that: Execute diversion induction strategies through diversion equipment, and continuously optimize road network safety assessment results based on diversion effects, including: Obtain the results of diversion suggestion execution; diversion suggestion execution results include the operating status of the diversion equipment, the actual number of diverted vehicles, and the driver feedback rate; Based on the results of the diversion suggestion execution, the safety index and efficiency index are calculated; the safety index is: ； The efficiency index is: ； in, Indicates safety indicators; represents the accident rate before optimization; represents the accident rate after optimization; represents the efficiency index; Indicates the lane change preparation time before optimization; Indicates the optimized lane change preparation time; Based on safety and efficiency indicators, optimize multiple diversion safety level intervals and / or multiple confluence safety level intervals; Conduct road network safety assessment based on multiple new diversion safety levels and / or multiple new merging safety levels.

6. A road diversion suggestion generation system based on road network safety assessment, characterized in that: It includes scenario module, analysis module, measurement generation module and optimization module; The scenario module is used to collect traffic data through various types of sensors and process the traffic data to obtain scenario-based traffic data; the scenario-based traffic data includes diverging area data and merging area data; The analysis module is used to analyze the road network safety of the scenario-based traffic data to obtain a road network safety analysis result; the road network safety analysis result includes a diverging area analysis result and a merging area analysis result; The measurement generation module is used to perform scenario-based diversion processing based on the road network safety assessment results to obtain a diversion induction strategy; the diversion induction strategy includes a diversion area induction strategy and a merging area induction strategy; The optimization module is used to execute the diversion induction strategy through the diversion equipment, and continuously optimize the road network safety assessment results based on the diversion effect.

7. The road diversion suggestion generation system based on road network safety assessment according to claim 6 is characterized in that: Diversion area data includes basic traffic data, environmental data, and equipment status data: Basic traffic data includes traffic volume, speed, and vehicle type distribution; Environmental data includes visibility and weather type; Equipment status data includes the operating status of guidance lights, variable message boards and broadcasting; Merging data includes traffic condition data and time data: Traffic condition data includes merging lane flow, gate flow, mainline speed, gate speed, and lane occupancy rate; Event data includes the frequency and duration of merging conflicts.

8. The road diversion suggestion generation system based on road network safety assessment according to claim 6 is characterized in that: The scenario module includes a data processing unit and a label adding unit; The data processing unit is used to clean the traffic data, remove abnormal data, and obtain pre-processed traffic data; The label adding unit is used to add scene labels to the pre-processed traffic data to obtain scene-based traffic data; the scene labels include diverging and merging labels and weather labels; the diverging and merging labels are related to the diverging and merging functions of the road; the weather labels are related to visibility.

9. The road diversion suggestion generation system based on road network safety assessment according to claim 6 is characterized in that: The analysis and judgment module includes a diversion area analysis and judgment unit and a confluence area analysis and judgment unit; The diversion area analysis unit is used to obtain the diversion area analysis result, including: Based on the early warning events in the diversion area data, a multi-dimensional statistical analysis is performed to obtain a heat map of the frequency of events; the early warning events are related to visibility and traffic volume; Matching the current visibility and the current traffic volume to multiple diversion safety levels to obtain the current diversion safety level; Based on the current diversion safety level, construct the diversion area assessment results; The merging area assessment unit is used to obtain a merging area assessment result, including: The random forest model is used to process the merging area data to obtain the predicted probability of merging accidents; Match the current merging accident prediction probability and the current lane occupancy rate to multiple merging safety levels to obtain the current merging safety level; Based on the current merging safety level, the merging area assessment results are constructed. The merging area assessment results include the predicted probability of merging accidents, the merging safety level, and the average speed in the merging area: ； in, represents the predicted probability of merging accidents; i represents the decision tree variable; n represents the total number of decision trees; is the weight of the decision tree for lesson i; represents the output of the i-th decision tree; X is the input feature vector; The average speed in the merging area is: ； in, represents the average speed in the merging area; Indicates the mainline speed; Represents the predicted probability of merging accidents.

10. The road diversion suggestion generation system based on road network safety assessment according to claim 6 is characterized in that: The optimization module includes an execution result acquisition unit, an index calculation unit, an optimization unit and a loop unit; The execution result is used by the acquisition unit to obtain the diversion suggestion execution result; the diversion suggestion execution result includes the operating status of the diversion device, the actual number of diverted vehicles and the driver feedback rate; The indicator calculation unit is used to calculate the safety indicator and efficiency indicator based on the diversion suggestion execution result; the safety indicator is: ； The efficiency index is: ； in, Indicates safety indicators; represents the accident rate before optimization; represents the accident rate after optimization; represents the efficiency index; Indicates the lane change preparation time before optimization; Indicates the optimized lane change preparation time; The optimization unit is used to optimize multiple diversion safety level intervals and / or multiple confluence safety level intervals based on the safety index and the efficiency index; The circulation unit is used to perform road network safety assessment based on multiple new diverging safety levels and / or multiple new merging safety levels.

Images