Real-time traffic flow signal control system based on multi-sensor fusion

Abstract

The present invention relates to the field of intelligent transportation technology, and specifically discloses a real-time traffic flow signal control system based on multi-sensor fusion. The present invention effectively ensures the consistency of traffic flow data by performing timestamp detection, missing value filling and data normalization on traffic flow data collected by different sensors, and further performs traffic flow data fusion through the Kalman filter algorithm, continuously corrects the measurement noise of multi-source sensors, and then outputs smoothed traffic flow data, providing strong data support for the analysis of subsequent signal light control strategies; dynamically solves the optimal signal light cycle length through the Webster formula, takes into account historical laws and real-time needs, avoids the lag of fixed timing, compares the actual signal light cycle length with the optimal signal light cycle length, dynamically adjusts the signal light cycle length, and then matches the signal light timing plan according to the traffic congestion index, effectively improving the intelligence, adaptability and overall road traffic capacity of real-time signal light control.

Description

Technical Field

[0001] The present invention relates to the field of intelligent transportation technology, and in particular to a real-time traffic flow signal control system based on multi-sensor fusion. Background Art

[0002] With the acceleration of urbanization and the continued growth of motor vehicle ownership, urban traffic congestion is becoming increasingly severe. Traffic flow management and signal control have become core tasks of urban traffic management. Traditional traffic signal control systems rely primarily on a single type of sensor (such as geomagnetic coils) to collect traffic flow data. These sensors only capture vehicle flow data and fail to capture multi-dimensional information such as pedestrian flow, vehicle type ratios, and vehicle speeds. Furthermore, these systems are susceptible to interference from factors such as road construction and weather, leading to data gaps or errors, which in turn affect the accuracy of signal control.

[0003] Furthermore, traditional traffic signal control systems mostly use fixed timing schemes, which are based on historical traffic flow statistics and fail to perceive dynamic changes in traffic conditions in real time. For example, during peak hours in the morning and evening, when traffic and pedestrian flows at intersections increase significantly, fixed timing schemes can easily lead to unused green lights in one direction and backlogs in the other. Furthermore, if peak timing schemes are still used during off-peak hours, road resources are wasted. This one-size-fits-all control approach is unable to adapt to the temporal and spatial fluctuations of traffic flow, resulting in low intersection efficiency, frequent congestion, and insufficient control rationality.

[0004] In order to solve the above-mentioned defects, a technical solution is now provided. Summary of the Invention

[0005] The purpose of the present invention is to provide a real-time traffic flow signal control system based on multi-sensor fusion to solve the problems raised in the above background.

[0006] The purpose of the present invention can be achieved through the following technical solutions: a real-time traffic flow signal control system based on multi-sensor fusion, comprising:

[0007] The data acquisition module is used to integrate multiple sensors such as geomagnetic sensors, cameras, millimeter-wave radar sensors, and meteorological sensors to collect the original traffic flow data of the target intersection during the current monitoring period and transmit it to the data processing module. It also collects the meteorological status data of the target intersection during the current monitoring period and transmits it to the server for storage;

[0008] The data processing module is used to receive the original traffic flow data of the target intersection during the current monitoring period and perform preprocessing operations to obtain the preprocessed traffic flow data of the target intersection during the current monitoring period and transmit it to the data fusion module;

[0009] A data fusion module is used to receive the pre-processed traffic flow data of the target intersection during the current monitoring period and perform recursive estimation and fusion of the data using the Kalman filter algorithm to obtain the optimal state estimate of the target intersection during the current monitoring period, and extract the optimal estimated traffic flow data of the target intersection during the current monitoring period and transmit it to the traffic congestion analysis module;

[0010] The traffic congestion analysis module is used to receive the optimal estimated traffic flow data of the target intersection during the current monitoring period, and analyze the traffic congestion index correction factor of the target intersection during the current monitoring period, thereby obtaining the traffic congestion index of the target intersection during the current monitoring period;

[0011] A signal cycle analysis module is used to analyze the optimal cycle duration of the signal light at the target intersection during the current monitoring period based on the traffic data of the target intersection during the preset historical period and the optimal estimated traffic flow of the target intersection during the current monitoring period;

[0012] A control strategy generation module is used to analyze the traffic light control instructions and signal light timing plan of the target intersection during the current monitoring period based on the traffic congestion index and the optimal cycle duration of the signal light at the target intersection during the current monitoring period;

[0013] The control terminal is used to receive signal light control instructions and signal light timing plans and perform corresponding control operations through the signal light controller.

[0014] Compared with the prior art, the present invention has the following beneficial effects:

[0015] The present invention effectively covers all elements of traffic flow status by integrating multiple types of sensors, avoiding the limitations of a single sensor. At the same time, it performs timestamp detection, missing value filling and data normalization on traffic flow data collected by different sensors, effectively ensuring the consistency of traffic flow data. It further fuses traffic flow data through the Kalman filter algorithm, continuously corrects the measurement noise of multiple-source sensors, and then outputs smoothed traffic flow data, effectively eliminating the measurement errors of multiple-source sensors, improving data reliability, and providing strong data support for the subsequent analysis of traffic light control strategies.

[0016] The present invention combines road type, flatness, number of lanes and vehicle type ratio to calculate a first traffic congestion correction factor using the inverse tangent function, and combines this with meteorological indicators to calculate and match a second traffic congestion correction factor. This quantifies the impact of inherent road properties and meteorological conditions on congestion, making congestion assessment more realistic. Traffic flow, speed and space occupancy are converted into the length, width and height of a rectangular parallelepiped, and the volume value is used as the congestion benchmark index, intuitively reflecting the "scale" of intersection congestion, avoiding the one-sidedness of a single indicator (such as vehicle flow alone), and providing a more comprehensive basis for signal control.

[0017] The present invention calculates the flow ratio based on the statistical data of the mean phase switching loss duration and the mode of saturated flow based on historical data, combines the current optimal estimated vehicle flow, and dynamically solves the optimal signal light cycle duration through the Webster formula, taking into account both historical rules and real-time needs, avoiding the lag of fixed timing, comparing the actual signal light cycle duration with the optimal signal light cycle duration, dynamically adjusting the signal light cycle duration, and then matching the signal light timing plan according to the traffic congestion index, thereby improving the intelligence and adaptability of real-time signal light control to a certain extent, effectively reducing traffic congestion, and improving the overall traffic capacity of the road. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The present invention will be further described below with reference to the accompanying drawings.

[0019] Figure 1 It is a system block diagram of the present invention. DETAILED DESCRIPTION

[0020] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0021] See also Figure 1 As shown, the present invention is a real-time signal control system for traffic flow based on multi-sensor fusion, which includes: a data acquisition module 101, a data processing module 102, a data fusion module 103, a traffic congestion analysis module 104, a signal period analysis module 105, a control strategy generation module 106, a control terminal 107 and a server 108. The data acquisition module 101 is connected to the data processing module 102, the data processing module 102 is connected to the data fusion module 103, the data fusion module 103 is connected to the traffic congestion analysis module 104, the signal period analysis module 105 is connected to the control strategy generation module 106, the traffic congestion analysis module 104 is connected to the control strategy generation module 106, the control strategy generation module 105 is connected to the control terminal 107, and the data acquisition module 101, the traffic congestion analysis module 104, and the signal period analysis module 105 are connected to the server 108 respectively.

[0022] The data acquisition module 101 integrates multiple types of sensors such as geomagnetic sensors, cameras, millimeter-wave radar sensors, and meteorological sensors to collect the original traffic flow data of the target intersection during the current monitoring period and transmit it to the data processing module, and collects the meteorological status data of the target intersection during the current monitoring period and transmits it to the server for storage.

[0023] Among them, the original traffic flow data includes: vehicle flow, pedestrian flow, vehicle type ratio, vehicle speed, and vehicle queue length.

[0024] The data processing module 102 is used to receive the original traffic flow data of the target intersection in the current monitoring period and perform preprocessing operations to obtain the preprocessed traffic flow data of the target intersection in the current monitoring period and transmit it to the data fusion module.

[0025] Specifically, the execution process of the pre-processing operation is as follows:

[0026] 201: Timestamp detection: original traffic flow data of the target intersection during the current monitoring period , extract all raw traffic flow data recorded during the current monitoring period The timestamps are arranged according to the size of the moment to generate a timestamp sequence ;

[0027] The original traffic flow data Including traffic flow , traffic flow 、Model ratio , vehicle speed , length of vehicle queues ;

[0028] In a timestamp sequence, calculate the time interval between all two adjacent timestamps , when the time interval between two adjacent timestamps is greater than the preset maximum time interval threshold When , it is determined that the time interval between the original traffic flow data records corresponding to the two adjacent timestamps is too large, that is, there is data missing between the two adjacent timestamps, and step 202 is executed.

[0029] 202: Missing value calculation and filling: Let the missing timestamp be , and the adjacent known traffic flow data are and , calculated by linear interpolation formula Get missing timestamps and missing timestamps corresponding to missing values of raw traffic flow , and fill in missing values accordingly.

[0030] It should be noted that timestamp detection, missing value calculation and filling are used to coordinate multi-source data (data collected by different sensors) to maintain consistency in the time base. Its core goal is to solve data asynchrony problems caused by factors such as sensor sampling frequency, clock deviation, and data transmission delay, and to ensure the accuracy of subsequent data processing and fusion.

[0031] 203: Traversal filling: perform traversal operations on all adjacent time stamps in all original traffic flow data time stamp sequences in the same manner as steps 201-202, complete filling of all missing values, and obtain the traffic flow data of the target intersection after missing values are filled in during the current monitoring period.

[0032] 204: Data normalization: normalizing the traffic flow data after filling missing values in the current monitoring period of the target intersection to obtain pre-processed traffic flow data in the current monitoring period of the target intersection.

[0033] It should be noted that the normalization processing method is minimum-maximum normalization. Normalization is to ensure that the original traffic flow data have the same scale and range to facilitate subsequent fusion calculations.

[0034] The data fusion module 103 is used to receive the pre-processed traffic flow data of the target intersection in the current monitoring period and perform data recursive estimation and fusion through the Kalman filter algorithm to obtain the optimal traffic flow data estimation value of the target intersection at each moment in the current monitoring period, and extract the optimal traffic flow data estimation value of the target intersection in the current monitoring period and transmit it to the traffic congestion analysis module.

[0035] It should be noted that the optimal traffic flow data estimation value includes the optimal vehicle flow estimation value, the optimal pedestrian flow estimation value, the optimal driving speed estimation value, the optimal vehicle model ratio estimation value, and the optimal vehicle queue length estimation value;

[0036] The method for obtaining the optimal traffic flow data estimate of the target intersection during the current monitoring period is as follows:

[0037] The optimal traffic flow data estimation values of the target intersection at each moment in the current monitoring period are averaged to obtain the optimal traffic flow data estimation mean of the target intersection in the current monitoring period, and this is used as the optimal traffic flow data estimation value of the target intersection in the current monitoring period.

[0038] Specifically, the data recursive estimation fusion using the Kalman filter algorithm includes:

[0039] 301: Initialization: Preset initialization traffic flow data estimation value and initialization traffic flow data estimation error ;

[0040] It should be noted that the estimated value of the initialized traffic flow data can be set based on historical data or experience; the estimated error of the initialized traffic flow data represents the degree of uncertainty in the estimated value of the initialized traffic flow data. This value reflects the confidence in the estimated value of the initialized traffic flow data. The larger the value, the greater the deviation between the estimated value of the initialized traffic flow data and the true value, and the higher the uncertainty; the smaller the value, the greater the confidence in the estimated value of the initialized traffic flow data.

[0041] 302: Kalman gain calculation: according to the formula Calculate the Kalman gain corresponding to the traffic flow data , where Indicates the sensor measurement error corresponding to the current traffic flow data, that is, the current time is k, Represents the estimation error of traffic flow data at the previous moment, that is, the previous moment is k-1 moment;

[0042] It should be noted that the Kalman gain is used to weigh the estimation error of the traffic flow data at the previous moment and the measurement error of the current traffic flow data to determine the weight of the measurement value when updating the estimated value; the measurement error is obtained from the factory operating instructions of the above-mentioned various types of sensors.

[0043] 303: Prediction: Estimated value based on traffic flow data at the previous moment , Kalman gain of traffic flow data at the current moment And the current traffic flow data measurement value, using the formula Predict the optimal traffic flow data estimate at the current moment , where Indicates the traffic flow data measurement value at the current moment;

[0044] It should be noted that all traffic flow data of the target intersection after preprocessing in the current monitoring period are respectively used as the traffic flow data measurement values at each moment in the current monitoring period.

[0045] 304: Update: According to the formula Calculate the current traffic flow data estimation error ;

[0046] 305: Iterative operation: The optimal traffic flow data estimate and the traffic flow data estimate error at the current moment are used as inputs to step 302 at the next moment, and the execution operations of steps 302-304 are continuously repeated to obtain the optimal traffic flow data estimate of the target intersection at each moment during the current monitoring period.

[0047] It should be noted that the Kalman filter is a recursive optimal estimation algorithm that uses a prediction-update mechanism to continuously correct for sensor measurement noise (such as random interference and inherent sensor errors) and output smoothed traffic flow data. When multiple sensors collect data, they are subject to various random interferences and inherent errors. For example, geomagnetic sensors may be affected by changes in the surrounding magnetic field, and cameras in poor lighting may produce image blur, leading to data errors and inaccurate data. The Kalman filter algorithm can effectively reduce the impact of noise on the data, resulting in smoother and more accurate traffic flow data output.

[0048] In a specific embodiment, the present invention effectively covers all elements of traffic flow status by integrating multiple types of sensors, avoiding the limitations of a single sensor. At the same time, it performs timestamp detection, missing value filling and data normalization on traffic flow data collected by different sensors, effectively ensuring the consistency of traffic flow data, and further integrates traffic flow data through the Kalman filter algorithm, continuously correcting the measurement noise of multiple source sensors, and then outputting smoothed traffic flow data, effectively eliminating the measurement errors of multiple source sensors, improving data reliability, and providing strong data support for the subsequent analysis of traffic light control strategies.

[0049] The traffic congestion analysis module 104 is used to receive the optimal traffic flow data estimate of the target intersection during the current monitoring period, and analyze the traffic congestion index correction factor of the target intersection during the current monitoring period, thereby analyzing and obtaining the traffic congestion index of the target intersection during the current monitoring period.

[0050] Specifically, the analysis process of the traffic congestion index correction factor of the target intersection during the current monitoring period is as follows:

[0051] The target intersection road type, road roughness, and number of lanes are obtained from the server. The optimal estimated vehicle type ratio of the target intersection during the current monitoring period is extracted and matched with the traffic congestion index correction factor impact values corresponding to each road type, each road roughness, each lane number, and each vehicle type ratio stored in the server. The traffic congestion index correction factor impact values corresponding to the road type, road roughness, number of lanes, and vehicle type ratio during the current monitoring period of the target intersection are obtained, which are recorded as ;

[0052] It should be noted that the road smoothness is obtained by collecting data using a laser smoothness meter.

[0053] Substitute the preset inverse tangent function Calculate the first traffic congestion index correction factor of the target intersection during the current monitoring period ;

[0054] Extract the meteorological status data of the target intersection during the current monitoring period from the server, and screen to obtain various meteorological status indicators of the target intersection during the current monitoring period. Compare each meteorological status indicator with its corresponding standard indicator. If a certain climate and environmental indicator is greater than its corresponding standard indicator, the difference between the two is calculated to obtain a high indicator difference. If a certain climate and environmental indicator is less than its corresponding standard indicator, the difference between the two is calculated to obtain a low indicator difference.

[0055] It should be noted that meteorological conditions indicators include visibility, wind speed, rainfall, snowfall, and visibility.

[0056] Calculate the mean of each high index difference and low index difference to obtain the high index difference mean and the low index difference mean, input the high index difference mean and the low index difference mean into the image processor, convert the high index difference mean and the low index difference mean into numerical values according to a certain ratio, and input them into the broken line graph to obtain two corresponding points, connect the two points in sequence with line segments to obtain a broken line, make the two end points of the broken line perpendicular to the X-axis respectively, so that the broken line and the two perpendicular lines form a closed figure with the X-axis, and identify the area of the closed figure as the meteorological state index of the target intersection in the current monitoring period;

[0057] Matching the meteorological state index of the target intersection during the current monitoring period with the meteorological state index corresponding to each preset second traffic index correction factor to obtain the second traffic congestion index correction factor of the target intersection during the current monitoring period;

[0058] The first traffic congestion index correction factor and the second traffic congestion index correction factor are summed to obtain a comprehensive traffic congestion index correction factor.

[0059] Specifically, the analysis process of the traffic congestion index of the target intersection during the current monitoring period is as follows:

[0060] The traffic flow value, traffic speed value, and space occupancy rate of the target intersection during the current monitoring period are obtained, and the ratios thereof are calculated with the preset maximum traffic flow value, minimum traffic speed value, and maximum space occupancy rate, respectively, to obtain the traffic flow ratio, traffic speed ratio, and space occupancy rate ratio of the target intersection during the current monitoring period. The traffic flow ratio, traffic speed ratio, and space occupancy rate ratio of the target intersection during the current monitoring period are converted into lengths according to the preset ratios, and a cuboid is constructed with the lengths of the traffic flow ratio, traffic speed ratio, and queue length ratio as the length, width, and height, respectively. The value of the volume of the cuboid is extracted as the congestion benchmark index of the target intersection during the current monitoring period. The congestion benchmark index is multiplied by the comprehensive correction factor of the traffic congestion index to obtain the traffic congestion index of the target intersection during the current monitoring period.

[0061] It should be noted that the process of obtaining traffic flow value, traffic speed value and space occupancy rate is as follows: the optimal vehicle flow estimate and the optimal pedestrian flow estimate of the target intersection in the current monitoring period are extracted from the optimal traffic flow data estimate of the target intersection in the current monitoring period, and the sum is calculated to obtain the traffic flow; the optimal vehicle driving speed estimate and the optimal pedestrian driving speed estimate of the target intersection in the current monitoring period are extracted, and the sum is calculated to obtain the traffic speed value; the optimal vehicle queue length estimate of the target intersection in the current monitoring period is extracted, and the ratio is calculated with the total length of the target intersection to obtain the space occupancy rate of the target intersection in the current monitoring period.

[0062] In a specific embodiment, the present invention combines road type, flatness, number of lanes, and vehicle type ratio to calculate a first traffic congestion correction factor using an inverse tangent function, and combines this with meteorological indicators to calculate and match a second traffic congestion correction factor. This quantifies the impact of inherent road attributes and meteorological conditions on congestion, making congestion assessment more relevant to actual scenarios. Traffic flow, speed, and space occupancy are converted into the length, width, and height of a rectangular parallelepiped, and the volume value is used as a congestion benchmark index. This intuitively reflects the "scale" of intersection congestion, avoids the one-sidedness of a single indicator (such as only vehicle flow), and provides a more comprehensive basis for signal control.

[0063] The signal cycle analysis module 105 is used to analyze the optimal cycle duration of the traffic light of the target intersection in the current monitoring period based on the traffic data of the target intersection in the preset historical period and the optimal estimated vehicle flow of the target intersection in the current monitoring period.

[0064] Specifically, the analysis process of the optimal cycle duration of the traffic light at the target intersection during the current monitoring period is as follows:

[0065] Extract the total phase switching loss duration of each signal light cycle at the target intersection in the preset historical period from the server, and calculate the average value to obtain the average total phase switching loss duration of the target intersection in the preset historical period, and use it as the total phase switching loss duration of the target intersection in the current monitoring period, recorded as ;

[0066] Extract the number of vehicles passing the stop line during the green light period of each signal light cycle at the target intersection within a preset historical period from the server, obtain the saturated flow rate of each signal light cycle at the target intersection within the preset historical period using the saturated flow rate calculation formula, and extract the mode of the saturated flow rate within each signal light cycle length as the saturated flow rate of the target intersection in the current monitoring period;

[0067] The optimal estimated traffic flow ratio of the target intersection in the current monitoring period is calculated by comparing the optimal traffic flow estimate value with the saturated traffic flow, and the optimal estimated traffic flow ratio of the target intersection in the current monitoring period is obtained, which is recorded as ;

[0068] Webster's algorithm formula Calculate the optimal cycle duration of the traffic light at the target intersection during the current monitoring period .

[0069] The control strategy generation module 106 is used to analyze the traffic light control instructions and traffic light timing plan of the target intersection in the current monitoring period based on the traffic congestion index of the target intersection in the current monitoring period and the optimal cycle duration of the traffic light.

[0070] Specifically, the process of analyzing the signal light control instructions and signal light timing plan of the target intersection during the current monitoring period is as follows:

[0071] Extract the actual traffic light cycle duration of the target intersection during the current monitoring period and compare and analyze it with the optimal traffic light cycle duration of the target intersection during the current monitoring period. If the optimal traffic light cycle duration is inconsistent with the actual traffic light cycle duration, generate a traffic light control instruction and send the traffic light control instruction to the control terminal;

[0072] Extract the traffic congestion index of the target intersection during the current monitoring period, and match the traffic congestion index of the target intersection during the current monitoring period with the preset signal light timing plan corresponding to the optimal cycle duration of each signal light for each traffic congestion index to obtain the signal light timing plan of the target intersection during the current monitoring period, wherein the signal light timing plan includes the green light duration, the red light duration and the yellow light duration, and send the signal light timing plan to the control terminal.

[0073] The control terminal 107 is used to receive signal light control instructions and signal light timing plans and execute corresponding control operations through the signal light controller.

[0074] In a specific embodiment, the present invention calculates the flow ratio based on the statistical mean of the phase switching loss duration and the mode of the saturated flow based on historical data, combines the current optimal estimated traffic flow, and dynamically solves the optimal signal light cycle duration through the Webster formula, taking into account both historical rules and real-time needs, avoiding the lag of fixed timing, comparing the actual signal light cycle duration with the optimal signal light cycle duration, dynamically adjusting the signal light cycle duration, and then matching the signal light timing plan according to the traffic congestion index, thereby improving the intelligence and adaptability of real-time control of signal lights to a certain extent, effectively reducing traffic congestion, and improving the overall traffic capacity of the road.

[0075] The server 108 is used to store the meteorological status data of the target intersection during the current monitoring period, store the road type, road roughness and number of lanes of the target intersection, store the traffic congestion index correction factor impact value corresponding to each road type, each road roughness, each lane number and each vehicle type ratio, and store the total phase switching loss time within each traffic light cycle in a preset historical period and the number of vehicles passing the stop line during the green light period.

[0076] The above content is merely an example and explanation of the structure of the present invention. Those skilled in the art may make various modifications or additions to the described specific embodiments or replace them in a similar manner. As long as they do not deviate from the structure of the invention or exceed the scope defined by the claims, they should all fall within the scope of protection of the present invention.

Claims

1. A real-time traffic flow signal control system based on multi-sensor fusion, characterized by: include: The data acquisition module is used to collect the original traffic flow data and meteorological status data of the target intersection during the current monitoring period and transmit them to the data processing module and the server respectively; The data processing module is used to receive the original traffic flow data of the target intersection during the current monitoring period and perform pre-processing operations to obtain the traffic flow data and transmit it to the data fusion module; The data fusion module uses the pre-processed traffic flow data of the target intersection in the current monitoring period to perform data recursive estimation and fusion through the Kalman filter algorithm to obtain the optimal traffic flow data estimate value of the target intersection at each moment in the current monitoring period, and extracts the optimal traffic flow data estimate value of the target intersection in the current monitoring period from it and transmits it to the traffic congestion analysis module; Traffic congestion analysis module, used to receive the estimated value of the optimal traffic flow data of the target intersection during the current monitoring period, and analyze the traffic congestion index correction factor to obtain the traffic congestion index; The analysis process of the traffic congestion index correction factor of the target intersection during the current monitoring period is as follows: Extract the road type, road roughness, number of lanes, and vehicle type ratio of the target intersection during the current monitoring period, and match them with the traffic congestion index correction factor impact values corresponding to each road type, road roughness, number of lanes, and vehicle type ratio stored in the cloud database to obtain the traffic congestion index correction factor impact values corresponding to the road type, road roughness, number of lanes, and vehicle type ratio of the target intersection during the current monitoring period; Substituting the preset arctangent function into the calculation, the first traffic congestion index correction factor of the target intersection during the current monitoring period is obtained; Extract the meteorological status data of the target intersection during the current monitoring period, and screen to obtain the meteorological status indicators of the target intersection during the current monitoring period. Compare each meteorological status indicator with its corresponding standard indicator. If a certain climate and environmental indicator is greater than its corresponding standard indicator, the difference between the two is calculated to obtain a high index difference. If a certain climate and environmental indicator is less than its corresponding standard indicator, the difference between the two is calculated to obtain a low index difference. Calculate the mean of each high index difference and low index difference to obtain the high index difference mean and the low index difference mean, input the high index difference mean and the low index difference mean into the image processor, convert the high index difference mean and the low index difference mean into numerical values according to a certain ratio, and input them into the broken line graph to obtain two corresponding points, connect the two points in sequence with line segments to obtain a broken line, make the two end points of the broken line perpendicular to the X-axis respectively, so that the broken line and the two perpendicular lines form a closed figure with the X-axis, and identify the area of the closed figure as the meteorological state index of the target intersection in the current monitoring period; Matching the meteorological state index of the target intersection during the current monitoring period with the meteorological state index corresponding to each preset second traffic index correction factor to obtain the second traffic congestion index correction factor of the target intersection during the current monitoring period; The first traffic congestion index correction factor and the second traffic congestion index correction factor are summed to obtain a comprehensive traffic congestion index correction factor; A signal cycle analysis module is used to analyze the optimal cycle duration of the signal light at the target intersection during the current monitoring period based on the traffic data of the target intersection during the preset historical period and the optimal estimated traffic flow of the target intersection during the current monitoring period; A control strategy generation module is used to analyze the traffic light control instructions and signal light timing plan for the target intersection during the current monitoring period based on the traffic congestion index and the optimal signal light cycle duration; The control terminal is used to receive signal light control instructions and signal light timing plans and perform corresponding control operations through the signal light controller.

2. The real-time traffic flow signal control system based on multi-sensor fusion according to claim 1 is characterized in that: The execution process of the pre-processing operation is as follows: 201: Timestamp detection: original traffic flow data of the target intersection during the current monitoring period , extract all raw traffic flow data recorded during the current monitoring period The timestamps are arranged according to the size of the moment to generate a timestamp sequence ; The original traffic flow data Including traffic flow , traffic flow 、Model ratio , vehicle speed and vehicle queue length ; In a timestamp sequence, calculate the time interval between all two adjacent timestamps , when the time interval between two adjacent timestamps is greater than the preset maximum time interval threshold When , it is determined that the time interval between the original traffic flow data records corresponding to the two adjacent timestamps is too large, that is, there is data missing between the two adjacent timestamps, and step 202 is executed; 202: Missing value calculation and filling: Let the missing timestamp be , and its adjacent known data points are and , calculated by linear interpolation formula Get missing timestamps and missing timestamps corresponding to missing values of raw traffic flow , and fill in missing values accordingly; 203: Traverse and fill: perform a traversal operation on all adjacent time stamps in all original traffic flow data time stamp sequences in the same manner as steps 201-202, complete filling of all missing values, and obtain the traffic flow data of the target intersection after missing values are filled in during the current monitoring period; 204: Data normalization: normalizing the traffic flow data after filling missing values in the current monitoring period of the target intersection to obtain pre-processed traffic flow data in the current monitoring period of the target intersection.

3. The real-time traffic flow signal control system based on multi-sensor fusion according to claim 1 is characterized in that: The data recursive estimation fusion by the Kalman filter algorithm includes: 301: Initialization: Preset initialization traffic flow data estimation value and initialization traffic flow data estimation error ; 302: Kalman gain calculation: according to the formula Calculate the Kalman gain corresponding to the traffic flow data , where Indicates the sensor measurement error corresponding to the current traffic flow data, that is, the current time is k, Represents the estimation error of traffic flow data at the previous moment, that is, the previous moment is k-1 moment; 303: Prediction: Estimated value based on traffic flow data at the previous moment , Kalman gain of traffic flow data at the current moment And the current traffic flow data measurement value, using the formula Predict the optimal traffic flow data estimate at the current moment , where Indicates the traffic flow data measurement value at the current moment; 304: Update: According to the formula Calculate the current traffic flow data estimation error ; 305: Iterative operation: The optimal traffic flow data estimate and the traffic flow data estimate error at the current moment are used as inputs to step 302 at the next moment, and the execution operations of steps 302-304 are continuously repeated to obtain the optimal traffic flow data estimate of the target intersection at each moment during the current monitoring period.

4. The real-time traffic flow signal control system based on multi-sensor fusion according to claim 1 is characterized in that: The analysis process of the traffic congestion index of the target intersection during the current monitoring period is as follows: The traffic flow value, traffic speed value, and space occupancy rate of the target intersection during the current monitoring period are obtained, and the ratios thereof are calculated with the preset maximum traffic flow value, minimum traffic speed value, and maximum space occupancy rate, respectively, to obtain the traffic flow ratio, traffic speed ratio, and space occupancy rate ratio of the target intersection during the current monitoring period. The traffic flow ratio, traffic speed ratio, and space occupancy rate ratio are converted into lengths according to the preset ratios, and a cuboid is constructed with the lengths of the traffic flow ratio, traffic speed ratio, and space occupancy rate ratio as the length, width, and height, respectively. The value of the volume of the cuboid is extracted as the congestion benchmark index of the target intersection during the current monitoring period. The congestion benchmark index is multiplied by the comprehensive correction factor of the traffic congestion index to obtain the traffic congestion index of the target intersection during the current monitoring period.

5. The real-time traffic flow signal control system based on multi-sensor fusion according to claim 1 is characterized in that: The analysis process of the optimal cycle duration of the traffic light at the target intersection during the current monitoring period is as follows: Extract the total phase switching loss duration of each signal light cycle at the target intersection within the preset historical period, calculate the average value, and obtain the average of the total phase switching loss duration of the target intersection within the preset historical period. This average value is used as the total phase switching loss duration of the target intersection in the current monitoring period. Extract the number of vehicles passing the stop line during the green light period of each signal light cycle at the target intersection within the preset historical period, calculate the saturated flow rate of each signal light cycle at the target intersection within the preset historical period using the saturated flow rate calculation formula, and extract the mode of the saturated flow rate within each signal light cycle as the saturated flow rate of the target intersection in the current monitoring period; Calculate the ratio of the optimal estimated traffic flow value of the target intersection in the current monitoring period to the saturated traffic flow to obtain the optimal estimated traffic flow ratio of the target intersection in the current monitoring period; The optimal cycle duration of the traffic light at the target intersection during the current monitoring period is calculated using the Webster algorithm formula.

6. The real-time traffic flow signal control system based on multi-sensor fusion according to claim 1 is characterized in that: The process of analyzing the signal light control instructions and signal light timing plan of the target intersection during the current monitoring period is as follows: Extract the actual traffic light cycle duration of the target intersection during the current monitoring period and compare and analyze it with the optimal traffic light cycle duration of the target intersection during the current monitoring period. If the optimal traffic light cycle duration is inconsistent with the actual traffic light cycle duration, generate a traffic light control instruction and send the traffic light control instruction to the control terminal; Extract the traffic congestion index of the target intersection during the current monitoring period, and match the traffic congestion index of the target intersection during the current monitoring period with the preset signal light timing plan corresponding to the optimal cycle duration of each signal light for each traffic congestion index to obtain the signal light timing plan of the target intersection during the current monitoring period, wherein the signal light timing plan includes the green light duration, the red light duration and the yellow light duration, and send the signal light timing plan to the control terminal.

Images