Vehicle-infrastructure cooperative intersection signal lamp control method and device, medium and product

Abstract

The invention provides a vehicle-road cooperative intersection signal lamp control method and device, a medium and a product. The method comprises the following steps: receiving multi-source data uploaded by roadside equipment; the multi-source data comprises at least one of detection data of common passenger vehicles, pedestrian data and data of the number of people in a bus in each direction of an intersection which are received by the roadside equipment and are cooperatively acquired by multiple devices; determining the total number of queuing people in each direction of the intersection according to the multi-source data; and according to the total number of queuing people in each direction, distributing the time length of the signal lamp corresponding to each direction by using a preset time length distribution model. According to the scheme of the invention, through the steps of receiving the multi-source data, determining the total number of queuing people, distributing the time length of the signal lamps and the like, the problem of low passing efficiency caused by static adjustment of the control time length of the signal lamps at the intersection in the prior art is effectively solved.

Description

Technical Field

[0001] The present application relates to the technical field of traffic light control, and in particular to a method, device, medium and product for controlling traffic lights at intersections in a vehicle-road collaborative manner. Background Art

[0002] In the urban traffic system, intersections are key nodes that connect various roads and realize traffic flow. Their traffic conditions not only affect the time cost of individual travelers, but also directly affect the traffic smoothness and operation efficiency of the entire city. Traditionally, traffic control at intersections mainly relies on traffic lights, and the control method of traffic lights is mostly based on vehicle flow to dynamically adjust the phase duration of traffic lights. However, this control method has defects: it fails to fully consider the differences in the number of different traffic participants and the far-reaching impact of such differences on the traffic efficiency of intersections.

[0003] Specifically, under the traditional traffic light control mode, whether it is a bus full of passengers or a private car carrying only a few people, they are all regarded as equivalent traffic units in front of the traffic light and enjoy the same right of passage. This "one-size-fits-all" approach ignores the differences in the actual needs of different vehicles for intersection resources, resulting in the inability to reasonably and efficiently allocate intersection resources. This problem is particularly prominent during peak hours or traffic congestion, often resulting in low traffic efficiency at intersections and increased traffic congestion. Summary of the invention

[0004] The purpose of this application is to provide a vehicle-road collaborative intersection signal light control method, device, medium and product to avoid the problem of static adjustment of intersection signal light control duration in the prior art, which leads to low intersection traffic efficiency during peak hours or traffic congestion.

[0005] To achieve the above objectives, in a first aspect, an embodiment of the present application provides a method for controlling a traffic light at an intersection in a vehicle-road cooperative manner, comprising:

[0006] Receiving multi-source data uploaded by a roadside device; the multi-source data includes at least one of detection data of ordinary passenger vehicles in each direction of the intersection, pedestrian data, and data on the number of people in a bus collected by multiple devices in a collaborative manner and received by the roadside device;

[0007] Determine the total number of people queuing in each direction of the intersection based on the multi-source data;

[0008] According to the total number of people queuing in each direction, the traffic light duration corresponding to each direction is allocated using a preset duration allocation model.

[0009] Optionally, determining the total number of people queuing in each direction of the intersection according to the multi-source data includes:

[0010] In the case where the multi-source data includes the detection data, estimating the first number of people corresponding to the detection data according to the detection data, in combination with historical data and a preset estimation model; the historical data is used to represent the number of ordinary passenger cars and the actual number of passengers corresponding to the ordinary passenger cars in the same area and the same time period of the intersection within a preset historical time;

[0011] In the case where the multi-source data includes the pedestrian data and the data on the number of people in the bus, determining a second number of people corresponding to the pedestrian data and a third number of people corresponding to the data on the number of people in the bus; the pedestrian data is determined based on a pedestrian detection device; the data on the number of people in the bus is determined based on an onboard device of the bus;

[0012] The first number of people, the second number of people, and the third number of people in each direction of the intersection are summed to determine the total number of people queuing in each direction of the intersection.

[0013] Optionally, according to the total number of people queuing in each direction, a preset duration allocation model is used to allocate the duration of the traffic light corresponding to each direction, including:

[0014] According to the total number of people queuing in each direction, the total number of people at the intersection and the dynamic adjustment coefficient of each direction are determined; the dynamic adjustment coefficient is determined comprehensively based on the traffic congestion index of each direction, the weather conditions at the intersection, the current season, and the travel habits of the time period passing through the intersection; the traffic congestion index is calculated comprehensively based on the total number of people queuing in each direction, the length of the current vehicle queue and the speed data;

[0015] Determine the traffic demand value in each direction according to the ratio of the total number of people in the queue to the total number of people at the intersection;

[0016] The duration allocation model is constructed by multiplying the traffic demand value, the dynamic adjustment coefficient and the preset basic duration;

[0017] The duration allocation model is used to allocate the traffic light duration corresponding to each direction.

[0018] Optionally, after allocating the signal light duration corresponding to each direction according to the total number of people queuing in each direction using a preset duration allocation model, the method further includes:

[0019] When the emergency vehicle detection device detects that the emergency vehicle is at a preset distance from the intersection, or receives a limited communication request signal sent by the emergency vehicle, obtaining the driving direction and speed of the emergency vehicle;

[0020] Determining an optimal signal light switching scheme according to the driving direction and speed of the emergency vehicle;

[0021] According to the optimal signal light switching scheme, the green light passage time of the signal light is extended for the intersection that the emergency vehicle needs to pass through, and after the emergency vehicle passes the current intersection, the step of allocating the signal light duration corresponding to each direction is performed according to the total number of people queuing in each direction using a preset duration allocation model.

[0022] Optionally, after allocating the signal light duration corresponding to each direction according to the total number of people queuing in each direction using a preset duration allocation model, the method further includes:

[0023] Obtaining the traffic demands of vehicles and pedestrians in different directions, as well as the road topology of the intersection and the traffic light topology corresponding to the intersection;

[0024] Determining the traffic priority of each direction according to the traffic demand, the road topology and the signal light topology;

[0025] According to the traffic priority, determining whether the current phase sequence of the signal light is conducive to smooth traffic flow;

[0026] If it is not conducive to the smooth flow of traffic, the phase sequence optimization algorithm is triggered to redetermine the phase sequence.

[0027] Optionally, after allocating the signal light duration corresponding to each direction according to the total number of people queuing in each direction using a preset duration allocation model, the method further includes:

[0028] Obtaining historical traffic data of the intersection;

[0029] According to the multi-source data and the historical traffic data, the time series analysis algorithm and the neural network algorithm model are input to predict the traffic flow of the intersection, and the prediction result of the intersection after the preset prediction time is determined;

[0030] According to the prediction results, calculate the basic duration, dynamic adjustment coefficient and phase sequence of the signal light timing;

[0031] The signal light is adjusted after the preset time according to the basic duration, the dynamic adjustment coefficient and the phase sequence of the signal light.

[0032] Optionally, the method further comprises:

[0033] Receive the vehicle's own driving intention and location information sent by the vehicle;

[0034] Analyze and process the received vehicle driving intention and location information to determine an analysis result; the analysis result includes the driving status of the vehicle and the estimated time of arrival at the intersection;

[0035] According to the analysis result and the duration allocation model, the switching time and duration of the signal light are adjusted in real time.

[0036] In a second aspect, the present application provides a vehicle-road cooperative intersection signal light control device, comprising:

[0037] A first receiving module is used to receive multi-source data uploaded by a roadside device; the multi-source data includes at least one of detection data of ordinary passenger cars in each direction of the intersection, pedestrian data, and data on the number of people in a bus received by the roadside device through multi-device collaborative collection;

[0038] A first determination module, used to determine the total number of people queuing in each direction of the intersection based on the multi-source data;

[0039] The first processing module is used to allocate the traffic light duration corresponding to each direction according to the total number of people queuing in each direction using a preset duration allocation model.

[0040] In a third aspect, the present application provides a readable storage medium having a program or instruction stored thereon, which implements the steps in the method described above when the program or instruction is executed by a processor.

[0041] In a fourth aspect, the present application provides a computer program product, comprising computer instructions, which implement the steps of the method described above when executed by a processor.

[0042] The beneficial effects of the above technical solution of the present application are as follows:

[0043] The present application aims at the problem that static adjustment of the control time of intersection lights in the prior art leads to low traffic efficiency at intersections during peak hours or in traffic congestion, and provides a vehicle-road collaborative intersection light control method. The method first receives multi-source data uploaded by roadside equipment. These multi-source data include detection data of ordinary passenger cars in each direction of the intersection, pedestrian data, and data on the number of people in buses, etc. collected by multi-device collaboration received by the roadside equipment. Through the collection of multi-source data, the traffic conditions at the intersection can be grasped in real time and comprehensively, including key information such as vehicle flow, pedestrian flow, and the number of people in buses. According to the received multi-source data, the total number of people queuing in each direction of the intersection is determined. By analyzing these data, the traffic pressure in each direction, that is, the total number of people queuing, can be accurately calculated, which is an important basis for dynamically adjusting the duration of the signal light. Finally, according to the total number of people queuing in each direction, the preset duration allocation model is used to allocate the duration of the signal light corresponding to each direction. The duration allocation model here can dynamically adjust the duration of the signal light according to the actual traffic conditions (that is, the total number of people queuing) to ensure that the traffic flow can pass through the intersection more smoothly. During peak hours or traffic congestion, due to the large number of people queuing, the signal light duration will be extended accordingly to ease traffic pressure; when there is less traffic, the signal light duration will be shortened to improve the traffic efficiency of the intersection.

[0044] The real-time nature of this application enables traffic light control to be dynamically adjusted according to current traffic conditions, rather than relying on fixed, static control schemes, thereby improving the intelligence and adaptability of traffic signal control and further improving traffic efficiency at intersections. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 A flow chart of a method for controlling a traffic light at an intersection with vehicle-road collaboration provided in an embodiment of the present application;

[0046] Figure 2 A structural diagram of a vehicle-road collaborative intersection signal light control device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0047] In order to make the technical problems, technical solutions and advantages to be solved by the present application clearer, a detailed description will be given below with reference to the accompanying drawings and specific embodiments.

[0048] It should be understood that the references to "one embodiment" or "an embodiment" throughout the specification mean that the specific features, structures, or characteristics associated with the embodiment are included in at least one embodiment of the present application. Therefore, the references to "in one embodiment" or "in an embodiment" appearing throughout the specification do not necessarily refer to the same embodiment. In addition, these specific features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.

[0049] In the embodiments of the present application, the term "and / or" describes the association relationship of the associated objects, indicating that three relationships may exist. For example, A and / or B may represent: A exists alone, A and B exist at the same time, and B exists alone. The character " / " generally indicates that the associated objects before and after are in an "or" relationship.

[0050] In the embodiments of the present application, the term "plurality" refers to two or more than two, and other quantifiers are similar.

[0051] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0052] The embodiments of the present application provide a method, device, medium and product for controlling intersection signals with vehicle-road collaboration. The method and device are based on the same application concept. Since the method and device solve the problem in a similar way, the implementation of the device and the method can refer to each other, and the repeated parts will not be repeated.

[0053] Reference Figure 1 As shown, the embodiment of the present application provides a method for controlling a traffic light at an intersection with vehicle-road collaboration, comprising:

[0054] Step 11, receiving multi-source data uploaded by roadside equipment; the multi-source data includes at least one of detection data of ordinary passenger cars in each direction of the intersection, pedestrian data and data on the number of people in a bus collected by multiple devices in a collaborative manner and received by the roadside equipment.

[0055] In this application, roadside equipment is deployed at various key locations at intersections, including but not limited to the four corners of the intersection, the green belt in the middle of the road, etc. These devices are equipped with a variety of sensors, and the collaborative collection of multiple devices includes but is not limited to millimeter-wave radars, cameras, etc., which are used to collect detection data of ordinary passenger cars. Among them, the millimeter-wave radar accurately detects information such as the vehicle's position, speed, and direction of travel by transmitting and receiving millimeter-wave signals; the camera uses image recognition technology to identify the type of vehicle, license plate number, etc. The two work together to ensure the comprehensiveness and accuracy of data collection on ordinary passenger cars. This application uses roadside equipment to collect detection data of ordinary passenger cars in each direction of the intersection collaboratively by multiple devices and send it to the control center in the cloud.

[0056] In terms of pedestrian data collection, pedestrian detection sensors are installed at both ends of the crosswalk at the intersection, which can use infrared sensing or pressure sensing technology. When a pedestrian enters the sensing area, the sensor can immediately detect the presence of the pedestrian and record the number of pedestrians, entry time and other information. At the same time, combined with the camera installed at a high point at the intersection, the pedestrian posture recognition algorithm is used to track the movement trajectory of the pedestrian. This application uses the roadside equipment to send the detected pedestrian data to the control center in the cloud.

[0057] The number of people in the bus is obtained by installing on-board sensors or on-board devices (OBU) on the bus. For example, the number of people in the bus counted by the OBU is uploaded to the roadside device, and the roadside device sends it to the control center in the cloud. Or, for example, a bus infrared counter or a people counting system based on image processing. When passengers get on or off the bus, the sensor automatically counts and uploads the real-time number of people in the bus to the roadside device through the on-board communication device. The roadside device then aggregates these multi-source data collected from different devices and uploads them to the control center in the cloud.

[0058] Step 12: Determine the total number of people queuing in each direction of the intersection based on the multi-source data.

[0059] In an embodiment of the present application, for ordinary passenger cars, the vehicles in the queue are determined based on the vehicle position information collected by the roadside equipment. Ordinary passenger cars are distinguished by vehicle type identification and their number is counted. Then, the total number of people in the ordinary passenger cars is calculated in combination with the average passenger capacity of the vehicle (an average value can be set based on historical data or local traffic statistics). In terms of pedestrian data, the number of pedestrians counted by the pedestrian detection sensor is directly used as the number of pedestrians queuing in that direction of the intersection. For the data on the number of people in the bus, the uploaded real-time number of people in the bus is added to the total number of people in that direction. Finally, the total number of people in the ordinary passenger cars, the number of pedestrians queuing, and the number of people in the bus are added together to obtain the total number of people queuing in each direction of the intersection.

[0060] Step 13, according to the total number of people queuing in each direction, use a preset duration allocation model to allocate the traffic light duration corresponding to each direction.

[0061] In the embodiment of the present application, the preset duration allocation model is constructed based on a deep learning algorithm. First, a large amount of historical traffic data is collected, including the total number of queues in each direction of each intersection under different time periods and weather conditions and the corresponding optimal signal light duration settings (obtained through actual traffic flow monitoring and optimization experiments). These data are used to train the deep learning model so that the model learns the complex mapping relationship between the total number of queues and the signal light duration.

[0062] In practical applications, the total number of queues in each direction calculated in step 12 is input into the trained duration allocation model. The model outputs the signal duration corresponding to each direction based on the learned pattern. For example, for directions with a large number of queues, the model will allocate a longer signal duration to ensure that vehicles and pedestrians in that direction can pass through the intersection smoothly; and for directions with a small number of queues, a relatively short signal duration is allocated to improve the overall traffic efficiency of the intersection.

[0063] Compared with the prior art, the present application solves the problem that traditional intersection signal control is mostly based on a single vehicle detection data and cannot fully consider the traffic conditions at the intersection. It integrates ordinary passenger car detection data, pedestrian data and bus population data. Through multi-device collaborative collection, the information of each traffic participant at the intersection is fully obtained, providing a richer and more accurate data basis for signal light control, which can more truly reflect the actual traffic needs of the intersection. The present application comprehensively considers the total number of people in the queue with different numbers of traffic participants. The number of people in ordinary passenger cars, the number of pedestrians and the number of people in buses are included in the calculation, making the assessment of the traffic conditions at the intersection more comprehensive and scientific, and providing a more reasonable basis for the allocation of signal light duration. The present application adopts a duration allocation model based on deep learning, which is different from the traditional fixed timing or simple rule-based timing method. The deep learning model can automatically learn the complex relationship between the total number of people in the queue and the duration of the signal light in historical data, and adapt to dynamic changes in different time periods and different traffic conditions. It can intelligently allocate the most suitable traffic light duration based on the total number of people in the real-time queue, significantly improving the traffic efficiency at the intersection, reducing the waiting time of vehicles and pedestrians, and improving traffic smoothness and safety.

[0064] Optionally, in multi-device collaborative collection, the application can use a variety of detection equipment such as video detectors, millimeter wave radars, and lidars to collect data such as vehicle flow, vehicle type, and vehicle speed in all directions of the intersection in real time. In the vehicle-road collaborative solution, mobile edge computing (MEC) devices also participate in data collection and perform preliminary processing and analysis on the collected vehicle data.

[0065] The bus is equipped with an OBU device, which is connected to the on-board passenger detection system and can obtain and transmit the bus's passenger number data in real time and accurately.

[0066] Pedestrian detection device collection: By deploying cameras or sensors at intersections, using advanced image recognition algorithms and sensor technology, the number of pedestrians waiting to cross the road in all directions of the intersection is counted, and the behavioral characteristics of pedestrians, such as walking speed, gathering situation, etc., are analyzed.

[0067] Finally, roadside equipment data aggregation: The roadside communication unit (i.e., roadside equipment RSU) collects data from OBU devices and other intelligent devices on the roadside (such as meteorological sensors, road condition sensors, etc.), and uploads it to the control center or processing center of the data cloud in real time to achieve the aggregation of multi-source data.

[0068] Optionally, the vehicle detection equipment and pedestrian detection device send the collected data to the RSU via a wired network (such as optical fiber) or a 5G wireless network. After the RSU integrates and preliminarily processes the data, it is transmitted to the control center or processing center in the cloud in real time to ensure the timeliness and accuracy of the data.

[0069] OBU device data transmission: OBU devices use the Cellular Vehicle-to-Everything (C-V2X) communication technology to send the number of people in the bus to the RSU. The RSU then forwards this data to the control center or processing center in the cloud, realizing efficient data transmission between the vehicle and the road.

[0070] Optionally, step 12 of the present application includes:

[0071] Step 121, when the multi-source data includes the detection data, estimating the first number of people corresponding to the detection data according to the detection data, combined with historical data and a preset estimation model; the historical data is used to represent the number of ordinary passenger cars in the same area and the same time period of the intersection within a preset historical time and the actual number of passengers corresponding to the ordinary passenger cars;

[0072] Step 121, when the multi-source data includes the pedestrian data and the data of the number of people in the bus, determining a second number of people corresponding to the pedestrian data and a third number of people corresponding to the data of the number of people in the bus; the pedestrian data is determined based on a pedestrian detection device; the data of the number of people in the bus is determined based on an onboard device of the bus;

[0073] Step 123, summing the first number of people, the second number of people, and the third number of people in each direction of the intersection to determine the total number of people queuing in each direction of the intersection.

[0074] In the embodiment of the present application, ordinary passenger car detection data at different intersections, different time periods, different weather conditions, and different traffic conditions, the corresponding actual number of passengers data, and related auxiliary data such as time, weather, and road sections are widely collected. The collected data is annotated to clarify the actual number of passengers corresponding to the number of ordinary passenger cars in each data sample as a supervisory label for model training. According to the characteristics of the data and the complexity of the problem, a linear regression model, a decision tree regression model, a neural network regression model, etc. are selected. Taking the decision tree regression model as an example, it can handle nonlinear relationships and can intuitively display the decision rules between features and outputs. According to the type of model selected, the corresponding model structure is constructed. For example, for the neural network regression model, the number of layers of the network, the number of neurons in each layer, the activation function and other parameters are determined. The collected data is randomly divided according to a certain proportion (such as 70% for the training set and 30% for the test set) to ensure that the data distribution of the training set and the test set is similar. Set the parameters of model training, such as learning rate, number of iterations, regularization parameters, etc. For example, the learning rate is set to 0.01 and the number of iterations is set to 1000 times. Input the training set data into the model, and minimize the error (such as mean square error) between the model's prediction results and the actual label by continuously adjusting the model's parameters. During the training process, the model gradually learns the relationship between the number of ordinary passenger cars, models, time periods, weather and other characteristics and the actual number of passengers. Input the test set data into the trained model and calculate the model's evaluation indicators, such as mean square error (MSE), mean absolute error (MAE), etc. The performance of the model is judged by the evaluation indicators. If the model performance is not ideal, analyze the reasons and optimize it. Possible optimization measures include adjusting the model structure, increasing the amount of training data, adjusting training parameters, etc. For example, if the model is found to be overfitting, the regularization parameter can be increased or the complexity of the model can be reduced. Through the above process, the estimation model is determined.

[0075] In the case where the multi-source data includes the detection data, features are extracted from the real-time detection data: features closely related to the estimated number of passengers are extracted, for example, the relevant feature is the vehicle model, and there are obvious differences in the passenger capacity of different vehicle models, such as a small car generally has a passenger capacity of 2-5 people, while an MPV model may have a passenger capacity of 6-7 people. By identifying and classifying the vehicle model, basic information is provided for estimating the number of passengers. The time when the vehicle enters the detection area is also extracted, and the passenger carrying patterns of vehicles in different time periods are different. For example, carpooling may be more common during the morning rush hour on weekdays, and the passenger capacity is relatively high.

[0076] Extract features from historical data: In addition to the two core data of the number of ordinary passenger cars and the actual number of passengers, we also extract features such as time period (distinguishing between weekdays / weekends, morning, noon and evening peaks, etc.), weather conditions (sunny, rainy, snowy, etc., different weather conditions affect travel methods and willingness to carpool, such as more people may choose to carpool on rainy days), and special holidays (travel demand and methods during holidays are different from weekdays, for example, during the Spring Festival, vehicles returning home may be full of passengers). These historical data features will be combined with real-time detection data features to provide more comprehensive information for the estimation model.

[0077] The features extracted from the real-time detection data and historical data are integrated to form a complete feature data set. For example, the real-time detected vehicle model, the time of entering the detection area, and the corresponding time period, weather conditions, special holidays and other features in the historical data are combined to construct a feature vector containing multiple influencing factors as the input data of the estimation model. The input data after preprocessing and feature integration is used as the input of the estimation model, and the estimation model analyzes and calculates the input data and outputs the estimated first number of people. The final estimation result is the estimated value of the number of passengers in an ordinary passenger car corresponding to the current detection data, that is, the first number of people is determined.

[0078] For example, the estimation model generally makes a preliminary estimate based on an average of 2 people per vehicle. However, considering the differences in actual conditions in different time periods and on different road sections, the estimation coefficient is adjusted dynamically. For example, during rush hours in the morning and evening on weekdays, the average number of passengers in vehicles in commercial areas may increase due to carpooling. At this time, by analyzing historical data and real-time traffic flow, the estimation coefficient is adjusted to 2.5 people; and in non-busy sections late at night, the coefficient may be adjusted to 1.5 people. The specific adjustment process is automatically completed by the machine learning model based on the comparative analysis of real-time collected data and historical data, thereby improving the accuracy of the estimate.

[0079] The bus transmits the actual number of people in the bus in real time through the on-board OBU device. After receiving this data, the roadside device directly uses it as the accurate statistical result of the number of bus passengers, that is, to determine the third person. For example, when a bus enters the detection range of the intersection, its OBU device immediately sends the current number of passengers in the car to the roadside device, ensuring the timeliness and accuracy of the bus passenger data.

[0080] Pedestrian detection devices, such as millimeter-wave radars or high-definition video monitoring equipment, are deployed at intersections. These devices use target recognition algorithms to identify non-motor vehicles and pedestrians waiting at intersections. By analyzing and counting the recognition results, estimates of the number of non-motor vehicles and pedestrians in each direction are formed. For example, video monitoring equipment extracts and counts the features of pedestrians in the image, combined with pedestrian movement trajectory analysis, to accurately count the number of pedestrians waiting in a certain direction at the intersection. Millimeter-wave radar estimates the number and location of pedestrians by detecting millimeter-wave signals reflected by the human body, thereby determining the second number of people.

[0081] For example, the pedestrian detection device counts that the number of pedestrians waiting in a certain direction is 100, that is, the second number is 100; the bus on-board equipment transmits that the number of people in a certain bus is 30, that is, the third number is 30.

[0082] The first, second, and third person counts in each direction of the intersection are summed to determine the total number of people queuing in each direction of the intersection. For example, if the first person count in a certain direction is estimated to be 200, the second person count is 100, and the third person count is 30, then the total number of people queuing in that direction is 200+100+30=330. These accurate total number of people queuing data will provide strong data support for the subsequent allocation of traffic light durations, so as to achieve more reasonable and efficient intersection traffic control.

[0083] Optionally, before determining the first number of people, the second number of people, and the third number of people, the present application constructs a data fusion model, and uses the data fusion model to assist in determining the first number of people, the second number of people, and the third number of people.

[0084] Specifically, collect at least one year of historical multi-source traffic data at intersections, including historical data of ordinary passenger car detection data (vehicle location, speed, type, etc.), historical data of pedestrian data (number of pedestrians, location, movement trajectory), historical data of the number of people in buses, etc. Clean the collected historical multi-source traffic data to remove obviously erroneous or incomplete data records. For example, remove data points with abnormal speed (such as exceeding the road speed limit too much) in vehicle detection data. Sort and organize the data by time (hour, weekday / weekend, season, etc.) and intersection area for subsequent analysis.

[0085] Use data mining and machine learning algorithms to build a data fusion model. This data fusion model can automatically identify the association between different types of data, such as ordinary passenger car detection data, pedestrian data, and bus population data, by learning a large amount of historical multi-source traffic data. For example, the association rule mining algorithm is used to discover the potential connection between ordinary passenger car flow and pedestrian flow in different time periods. At the same time, the data cleaning algorithm is used to eliminate redundant information in the data, such as repeated vehicle detection data; the error correction model is used to correct data that may have errors, so as to improve data quality and lay the foundation for subsequent accurate population estimation and signal light duration allocation. Among them, select a suitable association rule mining algorithm, such as the Apriori algorithm. Input the preprocessed multi-source data into the algorithm model. Set the support and confidence thresholds, for example, the support is set to 0.1 (indicating that at least 10% of the data set contains the rule), and the confidence is set to 0.8 (indicating that 80% of the data that meet the premise conditions also meet the result conditions). Through algorithms, the correlation between different types of data is mined. For example, during the morning rush hour on weekdays, when the flow of ordinary passenger cars in a certain intersection area increases by 20%, the pedestrian flow will increase by 15% accordingly.

[0086] The density-based spatial clustering algorithm (DBSCAN) is used to detect and remove duplicate vehicle detection data. The algorithm divides the data points that are densely connected in space into a cluster, and marks and deletes isolated data points (duplicate data often appear as isolated points).

[0087] Construct an error correction model, such as using a linear regression model to correct pedestrian detection data that may have errors. Use historically accurate pedestrian detection data as the training set, take the current detected pedestrian data features (such as the number of detected pedestrians, the time interval between pedestrians, etc.) as input, and the model outputs the corrected number of pedestrians.

[0088] When determining the first number of people, the data fusion model pre-processes the detection data of ordinary passenger cars to remove outliers, making the data more accurate and preparing for the subsequent combination of historical data and estimation models. The association rule mining algorithm can reveal the relationship between the number of ordinary passenger cars and other traffic elements in different time periods and sections, assist in determining the estimation coefficient, and optimize the estimation of the first number of people. For example, mining the correlation between ordinary passenger cars and pedestrian flow during the morning rush hour on weekdays can enable the model to comprehensively consider this correlation when estimating the first number of people, making the estimation more realistic.

[0089] When determining the second number of people, the data fusion model cleans and corrects the errors of the data collected by the pedestrian detection device. For example, the linear regression model is used to correct the pedestrian detection data to improve its accuracy and ensure the reliability of the second number of people. At the same time, the pedestrian data mined by the model is associated with other traffic data, which can further verify and optimize the calculation of the second number of people. For example, it is found that the pedestrian flow is related to the frequency of bus arrivals during a specific period of time, which can assist in judging the rationality of pedestrian data.

[0090] When determining the third person, the data fusion model does not directly process the data of the number of people in the bus, but it plays an indirect role by ensuring the accuracy of the communication link data between the OBU device and the roadside equipment. The data fusion model ensures the accuracy of other traffic data and helps the roadside equipment to verify the rationality when receiving the data of the number of people in the bus, such as judging whether the data is abnormal in combination with the overall traffic flow at the intersection.

[0091] Optionally, step 13 of the present application includes:

[0092] According to the total number of people queuing in each direction, the total number of people at the intersection and the dynamic adjustment coefficient of each direction are determined; the dynamic adjustment coefficient is determined comprehensively based on the traffic congestion index of each direction, the weather conditions at the intersection, the current season, and the travel habits of the time period passing through the intersection; the traffic congestion index is calculated comprehensively based on the total number of people queuing in each direction, the length of the current vehicle queue and the speed data;

[0093] Determine the traffic demand value in each direction according to the ratio of the total number of people in the queue to the total number of people at the intersection;

[0094] The duration allocation model is constructed by multiplying the traffic demand value, the dynamic adjustment coefficient and the preset basic duration;

[0095] The duration allocation model is used to allocate the traffic light duration corresponding to each direction.

[0096] In the embodiment of the present application, the dynamic adjustment coefficient is an important parameter that comprehensively considers various real-time factors to measure the traffic complexity in each direction and the impact of special circumstances on the duration of the traffic light. Its determination involves the following key steps:

[0097] Calculate the traffic congestion index: Use sensors and monitoring equipment installed at intersections to collect the total number of people queuing in each direction P, the length of the current vehicle queue L, and the speed data V. The traffic congestion index C is calculated by a comprehensive function C = f(P, L, V), for example, a weighted linear combination method C = α*P + β*L + γ / V can be used, where α, β, and γ are weight coefficients, and α+β+γ=1 is satisfied. These weight coefficients need to be reasonably set according to the degree of influence of the actual traffic conditions.

[0098] Consider weather conditions: Weather conditions have a significant impact on traffic operations, which can be quantified as a value W. For example, a sunny day can be set to W=1, light rain can be set to W=0.9, heavy rain can be set to W=0.8, and so on.

[0099] Combine the current season and travel habits: Different seasons and time periods at intersections correspond to different travel habits, which are quantified into a value S T For example, during the tourist season or during peak hours in the morning and evening, the number of people and travel patterns are different from usual, and the corresponding S T The value will also change.

[0100] Comprehensively determine the dynamic adjustment coefficient, such as using the weighted summation method K = w1*C+w2*W+w3*S T , where w1, w2, w3 are weight coefficients, and w1+w2+w3=1.

[0101] The traffic demand value reflects the proportion of each direction in the entire intersection traffic flow, which is crucial for the reasonable allocation of signal light duration. total (P total =P1+P2+…+Pi, where i is the number of intersections) Di=P / P total , we can get the traffic demand value of the ith direction. This ratio clearly shows the relative size of traffic demand in each direction, providing direct data support for the subsequent allocation of traffic light duration.

[0102] The duration allocation model is the core formula for determining the duration of the signal light according to the traffic demand and actual situation in each direction. The preset basic duration is T0, which is determined based on an in-depth analysis of the historical traffic flow data of the intersection. During the analysis process, the differences in traffic flow at different time periods at the intersection (such as morning peak, evening peak, and off-peak period), on weekdays and non-working days, as well as road grades (main roads, secondary roads), intersection types (cross intersections, T-shaped intersections, etc.) and other factors are fully considered. For example, the basic duration of the intersection of the main road will be set longer due to the large traffic flow and the need to ensure continuity; while the basic duration of the intersection of the secondary road and the main road will be reasonably set according to the actual traffic conditions. The green light duration ti in each direction is calculated according to the formula ti = T0*Di*K. This formula combines the traffic demand value Di, the dynamic adjustment coefficient K and the preset basic duration T0, comprehensively considers the actual traffic conditions in each direction, and can allocate the signal light duration more scientifically. Using the above-built duration allocation model, the calculated green light duration for each direction is allocated to the corresponding direction. In this way, the traffic light system can dynamically adjust the duration of the traffic light according to the real-time traffic demand and special circumstances in each direction, thereby improving the traffic operation efficiency of the intersection and reducing traffic congestion.

[0103] In order to enable the duration of traffic lights to adapt to the dynamic changes of traffic conditions in a timely manner, this application establishes an effective dynamic adjustment mechanism, including:

[0104] Real-time monitoring data: Continuously monitor traffic data such as the number of people queuing in each direction, the length of vehicle queues, and vehicle speeds through sensors and monitoring equipment, while paying attention to changes in weather conditions, seasons, and time periods.

[0105] Dynamically update traffic conditions: timely process and analyze real-time monitored data, update parameters such as traffic congestion index, dynamic adjustment coefficient, and traffic demand value in each direction to accurately reflect the current actual traffic conditions.

[0106] Traffic light duration adjustment: Using a dynamic adjustment algorithm, the traffic light duration in each direction is adjusted in a timely manner according to the parameters updated in real time. For example, when the traffic flow in a certain direction increases suddenly, resulting in a significant increase in the number of people queuing in that direction and a significant increase in the traffic congestion index, the system automatically increases the green light duration in that direction to ensure that the traffic flow in that direction can pass smoothly; conversely, when the traffic flow in a certain direction decreases, the green light duration is reduced accordingly.

[0107] Optionally, after allocating the signal light duration corresponding to each direction according to the total number of people queuing in each direction using a preset duration allocation model, the method further includes:

[0108] When the emergency vehicle detection device detects that the emergency vehicle is at a preset distance from the intersection, or receives a limited communication request signal sent by the emergency vehicle, obtaining the driving direction and speed of the emergency vehicle;

[0109] Determining an optimal signal light switching scheme according to the driving direction and speed of the emergency vehicle;

[0110] According to the optimal signal light switching scheme, the green light passage time of the signal light is extended for the intersection that the emergency vehicle needs to pass through, and after the emergency vehicle passes the current intersection, the step of allocating the signal light duration corresponding to each direction is performed according to the total number of people queuing in each direction using a preset duration allocation model.

[0111] In an embodiment of the present application, emergency vehicle detection equipment based on radar technology and image recognition technology is deployed at the intersection to monitor the situation around the intersection in real time. Radar technology is used to detect the distance, speed and other information of the vehicle, and image recognition technology is used to identify the type of vehicle to determine whether it is an emergency vehicle (such as a fire truck, ambulance, police car, etc.). When an emergency vehicle is detected approaching an intersection and the distance reaches a preset distance, subsequent operations are triggered. After the emergency vehicle detection equipment collects the information about the approach of the emergency vehicle, the driving direction and speed of the emergency vehicle are obtained through vehicle-road collaborative technology. At the same time, the priority passage request signal sent by the emergency vehicle itself is received to further confirm the passage needs of the emergency vehicle.

[0112] After receiving the relevant information of the emergency vehicle, the cloud control center of the present application immediately triggers the phase forced switching program. According to the driving direction and speed of the emergency vehicle, combined with the current traffic conditions in all directions of the intersection (such as vehicle queues, real-time traffic flow, etc.), the preset algorithm is used to quickly calculate the optimal signal light switching plan. This plan must ensure that the emergency vehicle passes through the intersection quickly to the greatest extent possible while minimizing the impact on traffic in other directions under the premise of ensuring safety. According to the determined optimal signal light switching plan, the green light passage time of the signal light is extended for the intersection that the emergency vehicle needs to pass through to ensure that the emergency vehicle can pass smoothly. At the same time, the signal lights in other directions are adjusted accordingly, such as shortening the green light time or switching to a red light to avoid traffic conflicts. After the emergency vehicle passes the current intersection, the normal dynamic timing mode is restored. That is, according to the total number of people queuing in each direction, the preset time allocation model is used to reallocate the signal light duration corresponding to each direction to adapt to normal traffic flow changes.

[0113] Optionally, the application also continuously performs fault detection on relevant components such as emergency vehicle detection equipment and signal control systems, and monitors the operating status of the equipment in real time. When a relevant fault is detected, the signal light switching operation can be manually triggered through manual monitoring video to ensure that emergency vehicles can still pass smoothly and maintain the basic order of traffic at the intersection.

[0114] Optionally, after allocating the signal light duration corresponding to each direction according to the total number of people queuing in each direction using a preset duration allocation model, the method further includes:

[0115] Obtaining the traffic demands of vehicles and pedestrians in different directions, as well as the road topology of the intersection and the traffic light topology corresponding to the intersection;

[0116] Determining the traffic priority of each direction according to the traffic demand, the road topology and the signal light topology;

[0117] According to the traffic priority, determining whether the current phase sequence of the signal light is conducive to smooth traffic flow;

[0118] If it is not conducive to the smooth flow of traffic, the phase sequence optimization algorithm is triggered to redetermine the phase sequence.

[0119] In the embodiment of the present application, sensors (such as geomagnetic sensors, video detectors, etc.) are set at the intersection to count the traffic flow, queue length and other information of vehicles in different directions, so as to measure the traffic demand of vehicles; at the same time, pedestrian detection equipment (such as infrared sensors, pressure-sensitive floor tiles, etc.) is set near the crosswalk to count the traffic flow, waiting time, etc. of pedestrians to obtain the traffic demand of pedestrians. Utilize geographic information system (GIS) technology and data from actual surveys of intersections to obtain the road topology of the intersection, including the connection relationship of each road, the number of lanes, lane functions (such as left turn, straight, right turn lanes) and other information. From the database or configuration file of the signal light control system, obtain the signal light topology corresponding to the intersection, that is, the grouping of signal lights, the control logic of each signal light and the corresponding relationship with the lanes.

[0120] Combine the acquired traffic demand data of vehicles and pedestrians with the road topology and signal light topology. For example, consider a road with heavy traffic and long queues, and the lane corresponding to the road is in a key connection position in the road topology, and the signal light corresponding to the lane has a specific control method in the signal light topology. Combine these factors for analysis. Formulate priority assessment rules, such as assigning different weights based on factors such as road grade, traffic flow, pedestrian flow, and whether it is a bus lane, and calculate the traffic priority values ​​for each direction. For example, the main road has heavy vehicle flow, which is assigned a higher weight and has a relatively high traffic priority; the direction with heavy pedestrian flow also has its priority appropriately increased.

[0121] Based on the principles of traffic engineering and actual traffic operation experience, set standards for judging whether the current phase sequence of traffic lights is conducive to smooth traffic. For example, it is defined that the average waiting time of vehicles exceeds a certain threshold, the queue length in a certain direction is too long and affects the passage of vehicles in other directions, etc. are not conducive to smooth traffic. Compare the actual traffic operation indicators of each direction under the current phase sequence of traffic lights (such as vehicle waiting time, queue length, etc.) with the set smooth traffic standards. According to the priority of traffic, judge whether the current phase sequence can meet the efficient passage of each direction in the order of priority. If it cannot be met, it is judged as not conducive to smooth traffic.

[0122] According to the actual situation and needs of the intersection, select the appropriate phase sequence optimization algorithm. For example, for sections with relatively stable traffic flow and appropriate intersection spacing, a green wave control strategy can be adopted; for complex multi-road intersections, a phase sequence optimization method based on intelligent algorithms such as genetic algorithms and simulated annealing algorithms can be selected.

[0123] According to the acquired data such as road topology, signal light topology, and traffic demand, relevant parameters are set for the selected phase sequence optimization algorithm. For example, in the green wave control strategy, parameters such as the distance between intersections and the average speed of vehicles need to be set; in the intelligent algorithm, parameters such as population size, number of iterations, crossover probability, and mutation probability need to be set.

[0124] Run the phase sequence optimization algorithm to improve the overall efficiency of traffic flow and reduce vehicle waiting time. Through the iterative calculation of the algorithm, the optimal phase sequence of traffic lights is searched. For example, in the phase sequence optimization based on genetic algorithms, new phase sequence schemes are generated through continuous selection, crossover, and mutation operations, and their fitness is evaluated to gradually approach the optimal solution. The optimal phase sequence of traffic lights calculated by the optimization algorithm is applied to the actual traffic light control system, the phase sequence setting of the traffic lights is updated, and the optimal adjustment of the phase sequence of traffic lights is achieved.

[0125] This application integrates the traffic needs of vehicles and pedestrians in different directions, the road topology, and the traffic light topology. By comprehensively considering these factors to determine the traffic priority, the actual traffic conditions at the intersection can be more accurately reflected, breaking the previous limitations of a single or a few factors determining the timing of traffic lights. The solution of this application has the ability to dynamically determine whether the current phase sequence is conducive to smooth traffic. Once it is determined that it is not conducive to smooth traffic, the phase sequence optimization algorithm is quickly triggered. This real-time monitoring and dynamic adjustment mechanism can adapt to changes in traffic flow in a timely manner, significantly improve traffic operation efficiency, and effectively alleviate congestion. This application also introduces advanced algorithms such as green wave band control strategies and genetic algorithms, which greatly improve the scientificity and effectiveness of signal light phase sequence optimization.

[0126] Optionally, after allocating the signal light duration corresponding to each direction according to the total number of people queuing in each direction using a preset duration allocation model, the method further includes:

[0127] Obtaining historical traffic data of the intersection;

[0128] According to the multi-source data and the historical traffic data, the time series analysis algorithm and the neural network algorithm model are input to predict the traffic flow of the intersection, and the prediction result of the intersection after the preset prediction time is determined;

[0129] According to the prediction results, calculate the basic duration, dynamic adjustment coefficient and phase sequence of the signal light timing;

[0130] The signal light is adjusted after the preset time according to the basic duration, the dynamic adjustment coefficient and the phase sequence of the signal light.

[0131] In the embodiment of the present application, multiple ways to obtain historical traffic data include but are not limited to traffic flow sensors installed at intersections (such as geomagnetic sensors, laser radar sensors, etc.), surveillance cameras (obtaining vehicle and pedestrian flow information through video analysis technology), electronic toll collection systems (used to record the number and time of vehicles passing through the intersection) and the database of the traffic management department (storing long-term traffic flow statistics of the intersection). According to a certain time period (for example, in hours, days or weeks), historical traffic data related to the intersection is collected from the above data sources, including information such as vehicle flow, pedestrian flow, vehicle speed, queue length, etc. in different time periods. The collected multi-source data is cleaned and pre-processed to remove noise data and outliers, and then integrated to form a historical traffic data set in a unified format. According to the multi-source data and the historical traffic data, they are input into the time series analysis algorithm and the neural network algorithm model to predict the traffic flow of the intersection, and determine the prediction result of the intersection after the preset prediction time.

[0132] Select appropriate time series analysis algorithms, such as the Autoregressive Integrated Moving Average Model (ARIMA), the Seasonal Autoregressive Integrated Moving Average Model (SARIMA), etc. According to the characteristics and rules of historical traffic data, determine the parameters of the model (such as autoregressive order, difference order, moving average order, etc.), and build a time series analysis model. This model can capture the changing trend and periodic characteristics of traffic flow data over time.

[0133] Adopt neural network algorithms in deep learning, such as long short-term memory network (LSTM), gated recurrent unit (GRU), etc. Design the structure of the neural network, including the number of neurons in the input layer, hidden layer and output layer, activation function, etc. The input layer receives multi-source data and historical traffic data, the hidden layer extracts features and performs nonlinear transformation on the input data, and the output layer outputs the predicted traffic flow value.

[0134] The integrated historical traffic data set is divided into a training set and a validation set. The training set is used to train the time series analysis algorithm model and the neural network algorithm model. By continuously adjusting the model parameters, the error between the model's prediction results and the actual traffic flow data is minimized. During the training process, methods such as cross-validation are used to evaluate the performance of the model to prevent the model from overfitting.

[0135] The trained time series analysis algorithm model and neural network algorithm model are integrated to comprehensively utilize the advantages of the two models to improve the accuracy of traffic flow prediction. The current multi-source data and historical traffic data are input into the integrated model to predict the traffic flow at the intersection after a preset time (for example, 15 minutes, 30 minutes, etc. in the future) to obtain the prediction results, including information such as vehicle flow, pedestrian flow, and queue length in different directions.

[0136] According to the predicted traffic flow in different directions, combined with the principles of traffic engineering and relevant traffic design standards, the basic duration of the signal lights in each direction is calculated. For example, for directions with heavy traffic, the basic duration of the green light is appropriately increased to ensure that vehicles can pass through the intersection smoothly; for directions with heavy pedestrian flow, the basic duration of the green light for pedestrians is increased. The specific calculation method can adopt the Webster timing method, the saturated flow method, etc.

[0137] Taking into account the dynamic changes and uncertainties of traffic flow, a dynamic adjustment coefficient is introduced to adjust the basic duration in real time. The calculation of the dynamic adjustment coefficient can be based on multiple factors, such as the predicted fluctuation range of traffic flow, the degree of congestion at the intersection, the impact of special events (such as traffic accidents, large-scale events, etc.), etc. By establishing a calculation model for the dynamic adjustment coefficient, the dynamic adjustment coefficient can be calculated in real time based on the prediction results and real-time monitoring data.

[0138] The phase sequence of the traffic lights is determined based on the prediction results and the traffic flow characteristics of the intersection. The principle of determining the phase sequence of the traffic lights is to minimize the conflict between vehicles and pedestrians and improve the traffic efficiency of the intersection. For example, for intersections with a large number of left-turning vehicles, a special left-turn phase can be set; for intersections with a large pedestrian flow, the phase for pedestrians crossing the street can be reasonably arranged. The optimal phase sequence of the traffic lights is selected by simulating and evaluating different phase sequence schemes.

[0139] The signal light is adjusted after the preset time according to the basic duration, the dynamic adjustment coefficient and the phase sequence of the signal light.

[0140] A new signal light timing scheme is generated based on the calculated basic duration, dynamic adjustment coefficient and signal light phase sequence. The basic duration is multiplied by the dynamic adjustment coefficient to obtain the actual timing of the signal lights in each direction, and the lights are arranged according to the determined signal light phase sequence.

[0141] After the preset time arrives, the new signal light timing plan is sent to the signal light control system at the intersection. The signal light control system automatically adjusts the duration and phase sequence of the signal light according to the received timing plan to achieve real-time optimization control of the traffic flow at the intersection. At the same time, the traffic conditions at the intersection are continuously monitored, and the signal light timing plan is further adjusted and optimized according to the actual situation.

[0142] Through the above specific implementation method, the traffic flow at the intersection can be accurately predicted based on the historical traffic data and real-time multi-source data of the intersection, and the timing and phase sequence of the traffic lights can be adjusted in real time according to the prediction results to improve the traffic efficiency and traffic safety of the intersection.

[0143] Optionally, the method further comprises:

[0144] Receive the vehicle's own driving intention and location information sent by the vehicle;

[0145] Analyze and process the received vehicle driving intention and location information to determine an analysis result; the analysis result includes the driving status of the vehicle and the estimated time of arrival at the intersection;

[0146] According to the analysis result and the duration allocation model, the switching time and duration of the signal light are adjusted in real time.

[0147] In the embodiment of the present application, a high-sensitivity, low-latency dedicated communication module is deployed. The module follows a specific communication protocol that has been strictly encrypted and optimized for efficient data transmission, ensuring that the vehicle's own driving intention and location information sent by the vehicle-mounted communication device can be stably and accurately received in complex electromagnetic environments and different weather conditions. The vehicle's driving intention information not only includes specific expressions of conventional driving directions such as going straight, turning left, turning right, and turning around, but also covers special driving intention marks such as emergency avoidance and special mission passage; the location information relies on multi-constellation fusion positioning technologies such as the Global Positioning System (GPS) and the Beidou Satellite Navigation System (BDS), combined with a high-precision map matching algorithm, accurate to the sub-meter coordinate data of the vehicle's current geographic location, and can also obtain detailed location information such as the lane the vehicle is in and the distance from the edge of the road.

[0148] Based on the vehicle's current driving status, driving intention, and distance from the intersection, the adaptive prediction model is used by comprehensively considering factors such as road speed limits, real-time traffic congestion conditions (obtained through real-time traffic flow monitoring systems, floating vehicle data, etc.), road construction, and emergencies. For example, when traffic flow is stable, a prediction algorithm based on a dynamic model is used to accurately calculate the vehicle's power performance parameters; in complex traffic congestion scenarios, it switches to a statistical prediction model based on historical traffic data, incorporates a real-time road condition correction coefficient, calculates the estimated time the vehicle will arrive at the intersection, and provides a confidence interval for the time prediction to reflect the reliability of the prediction.

[0149] Obtain a pre-built duration allocation model, and determine the reasonable switching time and duration of each phase of the traffic light under different traffic scenarios based on factors such as historical traffic flow data at the intersection, road capacity, priority of different driving directions, traffic characteristics at different time periods, and so on.

[0150] This application can use the analyzed vehicle driving status and estimated arrival time at the intersection as input parameters, combined with real-time collected information such as intersection traffic flow and queue length, to input the duration allocation model. Based on these parameters and the current operating status of the traffic light, the model calculates the switching time and duration adjustment plan required for each phase of the traffic light. Subsequently, through a traffic light control system with redundant backup and rapid response mechanism, the switching time and duration of the traffic light are adjusted in real time according to the adjustment plan, so as to achieve optimized control of traffic flow at the intersection and improve overall traffic efficiency.

[0151] Optionally, this application can also build a communication platform through blockchain for cross-departmental collaboration, use smart contracts for regular communication, and combine information from multiple departments to layout traffic lights during urban planning. In terms of multi-source data fusion, a middle platform is built to construct a neural network model to optimize traffic light strategies. In the system implementation and monitoring part, edge computing deployment is adopted to carry out hardware redundancy design and software testing, and real-time data collection through vehicle-road collaboration is used to dynamically adjust control strategies.

[0152] In summary, this application uses multi-dimensional data fusion and vehicle-road collaboration to accurately estimate the number of people in line and reasonably allocate intersection resources. Traffic lights are optimized with the goal of maximizing the number of people passing through, reducing vehicle waiting time, increasing the number of people passing through per unit time, and improving traffic efficiency. This application can adjust traffic lights according to real-time traffic conditions to respond to sudden traffic changes, and has strong dynamic adaptability. This application uses existing infrastructure and does not require large-scale reconstruction, so it is highly feasible to implement.

[0153] Reference Figure 2 As shown, the embodiment of the present application provides a vehicle-road cooperative intersection signal light control device, comprising:

[0154] The first receiving module 21 is used to receive multi-source data uploaded by the roadside device; the multi-source data includes at least one of the detection data of ordinary passenger cars in each direction of the intersection, pedestrian data and data on the number of people in the bus received by the roadside device in a collaborative manner by multiple devices;

[0155] A first determination module 22, for determining the total number of people queuing in each direction of the intersection according to the multi-source data;

[0156] The first processing module 23 is used to allocate the traffic light duration corresponding to each direction according to the total number of people queuing in each direction by using a preset duration allocation model.

[0157] In the embodiment of the present application, the first determining module 22 includes:

[0158] A first processing unit is configured to estimate a first number of people corresponding to the detection data based on the detection data, in combination with historical data and a preset estimation model, when the multi-source data includes the detection data; the historical data is used to represent the number of ordinary passenger cars and the actual number of passengers corresponding to the ordinary passenger cars in the same area and the same time period of the intersection within a preset historical time;

[0159] a second processing unit, for determining a second number of people corresponding to the pedestrian data and a third number of people corresponding to the number of people in the bus, when the multi-source data includes the pedestrian data and the number of people in the bus; the pedestrian data is determined based on a pedestrian detection device; the number of people in the bus is determined based on an onboard device of the bus;

[0160] The third processing unit is used to sum the first number of people, the second number of people and the third number of people in each direction of the intersection to determine the total number of people queuing in each direction of the intersection.

[0161] In the embodiment of the present application, the first processing module 23 includes:

[0162] A first determination unit is used to determine the total number of people at the intersection and the dynamic adjustment coefficient of each direction according to the total number of people queuing in each direction; the dynamic adjustment coefficient is determined comprehensively according to the traffic congestion index of each direction, the weather conditions of the intersection, the current season, and the travel habits of the time period passing through the intersection; the traffic congestion index is calculated comprehensively according to the total number of people queuing in each direction, the length of the vehicle queue of the current vehicle, and the speed data;

[0163] A second determining unit, configured to determine a traffic demand value in each direction according to a ratio of the total number of people in the queue to the total number of people at the intersection;

[0164] A construction unit, configured to construct the duration allocation model by multiplying the traffic demand value, the dynamic adjustment coefficient and a preset basic duration;

[0165] The fourth processing unit is used to allocate the traffic light duration corresponding to each direction by using the duration allocation model.

[0166] In the embodiment of the present application, the above-mentioned vehicle-road cooperative intersection signal light control device also includes:

[0167] A first acquisition module, configured to acquire the travel direction and travel speed of the emergency vehicle when the emergency vehicle detection device detects that the emergency vehicle is at a preset distance from the intersection, or receives a limited communication request signal sent by the emergency vehicle;

[0168] A second determination module, used to determine an optimal signal light switching solution according to the driving direction and driving speed of the emergency vehicle;

[0169] The second processing module is used to extend the green light passage time of the traffic light at the intersection that the emergency vehicle needs to pass through according to the optimal traffic light switching plan, and after the emergency vehicle passes the current intersection, execute the step of allocating the traffic light duration corresponding to each direction according to the total number of people queuing in each direction using a preset duration allocation model.

[0170] In the embodiment of the present application, the above-mentioned vehicle-road cooperative intersection signal light control device also includes:

[0171] A second acquisition module is used to acquire the traffic demands of vehicles and pedestrians in different directions, as well as the road topology structure of the intersection and the traffic light topology structure corresponding to the intersection;

[0172] A third determination module is used to determine the traffic priority of each direction according to the traffic demand, the road topology and the signal light topology;

[0173] A third processing module is used to determine whether the current phase sequence of the signal light is conducive to smooth traffic according to the traffic priority;

[0174] The fourth processing module is used to trigger the phase sequence optimization algorithm and redetermine the phase sequence if it is not conducive to the smooth flow of traffic.

[0175] In the embodiment of the present application, the above-mentioned vehicle-road cooperative intersection signal light control device also includes:

[0176] A third acquisition module is used to acquire historical traffic data of the intersection;

[0177] A fifth processing module, for predicting the traffic flow of the intersection according to the multi-source data and the historical traffic data input into the time series analysis algorithm and the neural network algorithm model, and determining the prediction result of the intersection after a preset prediction time;

[0178] The sixth processing module is used to calculate the basic duration, dynamic adjustment coefficient and phase sequence of the signal light timing according to the prediction result;

[0179] The seventh processing module is used to adjust the signal light after the preset time according to the basic duration, the dynamic adjustment coefficient and the phase sequence of the signal light.

[0180] In the embodiment of the present application, the above-mentioned vehicle-road cooperative intersection signal light control device also includes:

[0181] The second receiving module is used to receive the driving intention and location information sent by the vehicle;

[0182] A fourth determination module is used to analyze and process the received vehicle driving intention and the location information to determine an analysis result; the analysis result includes the driving state of the vehicle and the estimated time of arrival at the intersection;

[0183] An eighth processing module is used to adjust the switching time and duration of the traffic light in real time according to the analysis result and the duration allocation model.

[0184] Among them, the implementation embodiments of the above-mentioned vehicle-road collaborative intersection signal light control method are all applicable to the embodiments of the vehicle-road collaborative intersection signal light control device, and can also achieve the same technical effects.

[0185] A readable storage medium of an embodiment of the present application stores a program or instruction thereon. When the program or instruction is executed by a processor, the steps in the intersection traffic light control method for vehicle-road collaboration as described above are implemented, and the same technical effect can be achieved. To avoid repetition, it will not be repeated here.

[0186] The processor is the processor in the vehicle-road cooperative intersection signal control method described in the above embodiment. The readable storage medium includes a computer-readable storage medium, such as a computer read-only memory (ROM), a random access memory (RAM), a disk or an optical disk, etc.

[0187] The present application also provides a computer program product, including computer instructions, which, when executed by a processor, implement the above Figure 1 The various processes of the method embodiment shown can achieve the same technical effect, and will not be described again here to avoid repetition.

[0188] The above exemplary embodiments are described with reference to the accompanying drawings, and many different forms and embodiments are feasible without departing from the spirit and teachings of the present application. Therefore, the present application should not be constructed as a limitation of the exemplary embodiments proposed herein. More specifically, these exemplary embodiments are provided so that the present application will be perfect and complete, and the scope of the present application will be conveyed to those who are familiar with the technology. In these figures, the component sizes and relative sizes may be exaggerated for clarity. The terms used here are only based on the purpose of describing specific exemplary embodiments and are not intended to be limiting. As used herein, unless the text clearly indicates otherwise, the singular forms "one", "an" and "the" are intended to include these multiple forms. It will be further understood that the terms "comprising" and / or "including" when used in this specification indicate the presence of the features, integers, steps, operations, components and / or components, but do not exclude the presence or increase of one or more other features, integers, steps, operations, components, components and / or their groups. Unless otherwise indicated, when stated, a range of values ​​includes the upper and lower limits of that range and any subranges therebetween.

[0189] The above is a preferred embodiment of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles described in the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. A vehicle-road cooperative intersection signal light control method, characterized in that: include: Receiving multi-source data uploaded by a roadside device; the multi-source data includes at least one of detection data of ordinary passenger vehicles in each direction of the intersection, pedestrian data, and data on the number of people in a bus collected by multiple devices in a collaborative manner and received by the roadside device; Determine the total number of people queuing in each direction of the intersection based on the multi-source data; According to the total number of people queuing in each direction, the traffic light duration corresponding to each direction is allocated using a preset duration allocation model.

2. The method according to claim 1, characterized in that: Determine the total number of people queuing in each direction of the intersection based on the multi-source data, including: In the case where the multi-source data includes the detection data, estimating the first number of people corresponding to the detection data according to the detection data, in combination with historical data and a preset estimation model; the historical data is used to represent the number of ordinary passenger cars and the actual number of passengers corresponding to the ordinary passenger cars in the same area and the same time period of the intersection within a preset historical time; In the case where the multi-source data includes the pedestrian data and the data on the number of people in the bus, determining a second number of people corresponding to the pedestrian data and a third number of people corresponding to the data on the number of people in the bus; the pedestrian data is determined based on a pedestrian detection device; the data on the number of people in the bus is determined based on an onboard device of the bus; The first number of people, the second number of people, and the third number of people in each direction of the intersection are summed to determine the total number of people queuing in each direction of the intersection.

3. The method according to claim 1, characterized in that According to the total number of people queuing in each direction, the traffic light duration corresponding to each direction is allocated using a preset duration allocation model, including: According to the total number of people queuing in each direction, the total number of people at the intersection and the dynamic adjustment coefficient of each direction are determined; the dynamic adjustment coefficient is determined comprehensively based on the traffic congestion index of each direction, the weather conditions at the intersection, the current season, and the travel habits of the time period passing through the intersection; the traffic congestion index is calculated comprehensively based on the total number of people queuing in each direction, the length of the current vehicle queue and the speed data; Determine the traffic demand value in each direction according to the ratio of the total number of people in the queue to the total number of people at the intersection; The duration allocation model is constructed by multiplying the traffic demand value, the dynamic adjustment coefficient and the preset basic duration; The duration allocation model is used to allocate the traffic light duration corresponding to each direction.

4. The method according to claim 1, characterized in that: After allocating the signal light duration corresponding to each direction according to the total number of people queuing in each direction using a preset duration allocation model, the method further includes: When the emergency vehicle detection device detects that the emergency vehicle is at a preset distance from the intersection, or receives a limited communication request signal sent by the emergency vehicle, obtaining the driving direction and speed of the emergency vehicle; Determining an optimal signal light switching scheme according to the driving direction and speed of the emergency vehicle; According to the optimal signal light switching scheme, the green light passage time of the signal light is extended for the intersection that the emergency vehicle needs to pass through, and after the emergency vehicle passes the current intersection, the step of allocating the signal light duration corresponding to each direction is performed according to the total number of people queuing in each direction using a preset duration allocation model.

5. The method according to claim 1, characterized in that: After allocating the signal light duration corresponding to each direction according to the total number of people queuing in each direction using a preset duration allocation model, the method further includes: Obtaining the traffic demands of vehicles and pedestrians in different directions, as well as the road topology of the intersection and the traffic light topology corresponding to the intersection; Determining the traffic priority of each direction according to the traffic demand, the road topology and the signal light topology; According to the traffic priority, determining whether the current phase sequence of the signal light is conducive to smooth traffic flow; If it is not conducive to the smooth flow of traffic, the phase sequence optimization algorithm is triggered to redetermine the phase sequence.

6. The method according to claim 1, characterized in that After allocating the signal light duration corresponding to each direction according to the total number of people queuing in each direction using a preset duration allocation model, the method further includes: Obtaining historical traffic data of the intersection; According to the multi-source data and the historical traffic data, the time series analysis algorithm and the neural network algorithm model are input to predict the traffic flow of the intersection, and the prediction result of the intersection after the preset prediction time is determined; According to the prediction results, calculate the basic duration, dynamic adjustment coefficient and phase sequence of the signal light timing; The signal light is adjusted after the preset time according to the basic duration, the dynamic adjustment coefficient and the phase sequence of the signal light.

7. The method according to claim 1, characterized in that The method further comprises: Receive the vehicle's own driving intention and location information sent by the vehicle; Analyze and process the received vehicle driving intention and location information to determine an analysis result; the analysis result includes the driving status of the vehicle and the estimated time of arrival at the intersection; According to the analysis result and the duration allocation model, the switching time and duration of the signal light are adjusted in real time.

8. A vehicle-road cooperative intersection signal light control device, characterized in that: include: A first receiving module is used to receive multi-source data uploaded by a roadside device; the multi-source data includes at least one of detection data of ordinary passenger cars in each direction of the intersection, pedestrian data, and data on the number of people in a bus received by the roadside device through multi-device collaborative collection; A first determination module, used to determine the total number of people queuing in each direction of the intersection based on the multi-source data; The first processing module is used to allocate the traffic light duration corresponding to each direction according to the total number of people queuing in each direction using a preset duration allocation model.

9. A readable storage medium having a program or instruction stored thereon, characterized in that: When the program or instruction is executed by a processor, the steps in the method according to any one of claims 1 to 7 are implemented.

10. A computer program product, characterized in that The method comprises computer instructions, which, when executed by a processor, implement the steps of the method according to any one of claims 1 to 7.

Images