A Hybrid Traffic State Estimation Method Based on Vehicle-Road Integration

Abstract

This invention discloses a hybrid traffic state estimation method based on vehicle-road integration, relating to the field of intelligent traffic control technology. The method includes: acquiring vehicle-side measurement data and roadside measurement data for the traffic state of a given period on a highway, respectively; substituting these data into vehicle-side and roadside state observers to obtain vehicle-side and roadside estimates of the traffic state for the current period; fusing the vehicle-side and roadside estimates as prior estimates of the traffic state for the current period using vehicle-side and roadside Kalman filters, and then performing posterior estimates of the traffic state for the current period; fusing the two posterior estimates to obtain a hybrid estimate of the traffic state for the current period; fusing the hybrid estimate with the vehicle-side and roadside measurement data for the current period and then inputting it into the state observer to obtain vehicle-side and roadside estimates of the traffic state for the next period, and performing estimations for subsequent periods. This embodiment can improve the accuracy of traffic state estimation.

Description

Technical Field

[0001] This invention relates to the field of intelligent traffic control technology, and in particular to a hybrid traffic state estimation method based on vehicle-road integration. Background Technology

[0002] With the continuous development of autonomous driving technology, the autonomy of vehicles is gradually improving. Vehicles are increasingly used as mobile sensors to collect road information and obtain more accurate perception information. However, due to blind spots in the distribution of sensors, it is currently difficult to obtain accurate and real-time information for some road sections.

[0003] In the prior art, patent application CN117496706A provides a design method and device for a hybrid traffic flow state estimator in an intelligent connected environment, and patent application CN113034904A provides a traffic state estimation method and device based on ETC data. However, these methods do not utilize information from vehicle-side sensors and do not fully consider the inherent defects of various state estimators, resulting in limited accuracy in traffic state estimation.

[0004] In view of this, the present invention is proposed. Summary of the Invention

[0005] This invention provides a hybrid traffic state estimation method based on vehicle-road integration to solve the technical problems described in the background.

[0006] In a first aspect, embodiments of the present invention provide a hybrid traffic state estimation method based on vehicle-road integration, comprising:

[0007] S10. Acquire vehicle-side measurement data and roadside measurement data of the traffic conditions on the highway for one cycle from the vehicle-side sensor and the roadside sensor, respectively.

[0008] S20. Substitute the vehicle-end measurement data as the vehicle-end measurement output of the previous cycle into the vehicle-end state observer to obtain the vehicle-end estimate of the traffic state in the current cycle; substitute the roadside measurement data as the roadside measurement output of the previous cycle into the roadside state observer to obtain the roadside estimate of the traffic state in the current cycle.

[0009] S30. The vehicle-side estimation and the roadside estimation are fused together as the prior estimation of the current period's traffic state by the vehicle-side Kalman filter and the roadside Kalman filter, and the vehicle-side Kalman filter and the roadside Kalman filter respectively perform the posterior estimation of the current period's traffic state.

[0010] S40. The posterior estimates of the vehicle-side Kalman filter and the roadside Kalman filter are fused to obtain a hybrid estimate of the traffic state for the current period.

[0011] S50. After fusing the hybrid estimate with the vehicle-side measurement data of the current period's traffic state, the vehicle-side measurement output of the current period is substituted into the vehicle-side state observer to obtain the vehicle-side estimate of the traffic state of the next period; after fusing the hybrid estimate with the roadside measurement data of the current period's traffic state, the roadside measurement output of the current period is substituted into the roadside state observer to obtain the roadside estimate of the traffic state of the next period.

[0012] S60. Based on the vehicle-side estimation and roadside estimation of the next cycle, return to S30, and repeat the process to obtain the mixed estimation of traffic conditions for each cycle.

[0013] In a second aspect, embodiments of the present invention provide an electronic device, the electronic device comprising:

[0014] One or more processors;

[0015] Memory, used to store one or more programs.

[0016] When the one or more programs are executed by the one or more processors, the one or more processors implement the hybrid traffic state estimation method based on vehicle-road integration as described in the first aspect.

[0017] Thirdly, embodiments of the present invention also provide a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the hybrid traffic state estimation method based on vehicle-road integration described in the first aspect.

[0018] In summary, this invention provides a hybrid traffic state estimation method based on vehicle-road integration. Its advantages include: it can accurately estimate the real traffic state of highways by utilizing both vehicle-mounted and roadside sensors to collect traffic information in mixed traffic flow scenarios, and by fusing a state observer and a Kalman filter. Specifically, in the fusion of the state observer and the Kalman filter, this embodiment fully considers their inherent problems in state estimation. Through cycle-by-cycle closed-loop feedback between the state observer and the Kalman filter, it overcomes their inherent shortcomings and improves the accuracy of state estimation. Specifically:

[0019] Conventional Kalman filters rely on open-loop models for prior estimation, which cannot be corrected by incorporating actual measurements. This leads to significant initial estimation errors, impacting estimation accuracy and convergence speed. To address this deficiency, this embodiment uses the traffic state estimate from the state observer as the prior estimate for the Kalman filter, introducing measured information from the state observer into the Kalman filter and avoiding errors introduced by the model prior estimate. In contrast, patent applications CN117496706A and CN113034904A fuse the state observer's estimate with the Kalman filter's model prior state estimate as the final pre-estimate (i.e., the final prior estimate). However, the Kalman filter's model prior state estimate is still based on a mixed traffic flow model, and its error remains uncontrollable, interfering with the measured information from the state observer. In this embodiment, the state observer estimate after vehicle-road fusion is directly used as the prior estimate of the Kalman filter. It integrates the vehicle-side measured information and the roadside measured information, which can improve the accuracy of the prior estimate and the final estimate of the Kalman filter and speed up the convergence speed.

[0020] However, when traffic sensors are missing or malfunctioning in certain road sections, the state observers receive no measurements. This lack of observation data limits the accuracy of state estimation, leading to inaccurate estimations and slow convergence. To address this deficiency, this embodiment feeds back the hybrid estimate, which combines vehicle-side and roadside data, to two types of state observers as a supplement to traffic sensor measurements. In road sections lacking sensor data, the hybrid estimate fills the gaps in measurements; in road sections with available sensor data, the fusion of the hybrid estimate and sensor data further improves the accuracy of the measurement information.

[0021] This process is repeated continuously, and through the closed-loop feedback of the state observer and Kalman filter in each cycle, the accuracy and convergence speed of the state estimation are continuously improved, thus realizing the integrated vehicle-road hybrid traffic state estimation. Attached Figure Description

[0022] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0023] Figure 1 This is a schematic diagram of a highway cell partitioning provided in an embodiment of the present invention;

[0024] Figure 2 This is a schematic diagram of the deployment of various sensors on a highway according to an embodiment of the present invention;

[0025] Figure 3 This is a flowchart of a hybrid traffic state estimation method based on vehicle-road integration provided in an embodiment of the present invention;

[0026] Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention;

[0027] Figure label:

[0028] 1-Video sensing device, 2-ETC gantry, 3-Roadside unit, 4-Roadside computing facility, 5-Autonomous vehicle. Detailed Implementation

[0029] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below. Obviously, the described embodiments are only a part of the embodiments of this invention, and not all of them. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0030] In the description of this invention, it should be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing the invention and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0031] In the description of this invention, it should also be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0032] This invention provides a hybrid traffic state estimation method based on vehicle-road integration. It utilizes both vehicle-mounted and roadside sensors to collect road segment information and integrates a state observer and a Kalman filter to jointly estimate the traffic state of highways. To illustrate this method, the hybrid state estimator supporting its implementation will be described first. In the field of traffic state estimation, both state observers and Kalman filters are commonly used state estimators. The former has a relatively simple structure, while the latter offers relatively higher accuracy. Therefore, this embodiment combines the advantages of both to construct a hybrid state estimator to achieve accurate estimation of the mixed traffic flow state of highways under different penetration rates of autonomous vehicles.

[0033] In one specific embodiment, the method for constructing the hybrid state estimator may include the following steps:

[0034] S110, High-speed cell partitioning.

[0035] This step divides the highway into several segments based on the locations of entrance and exit ramps, changes in curvature radius, and changes in lane number. Each segment is called a cell, and the divided cells are sequentially numbered to facilitate the marking of sensor deployment locations. Optionally, combined with... Figure 1 The location of entrance and exit ramps can be used as the boundary between two adjacent cells, and the location where the number of lanes changes can also be used as the boundary between two adjacent cells.

[0036] Furthermore, highways are typically equipped with various sensors to measure traffic conditions in real time. Figure 2 For example, these sensors may include: a video sensing device 1, used to collect traffic flow on the road and automatically convert and calculate traffic flow density; a roadside unit 3 installed on the ETC gantry 2, which is a device in the ETC system that communicates with the on-board unit to realize vehicle identification, and can obtain information such as vehicle speed and traffic flow through this device; and an autonomous vehicle 5, used to collect vehicle speed.

[0037] In addition, roadside computing facilities 4 are deployed along the highway. These facilities can store information collected by different types of sensors, perform operations in subsequent hybrid state observers and Kalman filters, and transmit the final estimation results to the central control center.

[0038] Modeling of mixed traffic flow on S120 expressway.

[0039] This step establishes a mixed traffic flow model for the highway based on the cells defined above. Taking traffic flow density as the traffic state variable as an example, the specific structure of the model is as follows:

[0040] (1)

[0041] in, The dimension is The traffic state vector (i.e., the traffic flow density vector). The number of cells representing the highway can be obtained by arranging the traffic flow density of each cell according to its cell number. This embodiment measures and estimates traffic conditions periodically. and They represent the first The first cycle and the first Traffic state vector for each cycle; The dimension is The system control input, The quantity of external inputs representing traffic flow density (e.g., if a highway has 3 vehicle merging points, then...) ); The dimension is Traffic sensor measurement output vector, The number of sensors is represented by the traffic flow density measured by each sensor, which is then arranged according to the cell number to obtain the result. (Among them, cells without sensors do not participate in the arrangement, for example) If only the 1st, 3rd, and 5th cells contain sensors, then these three data points are arranged into a 3D vector according to the cell order. Indicates the The measurement output vector of a traffic sensor in a given cycle; Represents the system matrix; Represents the input matrix; This represents the output matrix (its specific value depends on the number and location of the traffic sensors deployed). It is a constant matrix; Indicates the The penetration rate of autonomous vehicles over a period of time , , and Let represent the system matrix, input matrix, output matrix, and constant matrix under different autonomous vehicle penetration rates, respectively; Indicates the System noise per cycle, Indicates the The measurement noise for each cycle, both system noise and system noise are Gaussian white noise.

[0042] Furthermore, in this embodiment, traffic sensors on highways are divided into two categories: vehicle-side sensors and roadside sensors. Therefore, there are two corresponding output equations, as follows:

[0043] (2)

[0044] in, and These are the output matrices corresponding to the vehicle-mounted sensors and roadside sensors, respectively, with matrix dimensions of [missing information]. , Indicates the The measurement output vector of the vehicle-side sensor in each cycle. Indicates the The measurement output vector of the roadside sensor for each cycle.

[0045] Since different types of sensors collect different traffic information, they can be uniformly converted into traffic flow density through the conversion relationship between traffic parameters. For example, autonomous vehicles and on-board sensors, acting as mobile sensors, calculate the velocity of their respective cells, which is then converted into the traffic flow density of those cells as measured values. Therefore, the output matrices corresponding to different types of sensors can adopt a unified representation method, as follows:

[0046] (3)

[0047] in, express Middle and the first The element corresponding to the nth cell, that is, as long as the nth cell is... Each cell is equipped with a corresponding sensor, and the sensor is in The order of the sensors is as follows: ,but The Middle Line number The element of the column is 1 if it is not 0; similarly, express Middle and the first The specific values ​​of the elements corresponding to each cell are... similar.

[0048] S130, Determine the conversion between measured values ​​and traffic flow density.

[0049] First, we will introduce the conversion between roadside sensor measurements and traffic flow density.

[0050] 1) For cells with ETC gantries at both the entrance and exit: the traffic flow density of the cell can be calculated from the traffic flow collected by the ETC at the entrance and exit within one cycle, i.e.:

[0051] (4)

[0052] in, Indicates the Traffic flow density per cell T This indicates the duration of each cycle (in seconds). and They represent the first The cell in the first The first cycle and the first Vehicle density per cycle, Indicates the The length of a single cell (unit: meter). Indicates entering the first cycle within a period Traffic flow per cell Indicates the outflow within a period. Traffic flow per cell.

[0053] 2) For roadside video surveillance equipment: On average, one set of high-definition video surveillance equipment is deployed per kilometer along the highway. The traffic flow density per cell is calculated as follows:

[0054] (5)

[0055] in, It is relative to the first Traffic flow density collected by video surveillance equipment located upstream of the unit cell. It is relative to the first Traffic flow density obtained by video surveillance equipment located downstream of each cell.

[0056] The following explains the conversion between vehicle-side sensor measurements and traffic flow density. For a cell containing autonomous vehicles, the position and speed information of each vehicle can be obtained, and then the average speed of that cell can be calculated, i.e.:

[0057] (6)

[0058] in, Indicates the The average velocity of each cell b For the first The number of autonomous vehicles per cell For the first The speed of an autonomous vehicle.

[0059] Then, based on the relationship between velocity and density, the traffic flow density of the cell is further calculated.

[0060] S140, Construction of Hybrid State Observer.

[0061] After completing the highway modeling and output matrix design, this step constructs hybrid state observers for both the vehicle-side and roadside sensors. The hybrid state observer based on vehicle-side sensors (hereinafter referred to as the vehicle-side state observer) is detailed below:

[0062] (7)

[0063] in, It is the traffic state vector of the vehicle-side state observer. and They represent the first The first cycle and the first Traffic state vector of the vehicle-side state observer in each cycle (also known as vehicle-side estimation); It is the first The output vector of the vehicle-end state observer for each cycle This represents the gain matrix of the vehicle-end state observer.

[0064] The hybrid state observer based on roadside sensors (hereinafter referred to as the roadside state observer) is as follows:

[0065] (8)

[0066] in, This represents the traffic state vector of the roadside state observer. and They represent the first The first cycle and the first Traffic state vectors of roadside state observers in each cycle (also known as roadside estimates); It is the first The output vector of the roadside state observer for each cycle This represents the gain matrix of the roadside state observer.

[0067] Generally, vehicle-mounted sensors can obtain accurate information about their current cell, and then accurately estimate the traffic flow density of the current road segment through the vehicle-mounted state observer. However, they struggle to acquire information beyond the line-of-sight. Roadside sensors, on the other hand, can acquire information about the current cell and transmit this information to vehicles in the upstream road segment, thus compensating for the shortcomings of vehicle-mounted sensors. Combining the advantages of both types of sensors, this embodiment fuses the estimation results from the vehicle-mounted state observer and the roadside state observer to obtain a more accurate traffic flow density. This estimation result is then transmitted to the upstream road segment as beyond-line-of-sight information for the upstream segment, enabling precise implementation of the control strategy.

[0068] In one specific implementation, the fusion of the estimation results from two state observers may include the following steps:

[0069] Step 1: Assume that vehicle-side estimation and roadside estimation each have a weight. and Then weighted combination pairs can be generated. The weighted summation result of traffic flow density for:

[0070] (9)

[0071] Step 2: Calculate the vehicle-side estimation separately. and roadside estimation and The difference and :

[0072] (10)

[0073] Step 3: Perform cyclical calculations within each period to determine the standard deviation. The weight combination pair corresponding to the minimum value As the final weight combination:

[0074] (11)

[0075] Step 4: Substitute the final weight combination into formula (9) to calculate the estimated traffic flow. .

[0076] S150, constructed using a hybrid Kalman filter.

[0077] While constructing the hybrid state observer, a hybrid Kalman filter based on vehicle-end sensors (hereinafter referred to as the vehicle-end Kalman filter) and a hybrid Kalman filter based on roadside sensors (hereinafter referred to as the roadside Kalman filter) can be constructed simultaneously. Details are as follows:

[0078] 1) In a typical Kalman filter, the traffic state of the cells is estimated a priori based on the constructed hybrid traffic flow model. However, since the hybrid traffic flow model itself is an open-loop system, it lacks real-time feedback and is easily affected by external interference, leading to excessive estimation errors. Therefore, this embodiment uses the result of fusing the vehicle-side estimation and roadside estimation as the prior estimate of the Kalman filter, that is:

[0079] (12)

[0080] in, Indicates the Kalman filter for the first Prior estimates of traffic conditions for each cycle.

[0081] 2) Calculate the prior error covariance matrix of the vehicle-end Kalman filter and the roadside Kalman filter, respectively. and :

[0082] (13)

[0083] in, This indicates that the vehicle-side Kalman filter is at the 1st... The posterior error covariance matrix of the nth period (derived from the nth period) (Updated every cycle) This indicates that the vehicle-side Kalman filter is at the 1st... The prior error covariance matrix for each period; Indicates the roadside Kalman filter at the 1st The posterior error covariance matrix for each period, Indicates the roadside Kalman filter at the 1st The prior error covariance matrix for each period, Q This indicates system noise.

[0084] 3) Calculate the gain matrices of the vehicle-end Kalman filter and the roadside Kalman filter respectively. and :

[0085] (13-1)

[0086] in, and This indicates that the vehicle-side Kalman filter and the roadside Kalman filter are respectively in the first... Gain matrix for each period, and These represent the output matrices of the vehicle-side Kalman filter and the roadside Kalman filter, respectively. In this embodiment, , , R This indicates measurement noise.

[0087] 4) Based on the prior estimates of the two types of Kalman filters respectively and Prior error covariance matrix , and the gain matrix , Update the posterior state estimate to obtain the final estimates for the two classes of Kalman filters:

[0088] (14)

[0089] in, and These represent the effects of the vehicle-end Kalman filter on the first... Prior and posterior estimates of traffic conditions for each cycle In the vehicle-side Kalman filter, the first... Vehicle-end measurement output for each cycle; and These represent the roadside Kalman filters for the first... Prior and posterior estimates of traffic conditions for each cycle In the roadside Kalman filter, the first... The roadside measurement output for each cycle. In this embodiment, the measurement outputs of the state observer and the Kalman filter are consistent, i.e. , .

[0090] 5) Update the posterior error covariance matrix and :

[0091] (15)

[0092] in, and These represent the first and second Kalman filters at the vehicle end and roadside respectively. The posterior error covariance matrix for each period; This represents the identity matrix, where the diagonal elements are 1 and the other elements are 0.

[0093] To obtain a more accurate density estimate, this embodiment utilizes a fusion algorithm to perform a weighted summation of the two posterior estimates obtained through the Kalman filter. Specifically:

[0094] Step 1: Assume the posterior estimate of the vehicle-end Kalman filter. Posterior estimation of roadside Kalman filters Each has a weight and Then weight combination pairs can be generated. The weighted summation result of traffic flow density for:

[0095] (16)

[0096] Step 2: Calculate separately , respectively with The difference and ,Right now:

[0097] (17)

[0098] Step 3: Perform cyclical calculations within each period to determine the standard deviation. The minimum value corresponds to the pair of value combinations As the final weight combination:

[0099] (18)

[0100] Step 4: Substitute the final weight combination into formula (16) to calculate the final estimate of traffic flow within one cycle. :

[0101] (19)

[0102] The estimated value It is the final estimate of the actual traffic state for the current period by the entire hybrid state estimator. For ease of distinction, this embodiment refers to it as the hybrid estimate.

[0103] In practical applications, the estimation of mixed traffic flow density is performed sequentially, requiring iterative calculations by the state observer and Kalman filter. To obtain a more accurate traffic flow density, the calculation result of equation (19) can be fused with the measurement data collected by sensors at the vehicle end and roadside in subsequent state estimations, serving as the feedback value of the state observer in the next cycle. That is, the calculation for the next cycle is performed based on equations (7) and (8), thereby correcting the estimation result of the state observer and obtaining more accurate state values ​​for all cells. Specifically:

[0104] The value in equation (19) can be calculated from the output matrix. Corresponding output value :

[0105] (20)

[0106] in, The measurement output matrix corresponding to expression (19) can be designed as needed to satisfy the corresponding observable judgment matrix ( , (The full number is sufficient.) For convenience, You can take the identity matrix.

[0107] Then, compare equation (20) with the vehicle-side sensor measurements shown in equation (2). Roadside sensor measurements A fusion calculation is performed to obtain a new output value. Optionally, due to the absence of traffic sensors in some cells, this embodiment employs two methods. Fusion with measured values: Taking vehicle-side sensors as an example, if the highway's first... Vehicle-side measurements of the current period's traffic state within each cell If a sensor is missing (e.g., no sensor is deployed, or the sensor is faulty), then the cell's position will be affected. Output value It is directly used as the new output value of the cell. If the highway is numbered Vehicle-side measurements of the current period's traffic state within each cell If it exists, then the cell's... Output value and The weighted average is used as the new output value for that cell. Finally, by arranging the new output values ​​of each cell in order, we obtain the new output vector. The fusion calculation for roadside sensors is similar.

[0108] For example, the overall fusion calculation of all cells can be represented as:

[0109] (twenty one)

[0110] And the fusion and Substituting equations (7) and (8), the state observer for the next cycle is corrected. The vehicle-end state observer is corrected as follows:

[0111] (twenty two)

[0112] in, Indicates the first The output corresponding to the periodic mixed estimation Indicates the first The output corresponding to the periodic mixed estimation is fused with the vehicle-side measurement data to form a new output; The fusion function is represented by the fusion operation shown in formula (21), where the data in the fusion operation are all data from the same period.

[0113] The roadside condition observer has been corrected to:

[0114] (twenty three)

[0115] in, Indicates the first The output corresponding to the hybrid estimation in each cycle is fused with the new output from the roadside measurement data. Subsequent cycle vehicle-side estimation and roadside estimation are both based on the state observers shown in equations (22) and (23). The entire process described above represents the complete hybrid state estimator.

[0116] Based on the above hybrid state estimator Figure 3 This is a flowchart illustrating a hybrid traffic state estimation method based on vehicle-road integration, provided by an embodiment of the present invention. This method is applicable to hybrid traffic flow scenarios on highways and can be executed by the aforementioned roadside computing facilities or other electronic devices. Figure 3 As shown, the method specifically includes:

[0117] S10. Acquire vehicle-side measurement data and roadside measurement data of the traffic conditions on the highway for one cycle from the vehicle-side sensor and the roadside sensor, respectively.

[0118] For example, traffic flow density is still taken as the traffic state. Various sensors collect traffic data of each cell of the highway in real time. The collected data can be converted into traffic flow density through equations (4), (5), and (6). Among them, the traffic flow density from the vehicle-end sensor is called vehicle-end measurement data, and the traffic flow density from the roadside sensor is called roadside measurement data.

[0119] In this embodiment, the state estimation is performed periodically. In each estimation, the current period is called the current period, and the period in which the last estimation occurred is called the previous period. For example, if a state estimation is performed every 10 minutes, and the current time is 8:10, the traffic state between 8:00 and 8:10 is estimated. Then, 8:00-8:10 is called the current period, and 7:50-8:00 is called the previous period.

[0120] Each estimation requires acquiring vehicle-side measurement data from the previous period and the current period, as well as roadside measurement data, to estimate the actual traffic conditions for the current period.

[0121] S20. The vehicle-side measurement data of the previous cycle is used as the vehicle-side measurement output of the previous cycle and substituted into the vehicle-side state observer to obtain the vehicle-side estimate of the traffic state of the current cycle; the roadside measurement data of the previous cycle is used as the roadside measurement output of the previous cycle and substituted into the roadside state observer to obtain the roadside estimate of the traffic state of the current cycle.

[0122] Specifically, by arranging the vehicle-side measurement values ​​within each cell of the highway in the previous cycle according to cell number (cells with sensors are included in the arrangement, while cells without sensors are excluded), the measurement vector of the vehicle-side sensors in the previous cycle can be obtained. This vector is then used as... Substituting into equation (7) and performing the calculation, we obtain... This is the vehicle-side estimate of the traffic status for the current period.

[0123] Similarly, by arranging the roadside measurement values ​​within each cell of the highway in the previous cycle according to cell number (cells with sensors are included in the arrangement, while cells without sensors are excluded), the measurement vector of the roadside sensors in the previous cycle can be obtained. This vector is then used as... Substituting into formula (8), we get This is the roadside estimate of the traffic conditions for the current period.

[0124] S30. The vehicle-side estimation and roadside estimation are fused together as the prior estimation of the current period's traffic state by the vehicle-side Kalman filter and the roadside Kalman filter, and the vehicle-side Kalman filter and the roadside Kalman filter respectively perform the posterior estimation of the current period's traffic state.

[0125] Specifically, substituting the final weight combination determined by equations (9)-(11) into equation (9), the following can be calculated: and fusion value The Kalman filter for the current period is obtained through equations (13), (13-1), and (14). As Substituting into equation (14), and simultaneously arranging the vehicle-end measurement values ​​within each cell of the highway in the current cycle in sequence, we obtain the result as... Substituting into equation (14), the calculated result is... This is the posterior estimate of the current period's traffic state by the vehicle-side Kalman filter; similarly, ... As Substituting into equation (14), and simultaneously arranging the roadside measurement values ​​within each cell of the highway in the current cycle in sequence, we obtain the following: Substituting into equation (14), the calculated result is... This is the posterior estimate of the traffic state for the current period by the roadside Kalman filter.

[0126] S40. The posterior estimates of the vehicle-side Kalman filter and the roadside Kalman filter are fused to obtain a hybrid estimate of the traffic state for the current period.

[0127] Specifically, calculate according to formula (19) and fusion results This serves as a hybrid estimate of the traffic state for the current period. This completes the first estimation of the actual traffic state of the highway by the hybrid state estimator.

[0128] S50. After fusing the hybrid estimate with the vehicle-side measurement data of the current period's traffic state, the vehicle-side measurement output of the current period is substituted into the vehicle-side state observer to obtain the vehicle-side estimate of the traffic state of the next period; after fusing the hybrid estimate with the roadside measurement data of the current period's traffic state, the roadside measurement output of the current period is substituted into the roadside state observer to obtain the roadside estimate of the traffic state of the next period.

[0129] This step marks the beginning of the next estimation by the hybrid state estimator, which involves estimating the actual traffic conditions of the highway in the next cycle. Specifically, firstly, based on equation (20), the hybrid estimator... Calculate the corresponding output value Simultaneously, vehicle-mounted sensors acquire vehicle-mounted measurement data of the highway traffic conditions for the next cycle. And roadside measurement data of highway traffic conditions for the next cycle from roadside sensors. .

[0130] Then, according to equation (21), the values ​​of each cell are... and The output of the vehicle-side state observer is obtained through fusion. ;Will As Substituting into equation (22), we obtain the new... This refers to the vehicle-side estimation of traffic conditions for the next cycle.

[0131] At the same time, according to equation (21), each cell's and The data is then fused to obtain a new output from the roadside state observer. .Will As Substituting into equation (23), we obtain the new... This refers to the roadside estimation of traffic conditions for the next cycle.

[0132] S60. Based on the vehicle-side estimation and roadside estimation of the next cycle, return to S30. Repeat this process to obtain a mixed estimate of the traffic state for each cycle.

[0133] Specifically, return to S30 and repeat the operations from S30 to S60, that is, merge the vehicle-side estimate of the traffic state in the next cycle and the roadside estimate of the traffic state in the next cycle as the prior estimate of the traffic state in the next cycle by the vehicle-side Kalman filter and the roadside Kalman filter; and then perform posterior estimates of the traffic state in the next cycle by the two Kalman filters respectively.

[0134] The posterior estimates from the two Kalman filters are fused to obtain a hybrid estimate of the traffic state for the next cycle. This completes the current estimation by the hybrid state estimator.

[0135] Next, the hybrid state estimator continues with subsequent estimations. The hybrid estimate of the next cycle is fused with the vehicle-side measurement data of the next cycle, and this fusion is used as the vehicle-side measurement output of the next cycle. This output is then substituted into the vehicle-side state observer to obtain the vehicle-side estimate for the cycle after that. Similarly, the hybrid estimate of the next cycle is fused with the roadside measurement data of the next cycle, and this fusion is used as the roadside measurement output of the next cycle. This output is then substituted into the roadside state observer to obtain the roadside estimate for the cycle after that.

[0136] Based on the vehicle-side and roadside estimates for the next cycle, return to S30, and repeat the operations from S30 to S60 again, thus obtaining the hybrid estimate for each cycle. It can be seen that S10 and S20 are executed only in the first state estimate, while S30 to S60 are executed in the first state estimate and in every subsequent state estimate.

[0137] In summary, this embodiment provides a hybrid traffic state estimation method based on vehicle-road integration. In the mixed traffic flow scenario of highways, it utilizes both vehicle-mounted sensors and roadside sensors to collect traffic information and integrates a state observer and a Kalman filter to estimate the actual traffic state of the highway. Specifically, in the fusion of the state observer and the Kalman filter, this embodiment fully considers their inherent problems in state estimation. Through cycle-by-cycle closed-loop feedback between the state observer and the Kalman filter, it overcomes their inherent shortcomings and improves the accuracy of state estimation. Specifically:

[0138] Conventional Kalman filters rely on open-loop models for prior estimation, which cannot be corrected by incorporating actual measurements. This leads to significant initial estimation errors, impacting estimation accuracy and convergence speed. To address this deficiency, this embodiment uses the traffic state estimate from the state observer as the prior estimate for the Kalman filter, introducing measured information from the state observer into the Kalman filter and avoiding errors introduced by the model prior estimate. In contrast, patent applications CN117496706A and CN113034904A fuse the state observer's estimate with the Kalman filter's model prior state estimate as the final pre-estimate (i.e., the final prior estimate). However, the Kalman filter's model prior state estimate is still based on a mixed traffic flow model, and its error remains uncontrollable, interfering with the measured information from the state observer. In this embodiment, the state observer estimate after vehicle-road fusion is directly used as the prior estimate of the Kalman filter. It integrates the vehicle-side measured information and the roadside measured information, which can improve the accuracy of the prior estimate and the final estimate of the Kalman filter and speed up the convergence speed.

[0139] However, when traffic sensors are missing or malfunctioning in certain road sections, the state observers receive no measurements. This lack of observation data limits the accuracy of state estimation, leading to inaccurate estimations and slow convergence. To address this deficiency, this embodiment feeds back the hybrid estimate, which combines vehicle-side and roadside data, to two types of state observers as a supplement to traffic sensor measurements. In road sections lacking sensor data, the hybrid estimate fills the gaps in measurements; in road sections with available sensor data, the fusion of the hybrid estimate and sensor data further improves the accuracy of the measurement information.

[0140] This process is repeated continuously, and through the closed-loop feedback of the state observer and Kalman filter in each cycle, the accuracy and convergence speed of the state estimation are continuously improved, thus realizing the integrated vehicle-road hybrid traffic state estimation.

[0141] Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention, such as... Figure 4As shown, the device includes a processor 60, a memory 61, an input device 62, and an output device 63; the number of processors 60 in the device can be one or more. Figure 4 Taking a processor 60 as an example; the processor 60, memory 61, input device 62, and output device 63 in the device can be connected via a bus or other means. Figure 4 Taking the example of a connection between China and Israel via a bus.

[0142] The memory 61, as a computer-readable storage medium, can be used to store software programs, computer-executable programs, and modules, such as the program instructions / modules corresponding to the vehicle-road integrated hybrid traffic state estimation method in this embodiment of the invention. The processor 60 executes various functional applications and data processing of the device by running the software programs, instructions, and modules stored in the memory 61, thereby realizing the aforementioned vehicle-road integrated hybrid traffic state estimation method.

[0143] The memory 61 may primarily include a program storage area and a data storage area. The program storage area may store the operating system and at least one application program required for a given function; the data storage area may store data created based on terminal usage. Furthermore, the memory 61 may include high-speed random access memory and non-volatile memory, such as at least one disk storage device, flash memory, or other non-volatile solid-state storage device. In some instances, the memory 61 may further include memory remotely located relative to the processor 60, which can be connected to the device via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0144] Input device 62 can be used to receive input digital or character information, and to generate key signal inputs related to user settings and function control of the device. Output device 63 may include display devices such as a display screen.

[0145] This invention also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the vehicle-road integrated hybrid traffic state estimation method of any embodiment.

[0146] The computer storage medium of this invention can be any combination of one or more computer-readable media. A computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. A computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of computer-readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this document, a computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0147] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, capable of sending, propagating, or transmitting programs for use by or in connection with an instruction execution system, apparatus, or device.

[0148] Program code contained on a computer-readable medium may be transmitted using any suitable medium, including but not limited to wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0149] Computer program code for performing the operations of this invention can be written in one or more programming languages ​​or a combination thereof. Programming languages ​​include object-oriented programming languages—such as Java, Smalltalk, and C++—as well as conventional procedural programming languages—such as C or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0150] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the technical solutions of the embodiments of the present invention.

Claims

1. A hybrid traffic state estimation method based on vehicle-road integration, characterized in that, include: S10. Acquire vehicle-side measurement data and roadside measurement data of the traffic conditions on the highway for one cycle from the vehicle-side sensor and the roadside sensor, respectively. S20. Substitute the vehicle-end measurement data as the vehicle-end measurement output of the previous cycle into the vehicle-end state observer to obtain the vehicle-end estimate of the traffic state in the current cycle. The roadside measurement data is substituted into the roadside state observer as the roadside measurement output of the previous cycle to obtain the roadside estimate of the traffic state in the current cycle. S30. The vehicle-side estimation and the roadside estimation are fused together as the prior estimation of the current period's traffic state by the vehicle-side Kalman filter and the roadside Kalman filter, and the vehicle-side Kalman filter and the roadside Kalman filter respectively perform the posterior estimation of the current period's traffic state. S40. The posterior estimates of the vehicle-side Kalman filter and the roadside Kalman filter are fused to obtain a hybrid estimate of the traffic state for the current period; the hybrid estimate is the final estimate of the actual traffic state for the current period. S50. After fusing the hybrid estimate with the vehicle-side measurement data of the current period's traffic state, the vehicle-side measurement output of the current period is substituted into the vehicle-side state observer to obtain the vehicle-side estimate of the traffic state of the next period; after fusing the hybrid estimate with the roadside measurement data of the current period's traffic state, the roadside measurement output of the current period is substituted into the roadside state observer to obtain the roadside estimate of the traffic state of the next period. S60. Based on the vehicle-side estimation and roadside estimation of the next cycle, return to S30, and repeat the process to obtain the mixed estimation of traffic conditions for each cycle. In this context, the variable for each traffic state is traffic flow density, and the vehicle-side state observer is: in, and They represent the first The first cycle and the first Vehicle-side estimation of traffic conditions over a given period Indicates the first The output of the vehicle-end state observer in each cycle Indicates the first Vehicle-end measurement output for each cycle; Indicates the first System control input for each cycle; Indicates the first The penetration rate of autonomous vehicles over a period of time , , and They represent The corresponding system matrix, input matrix, constant matrix, and gain matrix; This represents the output matrix based on vehicle-mounted sensors; Accordingly, the step of fusing the hybrid estimate with the vehicle-mounted measurement data of the current period's traffic state, and then substituting this into the vehicle-mounted state observer as the vehicle-mounted measurement output of the current period to obtain the vehicle-mounted estimate of the traffic state for the next period, includes: The hybrid estimate is calculated based on the output matrix. Corresponding output ; The output Vehicle-side measurement data related to the current period's traffic conditions By merging, a new output is obtained. ,in, , This represents the fusion function; specifically, if the highway's first... The cell contains no vehicle-side sensors, and the first... individual cells Output value As the first The new output value of each cell If the highway is numbered Each cell contains vehicle-side sensors, which will... individual cells Output value With the Vehicle-side measurements of the current period's traffic status for each cell. Perform a weighted average to obtain the first... The new output value of each cell Arrange the new output values ​​of each cell in order to obtain the new output. ; The vehicle-side state observer is corrected as follows: in, Indicates the first The output corresponding to the mixed estimation of each period, Indicates the first The output corresponding to the hybrid estimation of each cycle is the new output after fusing the vehicle-side measurement data; The new output As Substituting the corrected vehicle-end state observer, we obtain This is the vehicle-side estimate of traffic conditions for the next cycle.

2. The method according to claim 1, characterized in that, Prior to S10, it also included: Divide the highway into cells; A mixed traffic flow model for highways is established based on the partitioned cells. Based on the number and location of vehicle-mounted sensors and the number and location of roadside sensors, the output matrix in the hybrid traffic flow model is constructed; Based on the hybrid traffic flow model and output matrix, a vehicle-side state observer based on vehicle-side sensors and a roadside state observer based on roadside sensors are constructed respectively. Based on the hybrid traffic flow model and output matrix, a vehicle-side Kalman filter based on vehicle-side sensors and a roadside Kalman filter based on roadside sensors are constructed respectively.

3. The method according to claim 1, characterized in that, The step of substituting the vehicle-end measurement data as the vehicle-end measurement output of the previous cycle into the vehicle-end state observer to obtain the vehicle-end estimate of the traffic state for the current cycle includes: After arranging the vehicle-end measurement values ​​within each cell of the highway in sequence, as... Substituting the data into the vehicle-side state observer, we obtain... This is the vehicle-side estimate of the traffic status for the current period.

4. The method according to claim 1, characterized in that, The vehicle-side Kalman filter is: in, and These represent the effects of the vehicle-end Kalman filter on the first... Prior and posterior estimates of traffic conditions for each cycle This indicates that the vehicle-side Kalman filter is at the 1st... Gain matrix for each period, This represents the output matrix of the vehicle-end Kalman filter; Indicates the first The penetration rate of autonomous vehicles over a period of time express The corresponding constant matrix; In the vehicle-side Kalman filter, the first... Vehicle-end measurement output for each cycle; Accordingly, S30 includes: The vehicle-side estimation and roadside estimation are fused, and the fusion result is used as... Substitute the vehicle-side Kalman filter; Arrange the vehicle-end measurement values ​​within each cell of the highway in the current period in order as follows: Substitute the vehicle-side Kalman filter; Calculate based on the formula after substitution. This is the posterior estimate of the traffic state for the current period by the vehicle-side Kalman filter.

5. The method according to claim 1, characterized in that, The fusion of the vehicle-side estimation and the roadside estimation includes: Multiple sets of weighted parameters are used to iteratively weight and fuse the vehicle-side estimation and the roadside estimation to obtain multiple fusion result sequences; The combination of weighted parameters corresponding to the fusion result sequence with the smallest standard deviation is taken as the optimal combination of weighted parameters; Based on the optimal weighted parameter combination, the vehicle-side estimation and roadside estimation are finally weighted and fused.

6. An electronic device, characterized in that, include: one or more processors; Memory, used to store one or more programs. When the one or more programs are executed by the one or more processors, the one or more processors implement the hybrid traffic state estimation method based on vehicle-road integration as described in any one of claims 1-5.

7. A computer-readable storage medium, characterized in that, It stores a computer program, which, when executed by a processor, implements the hybrid traffic state estimation method based on vehicle-road integration as described in any one of claims 1-5.

Images