An intelligent vehicle ramp merging control method based on driving style recognition

Abstract

The present invention provides a method for controlling ramp merging of intelligent vehicles based on driving style recognition. The method establishes multiple discrete points in a three-dimensional coordinate system based on historical motion state data of manually driven vehicles in different merging zones, clusters the multiple discrete points, and obtains multiple cluster sets and multiple cluster centers. Driving styles and motion control coefficients corresponding to the multiple cluster sets are set. A kinematic model of the intelligent vehicle based on kinematics is established. According to the motion control coefficients, the objective function is adjusted to obtain motion control instructions, and motion control is performed on the intelligent vehicle. The present invention improves the accuracy of motion control of the intelligent vehicle and effectively avoids vehicle collisions in ramp merging zones.

Description

Technical Field

[0001] The present invention relates to the field of intelligent vehicle ramp merging control, and in particular to an intelligent vehicle ramp merging control method based on driving style recognition. Background Art

[0002] Although the development of intelligent connected technology has led to an increasing number of intelligent vehicles taking to the roads, the current situation of mixed traffic between intelligent and human-driven vehicles will persist for a long time to come. In mixed traffic flows, the driving behavior of human-driven vehicles varies over time and space, influenced by the driver's habits, physiology, and decision-making awareness. For example, an aggressive human-driven vehicle entering a merging area may evolve to a moderate driving style as the merging process progresses, while a conservative human-driven vehicle may develop an aggressive driving style if it refuses to slow down and yield during the merging process. This temporal and spatial shift in the driving style of human-driven vehicles during the merging process increases the difficulty of accurate control for intelligent vehicles. This is particularly true in merging areas on expressways, where high vehicle speeds and the random variability of merging decisions by human-driven vehicles can lead to collisions if intelligent vehicles fail to adjust their driving behavior in a timely manner. Therefore, precisely controlling intelligent vehicles in mixed traffic flows such as ramp merging scenarios, adapting to the temporal and spatial variations in the driving style of human-driven vehicles, is key to improving traffic efficiency and reducing accident rates. Summary of the Invention

[0003] To address the above issues, the present invention proposes an intelligent vehicle ramp merging control method based on driving style recognition, which specifically includes the following steps:

[0004] S1. The intelligent vehicle collects historical motion status data of manually driven vehicles in different incoming zones of the incoming area;

[0005] Each historical motion status data includes: following distance ,speed , acceleration ;

[0006] Each historical motion state data corresponds to a discrete point in a three-dimensional coordinate system;

[0007] S2. Perform Gaussian mixture clustering on multiple discrete points in different input partitions to obtain three cluster sets and three cluster centers in each input partition;

[0008] Among them, each cluster set corresponds to a driving style;

[0009] Label each discrete point with a driving style;

[0010] S3. According to the three cluster centers in each incoming partition, the motion control coefficients of different cluster sets within the incoming partition are calibrated to obtain motion control coefficients corresponding to different cluster sets within each incoming partition;

[0011] S4. For each incoming partition, the discrete points in each cluster set of the three cluster sets corresponding to the incoming partition are divided into training samples and test samples in a ratio of 7:3;

[0012] Training a driving style recognition model using training samples, and testing the driving style recognition model using test samples to obtain a trained driving style recognition model;

[0013] S5. The intelligent vehicle obtains the motion state data of the manually driven vehicle at the current moment and identifies the driving style of the manually driven vehicle at the current moment through the driving style recognition model;

[0014] S6. The intelligent vehicle calculates the motion control instructions for the intelligent vehicle based on the current driving style of the manually driven vehicle, the current position of the intelligent vehicle corresponding to the incoming partition, and the current driving style of the intelligent vehicle;

[0015] S7. The intelligent vehicle controls the movement of the intelligent vehicle according to the motion control instruction;

[0016] S8. Repeat steps S5 to S7 until the smart vehicle exits the merging area.

[0017] Furthermore, step S1 specifically includes:

[0018] S11. The intelligent vehicle collects historical motion status data of multiple manually driven vehicles in different incoming zones in the incoming zone;

[0019] The different import partitions in the import area include: an import start area, an import execution area, and an import end area;

[0020] Vehicles traveling in the merging area include: intelligent vehicles, manually driven vehicles;

[0021] The historical motion state data of the manually driven vehicle includes: motion state data of the manually driven vehicle collected at predetermined intervals T when the manually driven vehicle sequentially passes through the merge start area, the merge execution area, and the merge end area, forming a plurality of historical motion state data;

[0022] Each historical motion status data includes: following distance ,speed , acceleration , location g, collection time t;

[0023] S12. According to the position g in each historical motion state data, determine the historical motion state data belongs to the import partition;

[0024] S13. For each incoming partition, the historical motion state data of the incoming partition is calibrated in a three-dimensional coordinate system to form a plurality of discrete points corresponding to the incoming partition;

[0025] Among them, the coordinates of each historical motion state data in the three-dimensional coordinate system are ( , ); Each historical motion state data corresponds to a discrete point in a three-dimensional coordinate system;

[0026] S14. For each incoming partition, execute step S13 to obtain multiple discrete points corresponding to each incoming partition.

[0027] Furthermore, step S2 specifically includes:

[0028] S21. Perform Gaussian mixture clustering on the multiple discrete points corresponding to each incoming partition to obtain three cluster sets and three cluster centers in the incoming partition;

[0029] Among them, each incoming partition corresponds to three cluster sets and three cluster centers;

[0030] Each cluster set includes multiple discrete points and a cluster center;

[0031] Each cluster center includes three component data: following distance component , velocity component , acceleration component ;

[0032] S22. Setting the driving style corresponding to each cluster set;

[0033] Said driving styles include: aggressive, moderate, and conservative;

[0034] S23. Label each discrete point in each cluster set with a driving style;

[0035] When the driving style corresponding to the cluster set is an aggressive type, marking the driving style of each discrete point in the cluster set as an aggressive type;

[0036] When the driving style corresponding to the cluster set is moderate, mark the driving style of each discrete point in the cluster set as moderate;

[0037] When the driving style corresponding to the cluster set is conservative, marking the driving style of each discrete point in the cluster set as conservative;

[0038] S24. Repeat steps S21 to S23 until all three incoming partitions are executed, forming three cluster sets and three cluster centers for each incoming partition, as well as a driving style corresponding to each cluster set and a driving style for each discrete point in each cluster set. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Various other advantages and benefits will become apparent to those skilled in the art upon reading the detailed description of the preferred embodiment below. The accompanying drawings are for illustration purposes only and are not to be considered as limiting the present invention. The same reference symbols are used throughout the drawings to represent the same components. In the drawings:

[0040] Figure 1 This is a structural diagram of the ramp merging zone. DETAILED DESCRIPTION

[0041] Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments described herein. Rather, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0042] In the description of the present invention, unless otherwise expressly specified or limited, terms such as "mounted," "connected," "connect," and "fixed" should be understood in a broad sense. For example, they may refer to connection, detachable connection, or integration; mechanical connection or electrical connection; direct connection or indirect connection through an intermediate medium; and internal communication between two components or interaction between two components. Those skilled in the art will understand the specific meanings of the above terms in the present invention based on specific circumstances.

[0043] The present invention proposes an intelligent vehicle ramp merging control method based on driving style recognition, which specifically includes the following steps:

[0044] S1. The intelligent vehicle collects historical motion status data of manually driven vehicles in different incoming zones of the incoming area;

[0045] Each historical motion status data includes: following distance ,speed , acceleration ;

[0046] Each historical motion state data corresponds to a discrete point in a three-dimensional coordinate system;

[0047] S2. Perform Gaussian mixture clustering on multiple discrete points in different input partitions to obtain three cluster sets and three cluster centers in each input partition;

[0048] Among them, each cluster set corresponds to a driving style;

[0049] Label each discrete point with a driving style;

[0050] S3. According to the three cluster centers in each incoming partition, the motion control coefficients of different cluster sets within the incoming partition are calibrated to obtain motion control coefficients corresponding to different cluster sets within each incoming partition;

[0051] S4. For each incoming partition, the discrete points in each cluster set of the three cluster sets corresponding to the incoming partition are divided into training samples and test samples in a ratio of 7:3;

[0052] Training a driving style recognition model using training samples, and testing the driving style recognition model using test samples to obtain a trained driving style recognition model;

[0053] S5. The intelligent vehicle obtains the motion state data of the manually driven vehicle at the current moment and identifies the driving style of the manually driven vehicle at the current moment through the driving style recognition model;

[0054] S6. The intelligent vehicle calculates the motion control instructions for the intelligent vehicle based on the current driving style of the manually driven vehicle, the current position of the intelligent vehicle corresponding to the incoming partition, and the current driving style of the intelligent vehicle;

[0055] S7. The intelligent vehicle controls the movement of the intelligent vehicle according to the motion control instruction;

[0056] S8. Repeat steps S5 to S7 until the smart vehicle exits the merging area.

[0057] Furthermore, step S1 specifically includes:

[0058] S11. The intelligent vehicle collects historical motion status data of multiple manually driven vehicles in different incoming zones in the incoming zone;

[0059] The different import partitions in the import area include: an import start area, an import execution area, and an import end area;

[0060] Vehicles traveling in the merging area include: intelligent vehicles, manually driven vehicles;

[0061] The historical motion state data of the manually driven vehicle includes: motion state data of the manually driven vehicle collected at predetermined intervals T when the manually driven vehicle sequentially passes through the merge start area, the merge execution area, and the merge end area, forming a plurality of historical motion state data;

[0062] Each historical motion status data includes: following distance ,speed , acceleration , location g, collection time t;

[0063] S12. According to the position g in each historical motion state data, determine the historical motion state data belongs to the import partition;

[0064] S13. For each incoming partition, the historical motion state data of the incoming partition is calibrated in a three-dimensional coordinate system to form a plurality of discrete points corresponding to the incoming partition;

[0065] Among them, the coordinates of each historical motion state data in the three-dimensional coordinate system are ( , ); Each historical motion state data corresponds to a discrete point in a three-dimensional coordinate system;

[0066] S14. For each incoming partition, execute step S13 to obtain multiple discrete points corresponding to each incoming partition.

[0067] Furthermore, step S2 specifically includes:

[0068] S21. Perform Gaussian mixture clustering on the multiple discrete points corresponding to each incoming partition to obtain three cluster sets and three cluster centers in the incoming partition;

[0069] Among them, each incoming partition corresponds to three cluster sets and three cluster centers;

[0070] Each cluster set includes multiple discrete points and a cluster center;

[0071] Each cluster center includes three component data: following distance component , velocity component , acceleration component ;

[0072] S22. Setting the driving style corresponding to each cluster set;

[0073] Said driving styles include: aggressive, moderate, and conservative;

[0074] S23. Label each discrete point in each cluster set with a driving style;

[0075] When the driving style corresponding to the cluster set is an aggressive type, marking the driving style of each discrete point in the cluster set as an aggressive type;

[0076] When the driving style corresponding to the cluster set is moderate, mark the driving style of each discrete point in the cluster set as moderate;

[0077] When the driving style corresponding to the cluster set is conservative, marking the driving style of each discrete point in the cluster set as conservative;

[0078] S24. Repeat steps S21 to S23 until all three incoming partitions are executed, forming three cluster sets and three cluster centers for each incoming partition, as well as a driving style corresponding to each cluster set and a driving style for each discrete point in each cluster set.

[0079] Furthermore, step S3 specifically includes:

[0080] S31. According to the three cluster sets corresponding to each incoming partition { } and three cluster centers { }、{ }、{ }, calculate three cluster sets The corresponding motion control coefficients { }、{ }、{ };

[0081] in,{ } is the first cluster set The corresponding cluster center; These are the three component data corresponding to the cluster center; the driving style corresponding to the first cluster set is aggressive;

[0082] { } is the second cluster set The corresponding cluster center; are the three component data corresponding to the cluster center; the driving style corresponding to the second cluster set is moderate;

[0083] { } is the third cluster set The corresponding cluster center; These are the three component data corresponding to the cluster center; the driving style corresponding to the third cluster set is conservative;

[0084] in, 、 、 The motion control coefficient corresponding to the following distance component; 、 、 motion control coefficients corresponding to velocity components; 、 、 Motion control coefficient corresponding to the acceleration component;

[0085] Where center is start, cond, or end;

[0086] When center is start, the cluster set and cluster center are the cluster set and cluster center corresponding to the inbound start area;

[0087] When center is cond, the cluster set and cluster center are the cluster set and cluster center corresponding to the incoming execution area;

[0088] When center is end, the cluster set and cluster center are the cluster set and cluster center corresponding to the end area;

[0089] in, ;

[0090] ;

[0091] ;

[0092] Among them, U is equal to A or M or C; center is equal to start, cond or end;

[0093] For example, Table 1 shows the cluster centers corresponding to different driving styles in different input partitions.

[0094] Table 1

[0095]

[0096] Table 2 shows the motion control coefficients of different driving styles corresponding to different input partitions.

[0097] Table 2

[0098]

[0099] S32. For each incoming partition, execute step S31 to obtain the motion control coefficient for each cluster set of each partition.

[0100] Furthermore, step S4 specifically includes:

[0101] S41. Pre-build a driving style recognition model;

[0102] S42. For each incoming partition, the discrete points in each of the three cluster sets corresponding to the incoming partition are divided into training samples and test samples in a ratio of 7:3;

[0103] Each training sample or test sample is a discrete point, and the data included in the discrete point are: following distance ,speed , acceleration , location g, collection time t, driving style;

[0104] S43. The driving style recognition model is trained using the training samples and tested using the test samples until the recognition accuracy of the driving style recognition model reaches a preset value k, thereby obtaining a trained driving style recognition model for the input partition.

[0105] S44 repeats steps S42 to S43, for each incoming partition are executed, to obtain three completed training driving style recognition models for the three incoming partitions;

[0106] The three trained driving style recognition models include: a starting area driving style recognition model, an execution area driving style recognition model, and an ending area driving style recognition model.

[0107] Furthermore, step S5 specifically includes:

[0108] S51 obtains the location of the smart vehicle, and determines the entry zone of the smart vehicle according to the location;

[0109] S52 obtains the current motion state data of the manually driven vehicle that is in conflict with the intelligent vehicle in the incoming zone where the intelligent vehicle is located;

[0110] The driving vehicle that has a merging conflict with the intelligent vehicle is the manually driven vehicle closest to the intelligent vehicle;

[0111] S53. Use the driving style recognition model corresponding to the merging zone where the intelligent vehicle is located to perform driving style recognition on the current motion state data of the manually driven vehicle that has a merging conflict with the intelligent vehicle to obtain the current driving style of the manually driven vehicle.

[0112] Furthermore, step S6 specifically includes:

[0113] Establishing the objective function , by adjusting 、 、 Make the objective function Minimum, get the output control vector , as the motion control instructions for intelligent vehicles; among them, is the predicted output control vector of the intelligent vehicle at the next moment after the current moment t;

[0114] Objective function Specifically:

[0115] ;

[0116] in, 、 、 is the motion control coefficient corresponding to the cluster set corresponding to the driving style of the human-driven vehicle that has a merging conflict with the intelligent vehicle in the current merging partition; is the error weight matrix; and is the control rights matrix; is the prediction step length; for The reference trajectory state vector of the intelligent vehicle at time t; Input from the motion planning module of the intelligent vehicle; for The prediction output at each moment is the trajectory state vector of the intelligent vehicle; for The predicted output control vector of the intelligent vehicle at time , for The predicted output control vector of the intelligent vehicle at time t;

[0117] Among them, the trajectory state vector of the intelligent vehicle is expressed as Indicates that its data format is: ;

[0118] The output control vector of the intelligent vehicle is used Indicates that its data format is: ;

[0119] in, is the position coordinate of the intelligent vehicle in the X-axis direction, is the position coordinate of the intelligent vehicle in the Y-axis direction of the coordinate axis; is the speed of the smart vehicle; is the heading angle of the smart vehicle; is the acceleration of the smart vehicle; is the front wheel turning angle of the smart vehicle;

[0120] ;

[0121] ;

[0122] ;

[0123] in, is the trajectory state vector middle The corresponding coefficients; is the trajectory state vector middle The corresponding coefficients are, is the trajectory state vector middle The corresponding coefficients are, is the trajectory state vector middle The corresponding coefficients; is the output control vector middle The corresponding first coefficient is, is the output control vector middle The corresponding first coefficient; is the output control vector middle The corresponding second coefficient is, is the output control vector middle The corresponding second coefficient.

[0124] Preferably, ;

[0125] ;

[0126] .

[0127] The beneficial effects of the invention are as follows:

[0128] 1. This invention divides the merging area into three merging subareas. The driving style of the manually driven vehicle in each merging subarea is identified in real time. The intelligent vehicle then makes real-time adjustments based on the driving style of the manually driven vehicle and motion control instructions, improving the real-time performance and accuracy of the intelligent vehicle's motion control.

[0129] 2. The present invention calibrates the historical motion state data of multiple manually driven vehicles in each incoming partition as discrete points in a three-dimensional coordinate system and performs cluster analysis to obtain multiple cluster sets and multiple cluster centers. The present invention also calculates the motion control coefficient corresponding to each cluster set in each incoming partition, as well as the driving style corresponding to each cluster center, thereby improving the accuracy of vehicle driving style recognition. Furthermore, the introduction of the motion control coefficient improves the accuracy of intelligent vehicle motion control command calculation.

[0130] 3. After the motion control coefficient is calculated, the motion control coefficient is used as the objective function The convergence parameter vector of the motion control coefficient is a real-time changing control coefficient, which improves the accuracy of real-time calculation of the motion control instructions of the intelligent vehicle.

[0131] The above description is only a preferred specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by any technician familiar with this technical field within the technical scope disclosed by the present invention should be covered by the scope of protection of the present invention.

Claims

1. An intelligent vehicle ramp merging control method based on driving style recognition, characterized in that: The method specifically comprises the following steps: S1. The intelligent vehicle collects historical motion status data of manually driven vehicles in different incoming zones of the incoming area; Each historical motion status data includes: following distance ,speed , acceleration ; Each historical motion state data corresponds to a discrete point in a three-dimensional coordinate system; S2. Perform Gaussian mixture clustering on multiple discrete points in different input partitions to obtain three cluster sets and three cluster centers in each input partition; Among them, each cluster set corresponds to a driving style; Label each discrete point with a driving style; S3. According to the three cluster centers in each incoming partition, the motion control coefficients of different cluster sets within the incoming partition are calibrated to obtain motion control coefficients corresponding to different cluster sets within each incoming partition; S4. For each incoming partition, the discrete points in each cluster set of the three cluster sets corresponding to the incoming partition are divided into training samples and test samples in a ratio of 7:3; Training a driving style recognition model using training samples, and testing the driving style recognition model using test samples to obtain a trained driving style recognition model; S5. The intelligent vehicle obtains the motion state data of the manually driven vehicle at the current moment and identifies the driving style of the manually driven vehicle at the current moment through the driving style recognition model; S6. The intelligent vehicle calculates the motion control instructions for the intelligent vehicle based on the current driving style of the manually driven vehicle, the current position of the intelligent vehicle corresponding to the incoming partition, and the current driving style of the intelligent vehicle; S7. The intelligent vehicle controls the movement of the intelligent vehicle according to the motion control instruction; S8. Repeat steps S5 to S7 until the smart vehicle exits the merging area; Step S6 specifically includes: Establishing the objective function , by adjusting 、 、 Make the objective function Minimum, get the output control vector , as the motion control instructions for intelligent vehicles; among them, is the predicted output control vector of the intelligent vehicle at the next moment after the current moment t; Objective function Specifically: ； in, 、 、 is the motion control coefficient corresponding to the cluster set corresponding to the driving style of the human-driven vehicle that has a merging conflict with the intelligent vehicle in the current merging partition; is the error weight matrix; and is the control rights matrix; is the prediction step length; for The reference trajectory state vector of the intelligent vehicle at time t; Input from the motion planning module of the intelligent vehicle; for The prediction output at each moment is the trajectory state vector of the intelligent vehicle; for The predicted output control vector of the intelligent vehicle at time , for The predicted output control vector of the intelligent vehicle at time t; Among them, the trajectory state vector of the intelligent vehicle is expressed as Indicates that its data format is: ; The output control vector of the intelligent vehicle is used Indicates that its data format is: ; in, is the position coordinate of the intelligent vehicle in the X-axis direction, is the position coordinate of the intelligent vehicle in the Y-axis direction of the coordinate axis; is the speed of the smart vehicle; is the heading angle of the smart vehicle; is the acceleration of the smart vehicle; is the front wheel turning angle of the smart vehicle; ； ； ； in, is the trajectory state vector middle The corresponding coefficients; is the trajectory state vector middle The corresponding coefficients are, is the trajectory state vector middle The corresponding coefficients are, is the trajectory state vector middle The corresponding coefficients; is the output control vector middle The corresponding first coefficient is, is the output control vector middle The corresponding first coefficient; is the output control vector middle The corresponding second coefficient is, is the output control vector middle The corresponding second coefficient.

2. The intelligent vehicle ramp merging control method based on driving style recognition according to claim 1 is characterized in that: Step S1 specifically includes: S11. The intelligent vehicle collects historical motion status data of multiple manually driven vehicles in different incoming zones in the incoming zone; The different import partitions in the import area include: an import start area, an import execution area, and an import end area; Vehicles traveling in the merging area include: intelligent vehicles, manually driven vehicles; The historical motion state data of the manually driven vehicle includes: motion state data of the manually driven vehicle collected at predetermined intervals T when the manually driven vehicle sequentially passes through the merge start area, the merge execution area, and the merge end area, forming a plurality of historical motion state data; Each historical motion status data includes: following distance ,speed , acceleration , location g, collection time t; S12. According to the position g in each historical motion state data, determine the historical motion state data belongs to the import partition; S13. For each incoming partition, the historical motion state data of the incoming partition is calibrated in a three-dimensional coordinate system to form a plurality of discrete points corresponding to the incoming partition; Among them, the coordinates of each historical motion state data in the three-dimensional coordinate system are ( , ); Each historical motion state data corresponds to a discrete point in a three-dimensional coordinate system; S14. For each incoming partition, execute step S13 to obtain multiple discrete points corresponding to each incoming partition.

3. The intelligent vehicle ramp merging control method based on driving style recognition according to claim 2 is characterized in that: Step S2 specifically includes: S21. Perform Gaussian mixture clustering on the multiple discrete points corresponding to each incoming partition to obtain three cluster sets and three cluster centers in the incoming partition; Among them, each incoming partition corresponds to three cluster sets and three cluster centers; Each cluster set includes multiple discrete points and a cluster center; Each cluster center includes three component data: following distance component , velocity component , acceleration component ; S22. Setting the driving style corresponding to each cluster set; Said driving styles include: aggressive, moderate, and conservative; S23. Label each discrete point in each cluster set with a driving style; When the driving style corresponding to the cluster set is an aggressive type, marking the driving style of each discrete point in the cluster set as an aggressive type; When the driving style corresponding to the cluster set is moderate, mark the driving style of each discrete point in the cluster set as moderate; When the driving style corresponding to the cluster set is conservative, marking the driving style of each discrete point in the cluster set as conservative; S24. Repeat steps S21 to S23 until all three incoming partitions are executed, forming three cluster sets and three cluster centers for each incoming partition, as well as a driving style corresponding to each cluster set and a driving style for each discrete point in each cluster set.

4. The intelligent vehicle ramp merging control method based on driving style recognition according to claim 3 is characterized in that: Step S3 specifically includes: S31. According to the three cluster sets corresponding to each incoming partition { } and three cluster centers { }、{ }、{ }, calculate three cluster sets The corresponding motion control coefficients { }、{ }、{ }; in,{ } is the first cluster set The corresponding cluster center; These are the three component data corresponding to the cluster center; the driving style corresponding to the first cluster set is aggressive; { } is the second cluster set The corresponding cluster center; are the three component data corresponding to the cluster center; the driving style corresponding to the second cluster set is moderate; { } is the third cluster set The corresponding cluster center; These are the three component data corresponding to the cluster center; the driving style corresponding to the third cluster set is conservative; in, 、 、 The motion control coefficient corresponding to the following distance component; 、 、 motion control coefficients corresponding to velocity components; 、 、 Motion control coefficient corresponding to the acceleration component; Where center is start, cond, or end; When center is start, the cluster set and cluster center are the cluster set and cluster center corresponding to the inbound start area; When center is cond, the cluster set and cluster center are the cluster set and cluster center corresponding to the incoming execution area; When center is end, the cluster set and cluster center are the cluster set and cluster center corresponding to the end area; in, ; ； ； Among them, U is equal to A or M or C; center is equal to start, cond or end; S32. For each incoming partition, execute step S31 to obtain the motion control coefficient for each cluster set of each partition.

5. The intelligent vehicle ramp merging control method based on driving style recognition according to claim 4 is characterized in that: Step S4 specifically includes: S41. Pre-build a driving style recognition model; S42. For each incoming partition, the discrete points in each of the three cluster sets corresponding to the incoming partition are divided into training samples and test samples in a ratio of 7:3; Each training sample or test sample is a discrete point, and the data included in the discrete point are: following distance ,speed , acceleration , location g, collection time t, driving style; S43. The driving style recognition model is trained using the training samples and tested using the test samples until the recognition accuracy of the driving style recognition model reaches a preset value k, thereby obtaining a trained driving style recognition model for the input partition. S44 repeats steps S42 to S43, for each incoming partition are executed, to obtain three completed training driving style recognition models for the three incoming partitions; The three trained driving style recognition models include: a starting area driving style recognition model, an execution area driving style recognition model, and an ending area driving style recognition model.

6. The intelligent vehicle ramp merging control method based on driving style recognition according to claim 1 is characterized in that: Step S5 specifically includes: S51 obtains the location of the smart vehicle, and determines the entry zone of the smart vehicle according to the location; S52 obtains the current motion state data of the manually driven vehicle that is in conflict with the intelligent vehicle in the incoming zone where the intelligent vehicle is located; The driving vehicle that has a merging conflict with the intelligent vehicle is the manually driven vehicle closest to the intelligent vehicle; S53. Use the driving style recognition model corresponding to the merging zone where the intelligent vehicle is located to perform driving style recognition on the current motion state data of the manually driven vehicle that has a merging conflict with the intelligent vehicle to obtain the current driving style of the manually driven vehicle.

Images