Man-machine mixed driving traffic flow intelligent network connection vehicle control method in expressway interleaving area

Abstract

The invention relates to an expressway interleaving area man-machine mixed driving traffic flow intelligent network connection vehicle control method, and belongs to the technical field of intelligent traffic, and the method comprises the steps: obtaining the historical state data of a target vehicle and surrounding vehicles; the historical state data are input into a completely trained lane changing intention prediction model, the lane changing intention of the target vehicle is obtained, and the completely trained lane changing intention prediction model comprises an input layer, a CNN layer, an LSTM layer, an Attention layer and an output layer which are connected in sequence; and when the lane changing intention is lane changing, controlling the target vehicle and the cooperative vehicle on the target lane to execute a cooperative lane changing strategy. According to the method, the lane changing intention prediction model of the expressway interleaving area based on the CNN, the LSTM and the attention mechanism is constructed for the complex traffic scene of the interleaving area, the lane changing intention of the intelligent network connection vehicle in the expressway interleaving area is effectively and accurately predicted, and the vehicle is controlled to execute the corresponding lane changing strategy.

Description

Technical Field

[0001] The present application relates to the field of intelligent transportation technology, and in particular to a method for controlling intelligent connected vehicles in human-machine mixed driving traffic flow in an expressway weaving area. Background Art

[0002] In recent years, with the rapid development of intelligent transportation systems and intelligent driving technologies, autonomous vehicles have gradually penetrated urban roads. It can be said that intelligent, connected vehicles will be a key participant and game-changer in future urban road traffic. While the full popularization of intelligent, connected vehicles will not be achieved in a short period of time, for the foreseeable future, the urban road traffic system will be characterized by a long-term mix of autonomous vehicles and human-driven vehicles, creating a complex mixed traffic environment.

[0003] The weaving area of ​​an expressway, where traffic flows converge between the main road and auxiliary roads or ramps, plays a decisive role in the overall expressway system's efficiency. Because drivers have extremely limited time and space to maneuver when merging vehicles in this area, it is often considered a traffic bottleneck on urban expressways. Compared to other parts of the expressway, the high vehicle density within this area, coupled with frequent deceleration, acceleration, and lane changes, makes traffic conditions extremely complex, highly prone to accidents and thus impacting traffic efficiency. Accurately and effectively predicting lane change intentions of intelligent connected vehicles is crucial for enabling them to make informed decisions, plan, and control traffic in advance, reduce traffic conflicts, and improve traffic efficiency and safety.

[0004] In the existing technology, when selecting input features for lane change intention prediction, the position, speed and acceleration information of the target vehicle are often used as model inputs, and the absolute motion information and relative motion information of surrounding vehicles are not integrated into consideration. In particular, for complex traffic scenarios in weaving areas, it is difficult to fully learn vehicle driving behaviors and their interactions, resulting in poor prediction accuracy, which in turn affects the behavioral decisions of intelligent connected vehicles and affects the traffic efficiency and safety in weaving areas of expressways. Summary of the Invention

[0005] In view of this, it is necessary to provide an intelligent connected vehicle control method for human-machine mixed driving traffic flow in the weaving area of ​​the expressway, so as to solve the problem in the existing technology that it is difficult to fully learn the vehicle driving behavior and its interaction relationship, resulting in poor prediction accuracy and affecting the traffic efficiency and safety in the weaving area of ​​the expressway.

[0006] To address the above issues, the present application provides a method for controlling intelligent connected vehicles in mixed human-machine driving traffic flow in an expressway weaving area, comprising: Obtain historical status data of the target vehicle and its surrounding vehicles; Inputting the historical state data into a well-trained lane-changing intention prediction model to obtain the lane-changing intention of the target vehicle, wherein the well-trained lane-changing intention prediction model includes an input layer, a CNN layer, an LSTM layer, an Attention layer, and an output layer connected in sequence; When the lane-changing intention is a lane-changing, the target vehicle and the cooperative vehicles in the target lane are controlled to execute a cooperative lane-changing strategy.

[0007] In some possible implementations, the fully trained lane change intention prediction model is obtained by training through the following steps: Obtain a CitySim data set, and extract a driving trajectory of each vehicle from the CitySim data set; Extracting a lane keeping trajectory in which no lane change occurs from the driving trajectory, taking the last moment of the lane keeping trajectory as a first end moment, taking a moment a preset time before the first end moment as a first start moment, and taking the time from the first start moment to the first end moment as a lane keeping scenario; Extracting the moment when the vehicle crosses the lane line from the driving trajectory as the second end moment, a moment a preset time before the second end moment as the second start moment, and a lane change scenario from the second start moment to the second end moment; extracting state data of a target vehicle and its surrounding vehicles in the lane changing scenario and the lane keeping scenario from the CitySim dataset, and determining a sample dataset based on the state data; An initial lane-changing intention prediction model is trained based on the sample data set to obtain a fully trained lane-changing intention prediction model.

[0008] In some possible implementations, extracting state data of the target vehicle and its surrounding vehicles in the lane changing scenario and the lane keeping scenario from the CitySim dataset, and determining a sample dataset based on the state data includes: Extracting initial state data of the target vehicle and its surrounding vehicles in the lane changing scenario and the lane keeping scenario from the CitySim dataset; Normalization is performed on the initial state data to obtain final state data, and a sample data set is determined based on the final state data.

[0009] In some possible implementations, extracting state data of the target vehicle and its surrounding vehicles in the lane changing scenario and the lane keeping scenario from the CitySim dataset includes: Based on the sliding window method, the state data of the target vehicle and its surrounding vehicles in the lane changing scenario and the lane keeping scenario are extracted from the CitySim dataset.

[0010] In some possible implementations, the lane-changing scenarios include left lane-changing scenarios and right lane-changing scenarios.

[0011] In some possible implementations, the surrounding vehicles include the front vehicle and the rear vehicle in the lane where the target vehicle is located, the left front vehicle, the left rear vehicle and the left rear vehicle in the left lane of the target vehicle, and the right front vehicle, the right vehicle and the right rear vehicle in the right lane of the target vehicle.

[0012] In some possible implementations, the status data includes historical status data of the target vehicle and historical interaction data with surrounding vehicles.

[0013] In some possible implementations, the method further includes: when the lane change intention is lane keeping, controlling the target vehicle to execute a follow-up strategy, wherein the follow-up strategy adopts an adaptive cruise control follow-up model.

[0014] In some possible implementations, the coordinated lane-changing strategy includes a two-stage coordinated strategy, comprising: controlling the target vehicle and the coordinated vehicles in the target lane to adjust their speeds based on a trajectory planning strategy until the distance between the target vehicle and surrounding vehicles meets the lane-changing requirement, and then controlling the target vehicle to change lanes; The vehicle trajectory in the trajectory planning strategy is described by the following formula:

[0015]

[0016]

[0017]

[0018]

[0019] Where, represents the longitudinal displacement of the vehicle, Indicates the vehicle spacing adjustment time, represents the undetermined coefficient, The first three items of represent the initial motion state of the vehicle. The last two items of represent the target motion state of the vehicle. ; The optimization objective function of the trajectory planning strategy is:

[0020] Where, For the target vehicle, For cooperative vehicles, is the target speed of the target vehicle, is the target speed of the cooperative vehicles, 、 、 They are 、 and The weight coefficient of is the desired speed of the vehicle, is the vehicle acceleration threshold, Target vehicle The acceleration of time, For cooperative vehicles acceleration of the moment; The constraints of the objective function are:

[0021] Where, For cooperative vehicles, For the target vehicle, for The longitudinal displacement of the vehicle, for The longitudinal displacement of the vehicle in front of the vehicle in the same lane is predicted based on the uniform velocity model. for Vehicle Captain, for The length of the vehicle in front of the vehicle in the same lane, for vehicle The acceleration of time, is the acceleration threshold, is the safety margin, express time The longitudinal displacement of the vehicle, for The distance between the target vehicle and surrounding vehicles at any moment, It is the safe distance between the target vehicle and surrounding vehicles, determined by the initial and final states of the vehicle during the lane change phase.

[0022] In some possible implementations, the cooperative lane-changing strategy includes a parallel cooperative strategy, which includes: when the lane-changing space on the lane away from the target vehicle's side by the cooperative vehicle in the target lane of the target vehicle is greater than a threshold, controlling the target vehicle and the cooperative vehicle to change lanes in parallel to one side.

[0023] The beneficial effects of the present application are as follows: the intelligent connected vehicle control method for mixed human-machine driving traffic flow in the weaving area of ​​the expressway provided by the present application constructs a lane change intention prediction model based on convolutional neural network (CNN), long short-term memory network (LSTM) and attention mechanism (Attention) for the complex traffic scenes in the weaving area. The lane change intention prediction model integrates the interaction data of the target vehicle and its surrounding vehicles, effectively and accurately predicts the lane change intention of the intelligent connected vehicle in the weaving area of ​​the expressway, and controls the vehicle to execute the corresponding strategy, thereby improving the traffic efficiency and safety of mixed human-machine driving traffic flow in the weaving area of ​​the expressway. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 A flow chart of an embodiment of a method for controlling intelligent connected vehicles in mixed human-machine driving traffic flow in an expressway weaving area provided by this application; Figure 2 Schematic diagram of the overall structure of the lane change intention prediction model provided in this application; Figure 3 A schematic flow chart of an embodiment of the lane change intention prediction model training steps provided in this application; Figure 4 Schematic diagram of the lane-changing scenario provided for this application; Figure 5 Schematic diagram of the lane keeping scenario provided for this application; Figure 6 Schematic diagram of the working principle of the sliding time window method provided in this application; Figure 7 Schematic diagram of a driving scenario in which a vehicle changes lanes to the left, provided for this application; Figure 8 For this application Figure 3 FIG. 5 is a flow chart of an embodiment of step S304 in FIG. DETAILED DESCRIPTION

[0025] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the embodiments described are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without making creative efforts are within the scope of protection of this application.

[0026] It should be understood that the schematic drawings are not drawn to scale. The flowcharts used in this application illustrate operations implemented according to some embodiments of the present application. It should be understood that the operations of the flowcharts may be implemented out of sequence, and steps that do not have a logical contextual relationship may be reversed in order or implemented simultaneously. In addition, those skilled in the art, guided by the contents of this application, may add one or more other operations to the flowcharts or remove one or more operations from the flowcharts. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities may be implemented in software form, or in one or more hardware modules or integrated circuits, or in different networks and / or processor systems and / or microcontroller systems.

[0027] The descriptions of "first" and "second" in the embodiments of this application are only used to describe the implicit purpose and should not be understood as indicating or implying their relative importance or implicitly indicating the number of technical features indicated. Therefore, technical features defined as "first" and "second" may explicitly or implicitly include at least one of such features. "And / or" describes the association relationship of associated objects, indicating that three relationships can exist. For example, "A and / or B" can mean: A exists alone, A and B exist at the same time, and B exists alone.

[0028] References herein to "embodiments" mean that a particular feature, structure, or characteristic described in connection with the embodiments may be included in at least one embodiment of the present application. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor does it constitute an independent or alternative embodiment that is mutually exclusive of other embodiments. It is understood, both explicitly and implicitly, by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0029] The present application provides a method for controlling intelligent connected vehicles in human-machine mixed driving traffic flow in the interweaving area of ​​an expressway, which is explained below.

[0030] Figure 1 This is a flow chart of an embodiment of the intelligent connected vehicle control method for mixed human-machine driving traffic flow in the expressway weaving area provided by this application, as shown in FIG. Figure 1 As shown in the figure, the intelligent connected vehicle control method for mixed human-machine driving traffic flow in the weaving area of ​​the expressway includes: S101, obtaining historical status data of the target vehicle and its surrounding vehicles; S102, inputting the historical state data into the fully trained lane-changing intention prediction model to obtain the lane-changing intention of the target vehicle, wherein the fully trained lane-changing intention prediction model includes an input layer, a CNN layer, an LSTM layer, an Attention layer, and an output layer connected in sequence; It should be noted that the overall structure of the lane-changing intention prediction model is as follows: Figure 2 As shown in the figure, when processing sample data, the prediction model first uses a one-dimensional convolutional neural network (1D CNN) layer for feature extraction. In 1D CNN, the convolution operation is performed unidirectionally along the time axis. In this way, the model can capture the local temporal features in the sequence data; then, the pooling layer performs dimensionality reduction on the output of the convolution layer to reduce the number of parameters and extract more critical features; the advantage of the CNN layer is that it can identify local patterns in the data and maintain the time series characteristics of the features. Compared with LSTM, CNN is more efficient in extracting local features; in addition, by using convolution kernels of different sizes, CNN can extract features of different scales from the data, which provides the model with richer information and helps to better understand and predict complex patterns in sequence data.

[0031] The extracted features are then input into the LSTM layer for sequence modeling. The LSTM network can learn important information from features at different positions and can process long sequences of data. In this application, the role of the LSTM layer is to extract the temporal information of the vehicle lane-changing features and learn the correlation features between sequences. The last hidden state of the LSTM layer is then passed to the attention layer for further analysis.

[0032] Next, the Attention layer enhances the model's ability to identify the importance of features by assigning different weights to different features. This mechanism allows the model to focus on features that are more critical to the classification task, thereby improving the model's prediction accuracy and robustness.

[0033] Finally, the features processed by the attention layer pass through a fully connected layer and are activated by the Softmax function to output the final classification probability. The Softmax layer converts the output of the model into a probability distribution, which can calculate the probability of the vehicle turning left, turning right, or staying in the current lane.

[0034] S103, when the lane change intention is to change lanes, controlling the target vehicle and the cooperative vehicles in the target lane to execute a coordinated lane change strategy; S104: When the lane change intention is lane keeping, control the target vehicle to execute a follow-and-relax strategy.

[0035] It should be noted that the following strategy of intelligent connected vehicles adopts the adaptive cruise control following model.

[0036] Compared with the existing technology, this application constructs a lane-changing intention prediction model based on convolutional neural networks (CNN), long short-term memory networks (LSTM) and attention mechanisms (Attention) for complex traffic scenarios in weaving areas. By integrating the interaction data of the target vehicle and its surrounding vehicles through the lane-changing intention prediction model, the lane-changing intention of intelligent connected vehicles in the weaving areas of expressways is effectively and accurately predicted, and the vehicles are controlled to execute corresponding strategies, thereby improving the traffic efficiency and safety of mixed human and machine driving in the weaving areas of expressways.

[0037] In order to obtain a better lane change intention prediction model, in some embodiments, such as Figure 3 As shown, the lane change intention prediction model in step S102 is obtained through the following steps: S301, obtaining a CitySim dataset, and extracting the driving trajectory of each vehicle from the CitySim dataset; It should be noted that lane-changing intention prediction is a classification problem, which aims to The target vehicle and surrounding vehicles’ status data are used to estimate the vehicle’s status during the prediction window. The probability of a lane change operation occurring, in this embodiment, , The value of determines the maximum prediction range of the prediction model. It is set to 5 seconds, which enables the evaluation of prediction models for longer forecast horizons while ensuring that a sufficient number of data samples are extracted from the selected dataset.

[0038] It should also be noted that the research object is Expressway A, an expressway intersection in a domestic city in the CitySim dataset. While the data volume is large, it can more realistically reflect the complex traffic conditions in the current domestic expressway intersections.

[0039] S302: Extracting a lane keeping trajectory in which the lane does not change from the driving trajectory, defining the last moment of the lane keeping trajectory as a first end moment, defining a moment a preset time before the first end moment as a first start moment, and defining a lane keeping scenario from the first start moment to the first end moment. S303: Extracting the moment when the vehicle crosses the lane line from the driving trajectory as the second end moment, a moment a preset time before the second end moment as the second start moment, and a lane change scenario from the second start moment to the second end moment; It should be noted that the lane changing scenario and lane keeping scenario are a set of time series from a single vehicle, and the time value ranges from 0 to , and then from arrive ,The time series includes 2s of historical trajectory data and 5s of future trajectory data; ,Lane-changing scenarios such as Figure 4 As shown in the figure, all samples in the lane changing scenario are marked as RLC (right lane change) or LLC (left lane change) according to the direction of the operation at the end of the lane changing scenario. Whether it is a left lane change or a right lane change depends on the direction of the vehicle and the change of the lane corresponding to the vehicle; the lane keeping scenario is as shown in the figure. Figure 5 As shown, the lane keeping scenario is defined as the lane change time TTLC of a single vehicle> A set of data samples.

[0040] S304. Extracting state data of the target vehicle and its surrounding vehicles in the lane-changing scenario and the lane-keeping scenario from the CitySim dataset based on a sliding window method, and determining a sample dataset based on the state data; It should be noted that the sliding window method is a commonly used method for processing time series data. Its working principle is as follows: Figure 6 As shown, Figure 6 middle refers to the length of the sliding time window, Indicates the sampling time interval, the sliding time window slides along the time axis each time Length, and intercept the fixed time length as In this embodiment, a 2-s sliding window is used to sample the vehicle trajectory with a sampling interval of 0.2s, and multiple lane-changing trajectory sequence sample sets are obtained:

[0041]

[0042] Where, represents the trajectory sequence sample set, Indicates the Trajectory sequence samples, Indicates the total number of sequence samples extracted, Indicates the features corresponding to the sequence, Indicates the labels corresponding to the predicted sequence samples, including left lane change, right lane change and lane keeping, , 0 means lane keeping, 1 means right lane change, and 2 means left lane change.

[0043] S305 : Training the initial lane-changing intention prediction model based on the sample data set to obtain a fully trained lane-changing intention prediction model.

[0044] Considering that lane changing is a strategy adopted by drivers to obtain a better driving experience during driving, this behavior involves changing from the current lane to another lane, usually accompanied by a certain lane changing speed to ensure driving continuity and safety. The reasons for drivers' lane changing intentions are diverse, mainly including the driving status of their own vehicle, the driving behavior of surrounding vehicles, the traffic environment, etc.

[0045] As Figure 7 Taking the driving scenario of a vehicle changing lanes to the left as an example, the target vehicle TV is traveling at a higher speed, while the leading vehicle PV is traveling relatively slowly. In order to seek a faster driving speed and a better driving experience, the target vehicle TV will have the intention to change lanes. In the left lane, the left front vehicle LPV and the left rear vehicle LFV are far apart, and the gap between them meets the safe lane-changing distance, which meets the lane-changing conditions. The vehicle may have the lateral driving intention to change lanes to the left. At the same time, the vehicle TV is close to the leading vehicle PV. In order to avoid a collision with the leading vehicle, the vehicle may have the longitudinal driving intention to slow down and change lanes. If the left rear vehicle LFV has a large acceleration, the lane-changing gap provided by the left lane will gradually decrease, and the safe lane-changing conditions will not be met. The target vehicle TV will keep going straight after comprehensive considerations.

[0046] It can be seen that when deciding whether to change lanes, the driver will consider multiple factors, among which position distance, speed, and surrounding traffic information are important considerations. This information helps the driver evaluate the relative position between the current lane and the target lane, as well as the dynamic changes that may be encountered during the lane change process. Therefore, in some embodiments, the state data includes historical state information of the target vehicle and historical interaction information with surrounding vehicles, wherein the historical state data of the target vehicle includes the vehicle's speed, position, and lane-related data, and the historical interaction information between the target vehicle and surrounding vehicles includes the relative speed and relative distance between the target vehicle and surrounding vehicles. The surrounding vehicles include the front and rear vehicles in the lane where the target vehicle is located, the left front vehicle, the left rear vehicle, and the left rear vehicle in the lane to the left of the target vehicle, and the right front vehicle, the right vehicle, and the right rear vehicle in the lane to the right of the target vehicle.

[0047] Considering that the state data includes speed, acceleration, and distance, the dimensions and units of these features are different. Since the dimension difference may affect the data analysis results, in some embodiments, the data is normalized to eliminate this effect. After normalization, all features will be adjusted to the same order of magnitude, which not only helps the model to more accurately capture the relative changes between features, but also helps to improve the convergence speed of model training. Specifically, Figure 8 As shown, step S304 includes: S801. Extracting initial state data of a target vehicle and its surrounding vehicles in a lane-changing scenario and a lane-keeping scenario from a CitySim dataset; S802: Normalize the initial state data to obtain final state data, and determine a sample data set based on the final state data.

[0048] It should be noted that in this embodiment, a linear normalization method is used to limit the range of eigenvalues ​​to [0, 1]. To ensure the stability and generalization ability of model training, only the statistical data of the training set are used to calculate the normalization parameters. Subsequently, these parameters are applied to the data of the validation set and the test set to ensure the consistency of the entire data preprocessing process. The specific calculation formula is as follows:

[0049]

[0050]

[0051] Where, Indicates the training set input data The minimum value of the features, Indicates the training set input data The maximum value of the features, Indicates the validation set input data Features, Indicates the test set input data Features.

[0052] In this example, the structure and input and output of each layer of the prediction model are introduced respectively: Input Layer: In the input layer, the historical state sequence of the target vehicle and the historical state sequences of surrounding vehicles are fused into the input feature sequence. The input is a three-dimensional vector, with the three dimensions representing the batch size (batch_size), the input sequence length (in_seq_len), and the number of features (in_channels). In this article, batch_size = 64, the frame rate (FPS) = 5, and in_seq_len = FPS * 5 * 2 = 10, and in_channels = 18. Therefore, the input vector shape of the input layer is (64, 10, 18).

[0053] CNN layer: In this layer, the input vector from the input layer first enters a one-dimensional convolutional layer. The convolutional layer has 36 output channels (out_channels) and a kernel size (kernel_size) of 3. Therefore, the output sequence length (out_seq_len) of the convolution operation is (in_seq_len - kernel_size + 1) = 10 - 3 + 1 = 8. Therefore, the shape of the vector after the one-dimensional convolutional layer is (64, 8, 36). The vector then enters a max pooling layer with a kernel size (kernel_size) of 3. Therefore, the output sequence length (out_seq_len) of the pooling operation is (8 - 3 + 1) = 6. Therefore, the final output vector shape of the CNN layer is (64, 6, 36).

[0054] LSTM layer: This layer takes as input the output vector from the CNN layer, with the input feature dimension input_size=36 and the hidden state dimension hidden_size=512. Information is passed and updated between time steps through a gating mechanism such as the input gate, forget gate, and output gate to generate the hidden state output for each time step. The final output vector shape of the LSTM layer is (64, 6, 512). Attention layer: This layer takes as input the output vector from the LSTM layer, with batch_size = 64, seq_len = 6, and hidden_size = 512. This input is first linearly transformed through a fully connected layer to generate a query vector (query), a key vector (key), and a value vector (value), all of shape (batch_size, seq_len, hidden_size). The query, key, and value vectors are then divided into multiple heads (num_heads), with the number of heads set to 32. For each head, the dimension head_dim = hidden_size / num_heads = 512 / 32 = 16. The dot product between the query vector (query) and the key vector (key) is then calculated to generate attention scores. These scores are then normalized using the softmax function to obtain attention weights (attention weights), whose shape is (batch_size, num_heads, seq_len, seq_len), or (64, 32, 6, 6). The attention weights are then applied to the value vector, resulting in a weighted output with a shape of (batch_size, num_heads, seq_len, head_dim), or (64, 32, 6, 16). Finally, the outputs from the multiple heads are concatenated to restore the original shape, resulting in an output shape of (batch_size, seq_len, hidden_size), or (64, 6, 512). When processing time series, the Attention layer weights the relationship between each position in the input sequence and every other position, enabling the model to better capture global dependencies.

[0055] Output layer: In sequence models, we are typically interested in the final state of the sequence, so we extract the output of the last time step, resulting in an output vector with a shape of (batch_size, hidden_size), or (64, 512). This vector is then fed into a fully connected layer and softmax normalized, resulting in the output of the entire model with a shape of (64, 3), where 64 represents the batch size and 3 represents the number of classification categories (left lane change, right lane change, lane keeping).

[0056] To determine a better lane-changing strategy and improve the efficiency and safety of mixed-vehicle traffic in weaving areas on expressways, in some embodiments, a coordinated lane-changing strategy includes a two-stage coordinated strategy. The two-stage coordinated strategy includes: controlling the speed of a target vehicle and cooperative vehicles in the target lane until the distance between the target vehicle and surrounding vehicles meets the lane-changing requirements, and then controlling the target vehicle to change lanes; It should be noted that the cooperative vehicle can be the rear vehicle or the front vehicle in the target lane, and can be an intelligent connected vehicle or a connected vehicle of a natural person who is willing to cooperate.

[0057] Among them, the vehicle trajectory in the trajectory planning strategy adopts the time The quartic polynomial description of :

[0058] Where, represents the longitudinal displacement of the vehicle, Indicates the vehicle spacing adjustment time, Denotes the unknown coefficient, let , using matrix form to express the initial and final state boundary constraints ,in:

[0059]

[0060] Where, The first three items of represent the initial motion state of the vehicle. The last two items represent the target motion state of the vehicle. To improve lane change smoothness, the target acceleration ; Select 、 As decision variables, construct an optimization problem, then the optimal coefficient satisfies: ; The design optimization objective function is:

[0061] Where, For the target vehicle, For cooperative vehicles, is the target speed of the target vehicle, is the target speed of the cooperative vehicles, 、 、 They are 、 and The weight coefficient of is the desired speed of the vehicle, is the vehicle acceleration threshold, Target vehicle The acceleration of time, For cooperative vehicles The acceleration at the moment; in the optimization function, When the item enters the lane change phase 、 The speed is close to the expected speed to reduce the impact of cooperative behavior on upstream traffic, The item makes the spacing adjustment time as short as possible. The term is a penalty term, which aims to prevent the optimal solution from falling on the boundary of the feasible region, thereby reducing the impact of human driving vehicle disturbance on the feasibility of the receding horizon planning; The constraints of the objective function are:

[0062] Where, For cooperative vehicles, For the target vehicle, for The longitudinal displacement of the vehicle, for The longitudinal displacement of the vehicle in front of the vehicle in the same lane is predicted based on the uniform velocity model. for Vehicle Captain, for The length of the vehicle in front of the vehicle in the same lane, for vehicle The acceleration of time, is the acceleration threshold, is the safety margin, express time The constraints of the first three rows of formulas for the longitudinal displacement of the vehicle are: during the spacing adjustment process, C1 and C2 do not collide with the vehicle in front of the lane and meet the comfort requirements; for The distance between the target vehicle and surrounding vehicles at any moment, is the safe distance between the target vehicle and surrounding vehicles, which is determined by the initial and final states of the vehicle during the lane change phase. The constraint in the last row of the formula means that at the end of the spacing adjustment phase, the distance between the target vehicle and surrounding vehicles must meet the minimum safe distance requirement for lane change to ensure that the trajectory planning during the lane change phase is solvable.

[0063] Furthermore, the lane changing strategy also includes a parallel cooperative strategy, which includes: when the lane changing space on the lane away from the target vehicle side of the cooperative vehicle in the target lane of the target vehicle is greater than a threshold, controlling the target vehicle and the cooperative vehicle to change lanes in parallel to one side.

[0064] It should be noted that if the lane changing space is greater than the threshold, it means that there is sufficient lane changing space.

[0065] It should also be noted that the parallel collaborative strategy has a higher priority than the two-stage collaborative strategy. That is, it is first determined whether the lane-changing space of the cooperative vehicle in the target lane of the target vehicle away from the lane on the side of the target vehicle is greater than the threshold. If it is met, the parallel collaborative strategy is executed; if it is not met, the two-stage collaborative strategy is executed.

[0066] The above is a detailed introduction to the intelligent connected vehicle control method for human-machine mixed driving traffic flow in the interweaving area of ​​an expressway provided by this application. This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method of this application and its core idea; at the same time, for technical personnel in this field, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as a limitation on this application.

[0067] The above is only a preferred specific implementation method of the present application, but the scope of protection of the present application is not limited thereto. Any changes or replacements that can be easily thought of by any technician familiar with this technical field within the technical scope disclosed in this application should be covered by the scope of protection of the present application.

Claims

1. A method for controlling intelligent connected vehicles in human-machine mixed traffic flow in an expressway weaving area, characterized in that: include: Obtain historical status data of the target vehicle and its surrounding vehicles; Inputting the historical state data into a well-trained lane-changing intention prediction model to obtain the lane-changing intention of the target vehicle, wherein the well-trained lane-changing intention prediction model includes an input layer, a CNN layer, an LSTM layer, an Attention layer, and an output layer connected in sequence; When the lane-changing intention is a lane-changing, the target vehicle and the cooperative vehicle on the target lane are controlled to execute a cooperative lane-changing strategy.

2. The intelligent connected vehicle control method for mixed human-machine traffic flow in the expressway weaving area according to claim 1 is characterized in that: The fully trained lane-changing intention prediction model is obtained through the following training steps: Obtain a CitySim data set, and extract a driving trajectory of each vehicle from the CitySim data set; Extracting a lane keeping trajectory in which the lane does not change from the driving trajectory, taking the last moment of the lane keeping trajectory as a first end moment, taking a moment a preset time before the first end moment as a first start moment, and taking a period from the first start moment to the first end moment as a lane keeping scene; Extracting the moment when the vehicle crosses the lane line from the driving trajectory as the second end moment, taking the moment of a preset time before the second end moment as the second start moment, and taking the time from the second start moment to the second end moment as a lane change scenario; extracting state data of the vehicle and its surrounding vehicles in the lane changing scenario and the lane keeping scenario from the CitySim data set, and determining a sample data set based on the state data; The initial lane-changing intention prediction model is trained based on the sample data set to obtain a fully trained lane-changing intention prediction model.

3. The intelligent connected vehicle control method for mixed human-machine traffic flow in the expressway weaving area according to claim 2 is characterized in that: Extracting state data of the target vehicle and its surrounding vehicles in the lane changing scenario and the lane keeping scenario from the CitySim data set, and determining a sample data set based on the state data, including: Extracting initial state data of the target vehicle and its surrounding vehicles in the lane changing scenario and the lane keeping scenario from the CitySim data set; The initial state data is normalized to obtain final state data, and a sample data set is determined based on the final state data.

4. The intelligent connected vehicle control method for mixed human-machine traffic flow in expressway weaving areas according to claim 2 is characterized in that: Extracting state data of the target vehicle and its surrounding vehicles in the lane changing scenario and the lane keeping scenario from the CitySim data set includes: Based on the sliding window method, the state data of the target vehicle and its surrounding vehicles in the lane changing scenario and the lane keeping scenario are extracted from the CitySim data set.

5. The intelligent connected vehicle control method for mixed human-machine traffic flow in expressway weaving areas according to claim 1 is characterized in that: The lane changing scenarios include left lane changing scenarios and right lane changing scenarios.

6. The intelligent connected vehicle control method for mixed human-machine traffic flow in expressway weaving areas according to claim 1 is characterized in that: The surrounding vehicles include the front vehicle and the rear vehicle in the lane where the target vehicle is located, the left front vehicle, the seat vehicle and the left rear vehicle in the left lane of the target vehicle, and the right front vehicle, the right vehicle and the right rear vehicle in the right lane of the target vehicle.

7. The intelligent connected vehicle control method for mixed human-machine driving traffic flow in expressway weaving area according to claim 1 is characterized in that: The state data includes historical state data of the target vehicle and historical interaction data with surrounding vehicles.

8. The intelligent connected vehicle control method for mixed human-machine traffic flow in expressway weaving areas according to claim 1 is characterized in that: The method further includes: when the lane change intention is lane keeping, controlling the target vehicle to execute a follow-and-relax strategy, wherein the follow-and-relax strategy adopts an adaptive cruise control follow-and-relax model.

9. The intelligent connected vehicle control method for mixed human-machine traffic flow in expressway weaving areas according to claim 1 is characterized in that: The cooperative lane-changing strategy includes a two-stage cooperative strategy, which includes: based on a trajectory planning strategy, controlling the target vehicle and the cooperative vehicles in the target lane to adjust the speed until the vehicle distance between the target vehicle and the surrounding vehicles meets the lane-changing requirements, and controlling the target vehicle to change lanes; The vehicle trajectory in the trajectory planning strategy is described by the following formula: In the formula, represents the longitudinal displacement of the vehicle, Indicates the vehicle spacing adjustment time, represents the undetermined coefficient, The first three items of represent the initial motion state of the vehicle. The last two items of represent the target motion state of the vehicle. ; The optimization objective function of the trajectory planning strategy is: In the formula, For the target vehicle, For collaborative vehicles, is the target speed of the target vehicle, is the target speed of the cooperative vehicles, , , They are , and The weight coefficient of is the expected speed of the vehicle, is the vehicle acceleration threshold, The target vehicle The acceleration of time, For collaborative vehicles The acceleration of the moment; The constraints of the objective function are: In the formula, For collaborative vehicles, For the target vehicle, for The longitudinal displacement of the vehicle, for The longitudinal displacement of the vehicle in front of the vehicle in the same lane is predicted based on the uniform velocity model. for Vehicle Captain, for The length of the vehicle in front of the vehicle in the lane, for vehicle The acceleration of time, is the acceleration threshold, is the safety margin, express time The longitudinal displacement of the vehicle, for The distance between the target vehicle and surrounding vehicles at all times, It is the safe distance between the target vehicle and surrounding vehicles, which is determined by the initial and final states of the vehicle during the lane change phase.

10. The intelligent connected vehicle control method for mixed human-machine traffic flow in expressway weaving area according to claim 9, characterized in that: The cooperative lane-changing strategy includes a parallel cooperative strategy, which includes: when the lane-changing space on the lane of the cooperative vehicle on the target lane of the target vehicle away from the target vehicle is greater than a threshold, controlling the target vehicle and the cooperative vehicle to change lanes in parallel to one side.

Images