Man-machine co-driving control right distribution method and device considering driving group characteristics

Abstract

The invention relates to the technical field of data processing, and discloses a man-machine co-driving control right distribution method and device considering driving group characteristics. The method comprises the following steps: acquiring vehicle operation data and extracting behavior characteristics, and constructing a driving group recognition model based on a deep neural network and dual-path characteristics; classifying driver behaviors to generate group labels, and obtaining control parameters according to the labels; and establishing a dual-index control right distribution model through the transverse deviation and the steering torque, and performing dynamic weight combination in combination with group identification frequency to obtain a distribution result. According to the invention, the efficiency and accuracy of man-machine co-driving control right distribution considering the characteristics of the driving group are improved.

Description

Technical Field

[0001] The present application relates to the field of data processing, and in particular to a method and device for allocating human-machine co-driving control rights taking into account the characteristics of a driving group. Background Art

[0002] Existing methods for allocating control rights in human-machine co-driving primarily rely on pre-set fixed allocation rules or simple weight calculations based on vehicle status and driver operating intentions. These methods typically employ a single control strategy, such as a linear allocation method based on vehicle lateral deviation or a proportional allocation method based on driver steering torque. Furthermore, some methods incorporate intelligent algorithms such as fuzzy control or neural networks in an attempt to increase the flexibility of control rights allocation. However, most of these methods fail to account for the diversity of driving groups and overlook the significant differences between drivers in terms of operating habits, reaction characteristics, and driving skills.

[0003] Existing technologies have the following shortcomings: First, the control rights allocation strategy is too rigid and cannot adapt to the operating preferences of different driving groups, resulting in frequent control disputes during human-machine collaboration; second, there is a lack of in-depth analysis and real-time recognition mechanism of driver behavior characteristics, which cannot accurately determine the type of driving group the current driver belongs to; third, the dynamic changes in driving group characteristics over time are not fully considered in the control rights allocation process, making it difficult to achieve a smooth transition of control rights; finally, when dealing with complex driving scenarios, existing methods often have the problem of unreasonable control rights allocation, affecting driving safety and comfort. Summary of the Invention

[0004] The present application provides a method and device for allocating human-machine co-driving control rights considering the characteristics of a driving group, which is used to improve the efficiency and accuracy of allocating human-machine co-driving control rights considering the characteristics of a driving group.

[0005] In a first aspect, the present application provides a method for allocating control rights for human-machine co-driving considering the characteristics of a driving group, the method comprising: Collecting vehicle operation data through on-board sensors, extracting vehicle-to-vehicle distance data, lateral displacement data, vehicle acceleration data, and driving operation frequency data from the vehicle operation data to obtain driving behavior characteristic data; Using the driving behavior feature data to establish a deep neural network model, and constructing a driving group recognition model based on a dual-path architecture of manually extracted features and CNN convolutional features; Classifying and distinguishing the real-time behaviors of drivers based on the driving group recognition model to generate general driving group labels, aggressive driving group labels, and novice driving group labels; Obtaining corresponding control parameters according to the driving group label, and calculating the vehicle lateral comprehensive deviation based on a normalized weighted combination of the preview point lateral deviation and the yaw angle deviation; By using the vehicle's lateral comprehensive deviation and the driver's steering torque, a control right allocation mathematical model is established through a double exponential function to obtain an intelligent system control weight coefficient and a driver control weight coefficient; According to the driving group label and the group identification frequency data within the time window, the intelligent system control weight coefficient and the driver control weight coefficient are dynamically weighted and combined to generate a control right allocation result.

[0006] In a second aspect, the present application provides a device for allocating control rights for human-machine co-driving that takes into account the characteristics of a driving group. The device for allocating control rights for human-machine co-driving that takes into account the characteristics of a driving group includes: an extraction module, configured to collect vehicle operation data through on-board sensors, extract vehicle-to-vehicle distance data, lateral displacement data, vehicle acceleration data, and driving operation frequency data from the vehicle operation data, and obtain driving behavior characteristic data; A construction module is used to establish a deep neural network model using the driving behavior feature data, and to construct a driving group recognition model based on a dual-path architecture of manually extracted features and CNN convolutional features; A discrimination module, configured to classify and discriminate the real-time behaviors of drivers based on the driving group recognition model, and generate general driving group labels, aggressive driving group labels, and novice driving group labels; a combination module, configured to obtain corresponding control parameters according to the driving group label, and calculate a vehicle lateral comprehensive deviation based on a normalized weighted combination of the preview point lateral deviation and the yaw angle deviation; a distribution module, configured to utilize the vehicle's lateral comprehensive deviation and the driver's steering torque to establish a control right distribution mathematical model through a double exponential function, thereby obtaining an intelligent system control weight coefficient and a driver control weight coefficient; A weighting module is used to dynamically weight the intelligent system control weight coefficient and the driver control weight coefficient according to the driving group label and the group identification frequency data within the time window to generate a control right allocation result.

[0007] In the technical solution provided by the present application, the present invention collects vehicle operation data through on-board sensors and extracts vehicle-to-vehicle distance data, lateral displacement data, vehicle acceleration data and driving operation frequency data, providing rich data support for the comprehensive analysis of driving behavior characteristics, ensuring the accurate capture of driver operation characteristics; using driving behavior feature data to establish a deep neural network model, adopting a dual-path architecture of manual feature extraction and CNN convolution features, improving the accuracy and robustness of feature extraction, and enhancing the model's recognition ability for different driving behavior patterns; using a driving group recognition model to classify and judge the driver's real-time behavior, generate general driving group labels, aggressive driving group labels and novice driving group labels, and realize the recognition of driver groups. Accurate identification and dynamic tracking of body characteristics; obtaining corresponding control parameters based on driving group labels, combining the normalized weighted combination of preview point lateral deviation and yaw angle deviation, realizing accurate calculation of vehicle lateral comprehensive deviation, and providing a reliable basis for control right allocation; using vehicle lateral comprehensive deviation and driver steering torque to construct a double exponential function control right allocation mathematical model, obtaining intelligent system control weight coefficient and driver control weight coefficient, ensuring smoothness and continuity of control right allocation; dynamically weighting and combining control weight coefficients according to driving group labels and group recognition frequency data within the time window, so that the control right allocation result can adapt to the dynamic changes of driver behavior characteristics, and improve the adaptability and coordination of the human-machine co-driving system. Through the organic combination of the above technical features, the present invention not only improves the accuracy of driving group identification, but also realizes adaptive allocation of control rights, effectively reduces control right conflicts in the process of human-machine co-driving, while ensuring driving safety and comfort, and providing personalized human-machine co-driving experience for different driving groups. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0009] Figure 1 This is a schematic diagram of an embodiment of a method for allocating control rights for human-machine co-driving taking into account the characteristics of a driving group in an embodiment of the present application; Figure 2 This is a schematic diagram of an embodiment of a device for allocating human-machine co-driving control rights taking into account the characteristics of the driving group in an embodiment of the present application. DETAILED DESCRIPTION

[0010] The embodiments of the present application provide a method and apparatus for allocating control rights for human-machine co-driving that takes into account the characteristics of the driving group. The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of this application and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units that are not explicitly listed or are inherent to these processes, methods, products or devices.

[0011] For ease of understanding, the specific process of the embodiment of the present application is described below. Figure 1 In one embodiment of the present application, a method for allocating control rights for human-machine co-driving considering the characteristics of the driving group includes: Step S101: collecting vehicle operation data through on-board sensors, extracting vehicle-to-vehicle distance data, lateral displacement data, vehicle acceleration data, and driving operation frequency data from the vehicle operation data to obtain driving behavior characteristic data; Step S102: Using the driving behavior feature data to establish a deep neural network model, and constructing a driving group recognition model based on a dual-path architecture of manually extracted features and CNN convolutional features; Step S103: classify and identify the real-time behaviors of the drivers based on the driving group recognition model to generate general driving group labels, aggressive driving group labels, and novice driving group labels; Step S104: Obtain corresponding control parameters based on the driver group label, and calculate the vehicle's comprehensive lateral deviation based on a normalized weighted combination of the preview point lateral deviation and the yaw angle deviation; Step S105: Using the vehicle's lateral comprehensive deviation and the driver's steering torque, a control right allocation mathematical model is established through a double exponential function to obtain an intelligent system control weight coefficient and a driver control weight coefficient; Step S106: Dynamically weight the intelligent system control weight coefficient and the driver control weight coefficient based on the driving group label and the group identification frequency data within the time window to generate a control right allocation result.

[0012] It is understandable that the execution subject of this application can be a human-machine co-driving control rights allocation device that takes into account the characteristics of the driving group, or it can be a terminal or a server, and the specific details are not limited here. The embodiment of this application is described by taking the server as the execution subject as an example.

[0013] Specifically, the vehicle's onboard sensor system collects raw data during operation. The system includes millimeter-wave radar, visual sensors, and a CAN bus. The millimeter-wave radar primarily collects distance information from the preceding vehicle, with a detection range of 0-200 meters and a sampling frequency of 20Hz. The visual sensor collects lane information for lateral displacement calculation, with a field of view of 120 degrees and a resolution of 1920×1080 pixels. The CAN bus collects vehicle status information, such as acceleration and speed, at a sampling frequency of 100Hz. The raw data is processed using a multi-sensor data fusion algorithm. This algorithm employs a Kalman filter to align and fuse the data collected by different sensors in both temporal and spatial dimensions, generating a unified data format. This data fusion yields vehicle-to-vehicle distance data with a resolution of 0.1 meters, lateral displacement data with an accuracy of centimeters, and vehicle acceleration data with an accuracy of 0.01 m / s². Based on this basic data, the frequency of driving maneuvers, including sudden braking, sudden acceleration, and sharp turns, is calculated over a 30-second sliding window. Sudden braking is defined as a deceleration greater than 4 m / s², sudden acceleration as an acceleration greater than 3 m / s², and sharp turns as a lateral acceleration greater than 2 m / s². After acquiring the driving behavior characteristic data, a deep neural network model is developed to construct a driver group recognition model. This model utilizes a dual-path architecture: a manual feature extraction path and a CNN convolutional feature path. In the manual feature extraction path, statistical features are extracted from the driving behavior characteristic data, including mean, standard deviation, kurtosis, and skewness, to form a 128-dimensional feature vector. The CNN convolutional path utilizes a three-layer convolutional architecture with kernel sizes of 3×3, 5×5, and 7×7, respectively. Each layer is followed by a max pooling layer, and finally a fully connected layer to produce a 256-dimensional feature vector. The two feature vectors are concatenated, passed through two fully connected layers, and finally fed into a softmax classification layer to form the driver group recognition model.

[0014] Based on the trained driving group recognition model, real-time driver behavior is classified and discriminated. The model outputs three category probabilities: general drivers, aggressive drivers, and novice drivers. A probability threshold of 0.6 is used for discrimination; when the probability of a category exceeds the threshold, it is assigned to that category. To avoid frequent switching, a 5-second window is used for smoothing, and the category with the highest frequency within the window is used as the final category label. After obtaining the driving group labels, control parameters are set based on the characteristics of each group. Taking a preview distance of 30 meters as an example, the lateral deviation at the preview point is calculated, and the vehicle's yaw angle deviation is also obtained. These two deviations are normalized and mapped to the range [0, 1]. Different weighting coefficients are used for different driving groups: the lateral deviation weight for the general driver group is 0.6, and the yaw angle weight for the aggressive driver group is 0.7 and 0.3, respectively; and the novice driver group is 0.5 and 0.5, respectively. The vehicle's comprehensive lateral deviation is obtained through weighted combination.

[0015] Based on the vehicle's lateral comprehensive deviation and the driver's steering torque, a mathematical model for control rights allocation is constructed. The model uses a double exponential function structure to establish exponential mappings for the comprehensive deviation and steering torque respectively. Different driving groups correspond to different exponential function parameters: the general driving group parameters are =4, =3, =0.5; the aggressive driving group parameter is =4, =10, =0.5; the parameter of the novice driving group is =20, =3, β3=0.5. These parameters are used to calculate the control weight coefficients for the intelligent system and the driver. Finally, the frequency data for group identification is statistically analyzed using a 10-second time window. Based on this frequency data, dynamic weight coefficients are calculated, and a weighted combination of the control weight coefficients for the intelligent system and the driver is performed. For example, within a 10-second window, if a driver's behavior is identified as belonging to the general driving group 6 times and the aggressive driving group 4 times, the general driving group will receive a weight of 0.6, and the aggressive driving group will receive a weight of 0.4. These weights are multiplied by their respective control weight coefficients and summed to obtain the final control rights allocation.

[0016] For example, when a driver is driving on a highway at a speed of 100 km / h, the collected data shows that the distance to the vehicle ahead fluctuates around 50 meters, the lateral deviation averages 0.2 meters, the acceleration varies steadily, and one sharp turn is recorded within 30 seconds. The driver group identification model outputs the following probability distributions: 0.75 for the average driver group, 0.15 for the aggressive driver group, and 0.1 for the novice driver group, identifying the driver as the average driver. At this point, the lateral deviation at the preview point is 0.3 meters, and the yaw angle deviation is 2 degrees. After normalization, these values ​​are 0.3 and 0.2, respectively, resulting in a weighted combined deviation of 0.26. The driver's steering torque is 2 N·m. The control weighting model calculates the intelligent system's control weight as 0.3 and the driver's control weight as 0.7. The average driver group identification result is maintained throughout the subsequent 10-second window, and the final control weighting remains at this level.

[0017] It should be noted that, in the embodiment of the present application, a control rights allocation model for human-machine co-driving is designed taking into account the driver's steering behavior and the vehicle's lateral deviation.

[0018]

[0019] Where, 、 Respectively represent the real-time control rights of the intelligent system and the driver, 、 、 is a parameter related to the characteristics of the driving group, represents the normalized steering torque, Represents the normalized vehicle lateral comprehensive deviation. 、 They are the normalized values ​​of the lateral deviation and yaw angle deviation at the preview point position, and the design parameters and satisfy .

[0020] Design corresponding driving group characteristic parameters for different driving groups.

[0021] (1) General driving group, characteristic parameters meet

[0022] (2) Aggressive driving group, characteristic parameters meet

[0023] (3) Novice driver group, characteristic parameters meet

[0024] For the control rights allocation model of the general driving group, as the comprehensive deviation As the size of the steering wheel increases, the control power of the intelligent system will gradually increase, and as the driver gradually participates in the steering task ( )hour, The growth rate is relatively low. For the control rights allocation model of the aggressive driving group, the comprehensive deviation When smaller (such as ), the driver always takes the lead, only when the comprehensive deviation The intelligent system intervenes only when a certain threshold is exceeded, which is more suitable for the driving style of aggressive drivers. For novice drivers, the intelligent system intervenes relatively early and quickly, compensating for their lateral control ability and improving driving safety.

[0025] Considering the evolution of drivers' driving behavior habits, the driving group label they belong to will change dynamically. Therefore, a flexible control rights allocation strategy is designed.

[0026] (1) The default label of the driving group is the general driving group; (2) According to the identification results of the driving group, a weighted combination method is used to achieve flexible allocation of human-machine control rights, that is, When the real-time identification result is the general driving group,

[0027] When the real-time identification result is an aggressive driving group,

[0028] Where, 、 is the dynamic weight coefficient, which is determined by the frequency of identification results in a time sliding window of length L, where L is a design parameter.

[0029] When the real-time identification result is a novice driver group,

[0030] Where, 、 is the dynamic weight coefficient, which is determined by the frequency of the identification results in the long time sliding window L.

[0031] In a specific embodiment, the process of executing step S101 may specifically include the following steps: (1) Obtaining raw driving data during vehicle driving through millimeter-wave radar and visual sensors, and extracting basic driving parameters from the raw driving data; (2) Perform multi-sensor data fusion processing on basic driving parameters to generate real-time following vehicle distance data, lateral displacement data, and vehicle acceleration data; (3) Based on the real-time following vehicle distance data and lateral displacement data, the number of sudden braking, sudden acceleration and sharp turning times are calculated through the sliding time window algorithm to obtain the driving operation frequency data; (4) Using vehicle acceleration data and driving operation frequency data, a data normalization algorithm is used to perform feature normalization processing to obtain standardized driving characteristics; (5) Preprocessing the standardized driving features through the CAN bus to generate a preprocessing feature sequence; (6) The preprocessed feature sequence is subjected to feature combination mapping operation to output driving behavior feature data.

[0032] Specifically, raw driving data is collected from the vehicle using millimeter-wave radar and visual sensors. The millimeter-wave radar operates in the 77 GHz frequency band, with a detection range of 0-200 meters, an angular resolution of 0.1 degrees, and a sampling frequency of 20 Hz. It primarily collects information such as the distance and relative speed of the vehicle ahead. The visual sensor uses a 1920×1080 high-definition camera with a frame rate of 30 fps and a field of view of 120 degrees. It primarily collects lane information and the positions of surrounding vehicles. Preliminary processing is performed on the collected raw driving data to extract basic driving parameters, including the raw value of the distance to the preceding vehicle, the raw value of the lane departure, and the raw value of the speed. Multi-sensor data fusion is used to integrate the basic driving parameters obtained from multiple sensors. Data fusion uses the Kalman filter algorithm to perform time alignment and spatial registration of data collected by different sensors. Time alignment uses linear interpolation to standardize data with different sampling frequencies to a single sampling frequency of 100 Hz. Spatial registration transforms data from different coordinate systems into a unified vehicle coordinate system using a sensor calibration matrix. After fusion processing, high-precision real-time vehicle-to-vehicle distance data is generated with an accuracy of 0.1 meter; lateral displacement data with an accuracy of centimeters; and vehicle acceleration data with an accuracy of 0.01m / s².

[0033] Based on the fused real-time following vehicle distance data and lateral displacement data, a sliding time window algorithm is used to calculate driving operation frequency. The sliding window length is set to 30 seconds, with a sliding step of 1 second. Within each window, the number of sudden braking, sudden acceleration, and sudden turning events are counted. The sudden braking criterion is a deceleration greater than 4 m / s² and lasting longer than 0.5 seconds; the sudden acceleration criterion is an acceleration greater than 3 m / s² and lasting longer than 0.5 seconds; and the sudden turning criterion is a lateral acceleration greater than 2 m / s² and lasting longer than 0.3 seconds. By detecting and counting these sudden behaviors, driving operation frequency data is generated. Feature normalization is performed on the vehicle acceleration data and driving operation frequency data. Using the z-score normalization method, the mean and standard deviation of each feature are calculated, transforming the data to a distribution with a mean of 0 and a standard deviation of 1. For the acceleration data, statistical features are calculated for the longitudinal and lateral accelerations. For the driving operation frequency data, the number of each type of operation within the 30-second window is converted to a frequency value per unit time. The standardization process ensures that features of different dimensions are comparable and obtains standardized driving features.

[0034] Standardized driving characteristics are preprocessed via the CAN bus. The CAN bus uses a transmission rate of 500kbps to collect vehicle status information such as speed, steering wheel angle, and accelerator pedal opening. The collected data is filtered for noise, and a median filter algorithm is used to eliminate outliers with a filter window size of 5 sampling points. The data is then downsampled to 10Hz while retaining key data features to generate a preprocessed feature sequence. Finally, the preprocessed feature sequence undergoes a feature combination mapping operation. Principal component analysis is used to extract the main feature components, retaining feature dimensions with a cumulative contribution rate of 95%. A feature selection algorithm is then used to calculate the importance score of each feature. Features with higher scores are selected to form the final feature set, which is then used to output the driving behavior feature data.

[0035] For example, on a busy road, the distance to the preceding vehicle recorded by millimeter-wave radar fluctuates between 45 and 65 meters, with an average distance of 55 meters. The visual sensor detects a lateral deviation of the vehicle relative to the lane centerline, varying between -0.3 and 0.3 meters. Multi-sensor fusion processing generates a stable sequence of following vehicle distance data, eliminating the measurement noise and uncertainty of individual sensors. Within a 30-second sliding window, two sudden braking maneuvers (maximum decelerations of 4.5 m / s² and 4.2 m / s², respectively), one sudden acceleration maneuver (maximum acceleration of 3.3 m / s²), and one sharp turn (maximum lateral acceleration of 2.3 m / s²) are detected. These raw data are normalized, resulting in feature values ​​distributed within the range [-3, 3]. Vehicle speeds collected via the CAN bus remain between 80 and 90 km / h, and the steering wheel angle varies within ±45 degrees. After feature combination mapping, the resulting driving behavior feature data contains 15 key dimensions, fully describing the driving behavior during that period.

[0036] In a specific embodiment, the process of executing step S102 may specifically include the following steps: (1) Input the driving behavior feature data into the pre-trained artificial feature extraction layer to generate an artificial feature vector; (2) Extract the driving behavior feature data through the CNN convolution layer to obtain the CNN feature vector; (3) Combine the artificial feature vector and the CNN feature vector in series to obtain the fused feature vector; (4) Perform deep neural network training based on the fused feature vector to obtain deep feature representation; (5) Using deep feature representation and prior knowledge to perform knowledge transfer learning and form knowledge-enhanced features; (6) The knowledge-enhanced features are classified and learned through a multi-layer perceptron to construct a driving group recognition model.

[0037] Specifically, the driving behavior feature data is input into a pre-trained artificial feature extraction layer. The artificial feature extraction layer contains multiple statistical feature extraction units that calculate statistical quantities such as the mean, standard deviation, kurtosis, and skewness of the driving behavior data. For the following distance data, features such as the average following distance, distance fluctuation range, and distance change rate are extracted; for the lateral displacement data, features such as the average offset, maximum offset, and offset frequency are extracted; and for the acceleration data, information such as the acceleration distribution characteristics and acceleration change rate is extracted. After artificial feature extraction, a 128-dimensional artificial feature vector is generated. Simultaneously, a CNN convolutional network is used to perform deep feature extraction on the driving behavior feature data. The CNN network architecture consists of three convolutional layers. The first layer uses 32 3×3 convolution kernels with a stride of 1, resulting in an output feature map of 32×32×32 size. The second layer uses 64 5×5 convolution kernels with a stride of 2, resulting in an output feature map of 16×16×64 size. The third layer uses 128 7×7 convolution kernels with a stride of 2, resulting in an output feature map of 8×8×128 size. Each convolutional layer is followed by a ReLU activation function and a max pooling layer with a pooling kernel size of 2×2 and a stride of 2. Finally, a fully connected layer converts the feature map into a 256-dimensional CNN feature vector.

[0038] The 128-dimensional manual feature vector and the 256-dimensional CNN feature vector are concatenated to form a 384-dimensional fused feature vector. This concatenation operation retains the advantages of both feature extraction methods: the manual feature extraction layer provides interpretable statistical features, while the CNN feature extraction layer captures the deep spatiotemporal characteristics of the data. The fused feature vector contains comprehensive information about driving behavior, providing a rich feature representation for subsequent deep learning. A deep neural network is constructed based on the fused feature vector for training. The network structure consists of three fully connected layers with 256, 128, and 64 nodes, respectively. The first and second layers use the ReLU activation function, and a Dropout layer is added to prevent overfitting with a dropout rate of 0.5. During training, a mini-batch training method with a batch size of 32 is used, the Adam optimizer is used, a learning rate of 0.001, and 100 training epochs are used. The network parameters are optimized using the backpropagation algorithm, ultimately resulting in a 64-dimensional deep feature representation.

[0039] Deep feature representation is used for knowledge transfer learning, incorporating prior knowledge from the field of driving behavior. This prior knowledge includes typical characteristic patterns of different driving groups, such as the stability of average drivers, the quick response of aggressive drivers, and the cautiousness of novice drivers. Through transfer learning, this prior knowledge is integrated with the deep feature representation to enhance the discriminative power of the features. Specifically, knowledge distillation is employed, using a pre-trained model as a teacher network to guide feature learning, resulting in a 96-dimensional knowledge-enhanced feature. Finally, the knowledge-enhanced feature is input into a multilayer perceptron for classification learning. The multilayer perceptron consists of two hidden layers, with 64 and 32 nodes, respectively, and uses the ReLU activation function. The output layer uses a softmax activation function and outputs three class probabilities: average drivers, aggressive drivers, and novice drivers. During the training phase, the cross-entropy loss function is used to optimize the network parameters via backpropagation. The training dataset consists of 5,000 labeled samples, with 20% used as a validation set and 10% as a test set. After training, the model achieved a classification accuracy of 92% on the test set.

[0040] For example, the input driving behavior feature data contains 15 dimensions. After processing through the artificial feature extraction layer, the calculated statistical features show an average following distance of 52 meters, a standard deviation of 5.3 meters, a mean lateral offset of 0.15 meters, and an acceleration kurtosis of 2.8, generating a 128-dimensional artificial feature vector. Simultaneously, the CNN network performs convolution on the input data. The first convolution layer extracts basic edge features, the second layer captures speed variation patterns, and the third layer identifies complete driving behavior segments, ultimately generating a 256-dimensional CNN feature vector. The two feature vectors are concatenated and trained through a deep neural network. The resulting deep feature representation clearly reflects the driver's operating characteristics. By combining prior knowledge for transfer learning, the knowledge-enhanced features better highlight group characteristics. The multi-layer perceptron classification results show that the probability of this driver belonging to the average driver group is 0.85, the probability of belonging to the aggressive driver group is 0.1, and the probability of belonging to the novice driver group is 0.05, ultimately classifying him as an average driver.

[0041] In a specific embodiment, the process of executing step S103 may specifically include the following steps: (1) Extract features from real-time driving behavior data using a driving group recognition model to generate real-time feature vectors; (2) Calculate the probability distribution of scores based on the real-time feature vector to form the probability value of group categories; (3) Use the group category probability value to set the threshold for multi-classification discrimination and output the initial group label; (4) Perform time series smoothing on the initial group labels to obtain smoothed group labels; (5) Perform group category mapping based on smoothed group labels to obtain standardized group labels; (6) The standardized group labels are processed through label fusion to generate general driving group labels, aggressive driving group labels and novice driving group labels.

[0042] Specifically, features are extracted from real-time driving behavior data using a pre-trained driving group recognition model. This process employs a dual-path feature extraction architecture. The manual feature extraction path performs statistical analysis on the driving data, including calculating the mean, standard deviation, maximum and minimum values ​​of the following distance, the range and frequency of lateral deviation, and the distribution characteristics of acceleration. The CNN feature extraction path extracts temporal features using a three-layer convolutional network with kernel sizes of 3×3, 5×5, and 7×7, respectively, followed by a max pooling operation. The features extracted by the dual paths are combined in series to generate a 384-dimensional real-time feature vector. A score probability distribution is calculated based on the real-time feature vector, and the softmax function is used to map the features into probabilities for the three driving group categories. The softmax function converts the output of the deep neural network into probability values ​​in the range [0, 1], where the sum of the probabilities of the three categories is 1. The probability calculation takes into account the weight contribution of each dimension in the feature vector, and a weighted combination is performed using the weight matrix of the fully connected layer to form the group category probability value.

[0043] A threshold is set for the calculated group category probability values ​​for multi-classification discrimination, and the threshold is set to 0.6. When the probability value of a certain category exceeds the threshold, the driver is judged to belong to the corresponding driving group. If the probability values ​​of all categories do not exceed the threshold, the category with the highest probability is selected as the judgment result. The initial group label is obtained through threshold discrimination. In order to reduce the control instability caused by frequent label switching, the initial group label is smoothed in time. A sliding time window of 5 seconds is used to perform statistical analysis on the label sequence within the window. The smoothing process uses the majority voting method to select the label with the most occurrences in the window as the label at the current moment. At the same time, a state transition constraint is introduced, requiring that the label changes at adjacent moments must meet certain smoothness conditions to obtain a smoothed group label.

[0044] Group category mapping is performed based on smoothed group labels, converting discrete label values ​​into a standardized representation. This mapping process considers the similarities and differences between different driving groups and uses one-hot encoding to convert group labels into a three-dimensional vector representation. The general driving group is mapped to [1, 0, 0], the aggressive driving group is mapped to [0, 1, 0], and the novice driving group is mapped to [0, 0, 1] to obtain standardized group labels. The standardized group labels are then fused to generate the final driving group labels. The fusion process considers temporal continuity and spatial consistency, accumulating historical labels using an exponentially weighted average. The weight coefficients decay exponentially over time, with more recent labels having higher weights. Through this fusion process, stable general driving group labels, aggressive driving group labels, and novice driving group labels are obtained.

[0045] For example, real-time driving behavior data for a driver shows an average following distance of 55 meters, a standard deviation of 4.8 meters, a peak lateral offset of 0.25 meters, and one sudden acceleration maneuver recorded within 30 seconds. Feature extraction yields a 384-dimensional feature vector. A softmax calculation yields probabilities for the three categories: 0.75 for the average driver group, 0.15 for the aggressive driver group, and 0.10 for the novice driver group. Because the probability of the average driver group exceeds the threshold of 0.6, the initial label is assigned to the average driver group. Within a 5-second window, five consecutive classifications consistently yield the average driver group. Smoothing maintains the label stability. Normalized mapping yields a vector representation of [1, 0, 0]. The final label fusion confirms that the driver belongs to the average driver group, providing a reliable basis for group characteristics in subsequent control allocation.

[0046] In a specific embodiment, the process of executing step S104 may specifically include the following steps: (1) Extract the preview point location data from the vehicle status information to obtain the real-time coordinates of the preview point; (2) Calculate the original value of the lateral deviation based on the real-time coordinates of the preview point to generate the lateral deviation of the preview point; (3) Calculate the original value of the yaw angle using the vehicle posture data to obtain the yaw angle deviation; (4) Normalize the lateral deviation and yaw angle deviation of the preview point to obtain the normalized deviation value; (5) Select the corresponding weighting coefficient according to the driving group label to generate the deviation weighting parameter; (6) The normalized deviation value and the deviation weighting parameter are combined to obtain the vehicle lateral comprehensive deviation.

[0047] Specifically, the preview point position data is extracted from the vehicle status information. The preview point is set 30 meters in front of the vehicle, and the vehicle's global position coordinates (x, y) and heading angle θ are obtained through on-board sensors. Based on the vehicle's current position and preview distance, the real-time coordinates (xp, yp) of the preview point are calculated using coordinate transformation. The coordinate transformation takes into account the influence of the vehicle's heading angle to ensure that the preview point is always in the direction of the vehicle's forward movement. Based on the real-time coordinates of the preview point, the lateral deviation is calculated. First, the reference trajectory equation of the lane centerline is obtained, and a third-order polynomial curve fitting is used to describe the shape of the lane line. The shortest distance from the preview point to the reference trajectory is calculated, which is the original value of the lateral deviation. In the specific calculation, the iterative nearest point algorithm is used to find the point on the reference trajectory closest to the preview point. The vertical distance between the two points is the lateral deviation of the preview point.

[0048] The yaw angle deviation is calculated using vehicle attitude data. Vehicle attitude data, including heading angle and yaw rate, is collected by an inertial measurement unit (IMU) at a sampling frequency of 100 Hz. The angle between the vehicle heading angle and the tangent of the reference trajectory is calculated to obtain the raw yaw angle value. Taking into account the vehicle's dynamic characteristics, a yaw rate term is introduced to compensate for the angle, improving the accuracy of the angle estimation and ultimately determining the yaw angle deviation. The preview point lateral deviation and yaw angle deviation are normalized. The normalized range for the lateral deviation is [-1, 1], corresponding to an actual deviation range of [-2m, 2m]. The normalized range for the yaw angle deviation is also [-1, 1], corresponding to an actual angle range of [-10°, 10°]. A linear normalization method is used to map the raw values ​​to a uniform numerical range, facilitating the subsequent weighted combination calculation to obtain the normalized deviation value.

[0049] Appropriate weighting coefficients are selected based on the characteristics of different driving groups. For the general driver group, the lateral deviation weight is 0.6, and the yaw angle deviation weight is 0.4, reflecting a moderate requirement for lane keeping. For the aggressive driver group, the lateral deviation weight is 0.7, and the yaw angle deviation weight is 0.3, emphasizing greater operational freedom. For the novice driver group, the lateral deviation weight is 0.5, and the yaw angle deviation weight is 0.5, reflecting a more conservative control strategy. Based on the current driving group label, the corresponding weighting coefficient combination is selected to generate the deviation weighting parameter. The normalized deviation value is combined with the deviation weighting parameter and a weighted summation method is used to calculate the vehicle's comprehensive lateral deviation. The comprehensive deviation value is also normalized to the range [0, 1]. Larger values ​​indicate a greater deviation from the desired driving state and require stronger control intervention.

[0050] For example, a vehicle is traveling on a curve at 80 km / h. The GPS obtains the vehicle's current position coordinates (100.5 m, 50.3 m) and a heading angle of 45 degrees. The coordinates of the preview point 30 meters ahead are calculated as (121.7 m, 71.5 m). The reference trajectory equation provided by the lane recognition system shows that the raw lateral deviation at the preview point is 0.8 meters. The vehicle attitude data indicates a current yaw angle of 3 degrees and a yaw rate of 0.2 rad / s. Normalization yields a normalized lateral deviation of 0.4 and a normalized yaw deviation of 0.3. The current driver group is identified as the average driver group. Using weighting coefficients of 0.6 and 0.4, the calculated lateral composite deviation is 0.36, indicating a controllable deviation requiring appropriate control intervention to maintain lane control.

[0051] In a specific embodiment, the process of executing step S105 may specifically include the following steps: (1) Establishing the first exponential function based on the vehicle's lateral comprehensive deviation to generate the deviation weight factor; (2) Using the driver's steering torque to establish a second exponential function, the steering weight factor is obtained; (3) Combine the deviation weight factor and the steering weight factor into a function to obtain the control weight reference value; (4) Calculate the initial weight of the intelligent system based on the control weight benchmark value to obtain the control weight coefficient of the intelligent system; (5) Calculate the driver's initial weight based on the complementary relationship between the intelligent system and the driver's weight; (6) Normalize the driver's initial weight to obtain the driver's control weight coefficient.

[0052] Specifically, in the process of allocating control rights for human-machine co-driving, considering the characteristics of the driving group, a first exponential function is first established based on the vehicle's lateral integrated deviation. This exponential function uses the lateral integrated deviation as the independent variable and generates a deviation weight factor through exponential mapping. Different exponential parameters are used for different driving groups: 4 for the general driving group, 4 for the aggressive driving group, and 20 for the novice driving group. The deviation weight factor's value range is limited to 0 to 1, reflecting the degree of influence of lateral deviation on control rights allocation. A second exponential function is established using the driver's steering torque. This function uses the steering torque as the input variable and generates a steering weight factor through exponential calculation. Different exponential parameters are set for different driving groups: 3 for the general driving group, 10 for the aggressive driving group, and 3 for the novice driving group. The steering weight factor is also limited to the range of 0 to 1, reflecting the strength of the driver's steering intention.

[0053] The deviation weight factor and steering weight factor are combined into a functional combination, using a product form to construct the combined function. During the combination process, the deviation weight factor reflects the vehicle's deviation degree, while the steering weight factor reflects the driver's operating intention. The product of the two forms the control weight baseline value, achieving the coupling of deviation and steering intention. The initial weight of the intelligent system is calculated based on the control weight baseline value. An adjustment factor of 0.5 is introduced during the calculation process to adjust the baseline value to ensure that the control weight of the intelligent system remains within a reasonable range. This adjusted value becomes the control weight coefficient of the intelligent system and is used for subsequent control rights allocation.

[0054] The driver's initial weight is calculated based on the complementary relationship between the intelligent system and the driver's weights. Under the constraint that the total control authority is 1, the driver's initial weight is equal to 1 minus the intelligent system's control weight coefficient. This calculation method ensures a smooth distribution of control authority between the human and the machine. The driver's initial weight is normalized, mapping the weight value to the range of 0 to 1. This normalization process takes into account the historical maximum and minimum weight values, and a standardized driver control weight coefficient is obtained through linear mapping.

[0055] For example, when a vehicle is traveling on a curved road, the measured lateral integrated deviation is 0.36, the driver's steering torque is 2 Nm, and the normalized steering torque value is 0.4. The current driver group is identified as the general group, and the calculated deviation weight factor is 0.23, and the steering weight factor is 0.70. After function combination, the control weight baseline value is 0.161. Further calculation results in an intelligent system control weight coefficient of 0.098, and an initial driver weight of 0.902. Finally, after normalization, the driver control weight coefficient is 0.85, indicating that the driver maintains a dominant position in the current scenario, with the intelligent system playing a supporting role.

[0056] In a specific embodiment, the process of executing step S106 may specifically include the following steps: (1) Counting the group recognition results based on the time sliding window to generate group recognition frequency data; (2) Calculate the dynamic weight coefficient using group identification frequency data to obtain the group weight parameter; (3) The intelligent system control weight coefficient and the group weight parameter are weighted together to obtain the final weight of the intelligent system; (4) Perform a weighted combination of the driver control weight coefficient and the group weight parameter to generate the driver's final weight; (5) Calculate the control rights allocation based on the final weight of the intelligent system and the final weight of the driver; (6) Normalize the control rights allocation calculation results to obtain the control rights allocation results.

[0057] Specifically, a time sliding window is used to count the group recognition results. The sliding window length is set to 10 seconds, with a sliding step of 1 second. The number of times the general driving group, the aggressive driving group, and the novice driving group are recognized in each window is recorded. The frequency of occurrence of each group is obtained through counting statistics to form group recognition frequency data. The dynamic weight coefficient is calculated based on the group recognition frequency data. The calculation process takes into account the proportion of each driving group in the window and converts the frequency data into weight parameters. Driving groups with higher occurrence frequencies are assigned larger weight values, while driving groups with lower occurrence frequencies are assigned smaller weight values. The group weight parameters are obtained through the mapping relationship from frequency to weight.

[0058] A weighted combination of the intelligent system's control weight coefficient and the group weight parameter is performed. During this combination, the control characteristics of different driving groups are considered, and the intelligent system's control weight is dynamically adjusted. When the identification result favors the novice driving group, the intelligent system's weight is increased; when the identification result favors the aggressive driving group, the intelligent system's weight is decreased, thereby obtaining the final intelligent system weight. A similar weighted combination is performed on the driver's control weight coefficient and the group weight parameter. This combination considers the impact of the driving group's characteristics on the driver's control authority. When the driver is identified as an aggressive driver, the driver's control weight is increased; when the driver is identified as a novice driver, the driver's control weight is appropriately decreased, ultimately generating the driver's final weight.

[0059] Control rights allocation is calculated based on the final weights of the intelligent system and the driver. This calculation ensures that the sum of the two weights is 1, while also taking into account the smoothness of weight changes to avoid operational inconsistencies caused by sudden changes in control rights. The control rights allocation calculation results are normalized to ensure that the final control rights allocation is within a reasonable range. This normalization process uses a linear mapping method to map the weight values ​​to a standard range, resulting in a standardized control rights allocation result.

[0060] For example, within a 10-second sliding window, the driver group identification results showed six occurrences for the general driver group, three for the aggressive driver group, and one for the novice driver group. Group weight parameters were calculated based on the frequency data: 0.6 for the general driver group, 0.3 for the aggressive driver group, and 0.1 for the novice driver group. The intelligent system's original control weight coefficient was 0.3, and the final weight after weighted combination was 0.25. The driver's original control weight coefficient was 0.7, and the final weight after weighted combination was 0.75. The control rights allocation remained unchanged after normalization, reflecting the driver-led, intelligent system-assisted control model in the current driving scenario.

[0061] The above describes the method for allocating human-machine co-driving control rights considering the characteristics of the driving group in the embodiment of the present application. The following describes the device for allocating human-machine co-driving control rights considering the characteristics of the driving group in the embodiment of the present application. Figure 2 In one embodiment of the present application, a device for allocating control rights for human-machine co-driving taking into account the characteristics of a driving group includes: Extraction module 201, for collecting vehicle operation data through on-board sensors, extracting vehicle-to-vehicle distance data, lateral displacement data, vehicle acceleration data, and driving operation frequency data from the vehicle operation data to obtain driving behavior characteristic data; A construction module 202 is configured to establish a deep neural network model using the driving behavior feature data, and to construct a driving group recognition model based on a dual-path architecture of manually extracted features and CNN convolutional features; A discrimination module 203 is configured to classify and discriminate the real-time behaviors of the drivers based on the driving group recognition model, and generate a general driving group label, an aggressive driving group label, and a novice driving group label; a combination module 204 for obtaining corresponding control parameters according to the driving group label, and calculating a vehicle lateral comprehensive deviation based on a normalized weighted combination of the preview point lateral deviation and the yaw angle deviation; An allocation module 205 is configured to utilize the vehicle's lateral comprehensive deviation and the driver's steering torque to establish a control right allocation mathematical model through a double exponential function to obtain an intelligent system control weight coefficient and a driver control weight coefficient; The weighting module 206 is configured to dynamically weight the intelligent system control weight coefficient and the driver control weight coefficient according to the driving group label and the group identification frequency data within the time window to generate a control right allocation result.

[0062] Through the coordinated cooperation of the above-mentioned components, the present invention collects vehicle operation data through on-board sensors and extracts vehicle-to-vehicle distance data, lateral displacement data, vehicle acceleration data, and driving operation frequency data, providing rich data support for the comprehensive analysis of driving behavior characteristics and ensuring the accurate capture of driver operation characteristics; using driving behavior feature data to establish a deep neural network model, adopting a dual-path architecture of manual feature extraction and CNN convolution features, improving the accuracy and robustness of feature extraction, and enhancing the model's ability to recognize different driving behavior patterns; using the driving group recognition model to classify and judge the driver's real-time behavior, generate general driving group labels, aggressive driving group labels, and novice driving group labels, realizing the recognition of driving Accurate identification and dynamic tracking of driver group characteristics; obtaining corresponding control parameters based on driving group labels, combining the normalized weighted combination of preview point lateral deviation and yaw angle deviation, realizing accurate calculation of vehicle lateral comprehensive deviation, and providing a reliable basis for control right allocation; using vehicle lateral comprehensive deviation and driver steering torque to construct a double exponential function control right allocation mathematical model, obtaining intelligent system control weight coefficient and driver control weight coefficient, ensuring smoothness and continuity of control right allocation; dynamically weighting and combining control weight coefficients according to driving group labels and group recognition frequency data within the time window, so that the control right allocation result can adapt to the dynamic changes of driver behavior characteristics, and improve the adaptability and coordination of the human-machine co-driving system. Through the organic combination of the above technical features, the present invention not only improves the accuracy of driving group identification, but also realizes adaptive allocation of control rights, effectively reduces control right conflicts in the process of human-machine co-driving, while ensuring driving safety and comfort, and providing personalized human-machine co-driving experience for different driving groups.

[0063] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for allocating human-machine co-driving control rights considering the characteristics of a driving group, characterized in that: The method for allocating human-machine co-driving control rights considering driving group characteristics includes: Collecting vehicle operation data through on-board sensors, extracting following vehicle distance data, lateral displacement data, vehicle acceleration data, and driving operation frequency data from the vehicle operation data to obtain driving behavior characteristic data; A deep neural network model is established using the driving behavior feature data, and a driving group recognition model is constructed based on a dual-path architecture of manually extracted features and CNN convolutional features; Classifying and distinguishing the real-time behaviors of drivers according to the driving group identification model to generate general driving group labels, aggressive driving group labels and novice driving group labels; Acquire corresponding control parameters according to the driving group label, and calculate the vehicle lateral comprehensive deviation based on the normalized weighted combination of the preview point lateral deviation and the yaw angle deviation; By using the vehicle lateral comprehensive deviation and the driver's steering torque, a control right allocation mathematical model is established through a double exponential function to obtain an intelligent system control weight coefficient and a driver control weight coefficient; According to the driving group label and the group identification frequency data within the time window, the intelligent system control weight coefficient and the driver control weight coefficient are dynamically weighted and combined to generate a control right allocation result.

2. The method for allocating human-machine co-driving control rights considering driving group characteristics according to claim 1 is characterized in that: The vehicle operation data is collected by the vehicle-mounted sensor, and the following vehicle distance data, lateral displacement data, vehicle acceleration data and driving operation frequency data are extracted from the vehicle operation data to obtain driving behavior characteristic data, including: Acquire raw driving data of the vehicle during driving by using a millimeter-wave radar and a visual sensor, and extract basic driving parameters from the raw driving data; Perform multi-sensor data fusion processing on the basic driving parameters to generate real-time following vehicle distance data, lateral displacement data and vehicle acceleration data; Based on the real-time following vehicle distance data and lateral displacement data, the number of emergency braking, the number of emergency acceleration and the number of emergency turning are calculated by a sliding time window algorithm to obtain driving operation frequency data; Using the vehicle acceleration data and the driving operation frequency data, a data normalization algorithm is used to perform feature normalization processing to obtain standardized driving features; Performing data preprocessing on the standardized driving features via a CAN bus to generate a preprocessing feature sequence; The preprocessed feature sequence is subjected to feature combination mapping operation to output the driving behavior feature data.

3. The method for allocating human-machine co-driving control rights considering driving group characteristics according to claim 1 is characterized in that: The method of establishing a deep neural network model using the driving behavior feature data and constructing a driving group recognition model based on a dual-path architecture of manually extracted features and CNN convolutional features includes: Inputting the driving behavior feature data into a pre-trained artificial feature extraction layer to generate an artificial feature vector; Extracting features from the driving behavior feature data through a CNN convolutional layer to obtain a CNN feature vector; The artificial feature vector and the CNN feature vector are combined in series to obtain a fused feature vector; Performing deep neural network training based on the fused feature vector to obtain deep feature representation; Using the deep feature representation and prior knowledge to perform knowledge transfer learning to form knowledge enhancement features; The knowledge enhancement features are classified and learned through a multi-layer perceptron to construct the driving group recognition model.

4. The method for allocating human-machine co-driving control rights considering driving group characteristics according to claim 1 is characterized in that: The method of classifying and judging the real-time behaviors of drivers according to the driving group recognition model to generate general driving group labels, aggressive driving group labels and novice driving group labels includes: Extracting features from real-time driving behavior data using the driving group recognition model to generate a real-time feature vector; Calculate the score probability distribution according to the real-time feature vector to form a group category probability value; Using the group category probability value to set a threshold for multi-classification discrimination, and output an initial group label; Performing time series smoothing processing on the initial group labels to obtain smoothed group labels; Perform group category mapping based on the smoothed group labels to obtain standardized group labels; The standardized group labels are subjected to label fusion processing to generate the general driving group label, the aggressive driving group label and the novice driving group label.

5. The method for allocating human-machine co-driving control rights considering driving group characteristics according to claim 1 is characterized in that: The step of acquiring corresponding control parameters according to the driving group label and calculating the vehicle lateral comprehensive deviation based on the normalized weighted combination of the preview point lateral deviation and the yaw angle deviation includes: Extract the preview point position data from the vehicle status information to obtain the real-time coordinates of the preview point; Calculate the original value of the lateral deviation based on the real-time coordinates of the preview point to generate the lateral deviation of the preview point; The original value of the yaw angle is calculated using the vehicle posture data to obtain the yaw angle deviation; Normalizing the preview point lateral deviation and the yaw angle deviation to obtain a normalized deviation value; Selecting a corresponding weighting coefficient according to the driving group label to generate a deviation weighting parameter; The normalized deviation value and the deviation weighting parameter are combined and calculated to obtain the vehicle lateral comprehensive deviation.

6. The method for allocating human-machine co-driving control rights considering driving group characteristics according to claim 1 is characterized in that: The method uses the vehicle lateral comprehensive deviation and the driver's steering torque to establish a control right allocation mathematical model through a double exponential function to obtain an intelligent system control weight coefficient and a driver control weight coefficient, including: Establishing a first exponential function according to the vehicle lateral comprehensive deviation to generate a deviation weight factor; Using the driver's steering torque to establish a second exponential function to obtain a steering weight factor; Performing a functional combination of the deviation weight factor and the steering weight factor to obtain a control weight reference value; Calculate the initial weight of the intelligent system according to the control weight reference value to obtain the control weight coefficient of the intelligent system; Calculate the initial weight of the driver based on the complementary relationship between the intelligent system and the driver's weight; The driver's initial weight is normalized to obtain the driver's control weight coefficient.

7. The method for allocating human-machine co-driving control rights considering driving group characteristics according to claim 1 is characterized in that: The step of dynamically weighting and combining the intelligent system control weight coefficient and the driver control weight coefficient according to the driving group label and the group identification frequency data in the time window to generate a control right allocation result includes: Count the group recognition results based on the time sliding window to generate group recognition frequency data; Calculating a dynamic weight coefficient using the group identification frequency data to obtain a group weight parameter; Performing a weighted combination of the intelligent system control weight coefficient and the group weight parameter to obtain a final weight of the intelligent system; Performing a weighted combination on the driver control weight coefficient and the group weight parameter to generate a final driver weight; Calculating the control right distribution according to the final weight of the intelligent system and the final weight of the driver; The control right allocation calculation result is normalized to obtain the control right allocation result.

8. A device for allocating human-machine co-driving control rights considering the characteristics of a driving group, characterized in that: The human-machine co-driving control rights allocation device considering the driving group characteristics includes: An extraction module is used to collect vehicle operation data through vehicle-mounted sensors, extract vehicle-to-vehicle distance data, lateral displacement data, vehicle acceleration data, and driving operation frequency data from the vehicle operation data, and obtain driving behavior characteristic data; A construction module is used to establish a deep neural network model using the driving behavior feature data, and to construct a driving group recognition model based on a dual-path architecture of manually extracted features and CNN convolutional features; A discrimination module, used to classify and discriminate the real-time behaviors of drivers according to the driving group recognition model, and generate general driving group labels, aggressive driving group labels and novice driving group labels; A combination module, used for obtaining corresponding control parameters according to the driving group label, and calculating the vehicle lateral comprehensive deviation based on the normalized weighted combination of the preview point lateral deviation and the yaw angle deviation; A distribution module, used to establish a control right distribution mathematical model by using the vehicle lateral comprehensive deviation and the driver's steering torque through a double exponential function to obtain an intelligent system control weight coefficient and a driver control weight coefficient; The weighting module is used to dynamically weight the intelligent system control weight coefficient and the driver control weight coefficient according to the driving group label and the group identification frequency data in the time window to generate a control right allocation result.

Images