Urban logistics vehicle intelligent driving-human driving mode switching method considering driving risk and energy consumption

Abstract

The invention relates to the technical field of intelligent driving, and discloses an urban logistics vehicle intelligent driving-human driving mode switching method considering driving risks and energy consumption. Comprising the steps of driver mode switching driving data extraction, mode switching interval acquisition considering driver styles and surrounding vehicle trajectory prediction, and mode switching opportunity optimization based on driving risks and limit energy consumption. The method comprises the following steps of: obtaining a risk interval and a risk change curve of a mode switching decision by combining a style identification result and utilizing a risk evaluation model according to historical characteristic data and surrounding vehicle track prediction in a driver mode switching process, and obtaining a time interval of the mode switching decision based on the risk interval and the risk change curve; and finally, an optimization model is established based on human driving mode speed prediction, and the switching opportunity is optimized. According to the method, the driving risk and the driving energy consumption are comprehensively considered, the problems of high driving energy consumption and insufficient safety caused by unreasonable man-machine co-driving mode switching of the intelligent driving logistics vehicle in a complex urban scene are effectively solved, and the passing efficiency of the urban logistics vehicle is improved.

Description

Technical Field

[0001] The present application relates to the field of intelligent driving technology, and in particular to a method for switching between intelligent driving and human driving modes for urban logistics vehicles that takes driving risks and energy consumption into consideration. Background Art

[0002] In urban roads with complex logistics and transportation environments, driving scenarios are diverse and highly time-varying. How to optimize energy consumption while maintaining the driving safety of human-machine co-driving vehicles is an important challenge currently faced. Therefore, under the premise of safe driving, it is urgent to study the predictive switching method of intelligent driving-human driving mode based on driving risks and energy consumption, and realize the optimized switching between intelligent driving mode and manual driving mode through real-time prediction and evaluation of driving risks and energy consumption, so as to minimize energy consumption.

[0003] However, the existing intelligent driving-human driving mode switching method mainly focuses on a single aspect of driving safety or energy consumption optimization, and fails to comprehensively consider the balance between the two. For example, the prior art published on November 5, 2021: the application publication number is CN 113602284 A, which specifically discloses a human-machine co-driving mode decision method, device, equipment and storage medium, wherein the method senses the driver's physiological characteristic indicators through a watch; determines the corresponding driving mode through a preset adaptive learning model according to the physiological characteristic indicators; obtains vehicle information within a preset range in front; generates a corresponding vehicle intention model through a preset behavior prediction algorithm according to the vehicle information; determines the risk level of the current vehicle according to the vehicle intention model; and determines the target mode switching decision according to the driving mode and risk level. This method solves the lag in human-machine co-driving mode switching based on the driver's objective physiological state and driving risk level prediction, but because different drivers have different subjective driving styles, only considering the objective state will lead to a poor driving experience for the driver, and the driving decision does not take the vehicle energy consumption into consideration, so the target mode switching decision has problems such as poor driving experience and high energy consumption. Summary of the invention

[0004] In order to solve the above problems, the present application provides a method for switching between intelligent driving and human driving modes for urban logistics vehicles that takes into account driving risks and energy consumption, so as to solve the problems of poor driving experience and high energy consumption in target mode switching decisions when facing scenarios with complex urban road conditions and multi-vehicle interference and interaction.

[0005] In order to achieve the above purpose, the technical solution adopted in this application is as follows:

[0006] The present application provides a method for switching between intelligent driving and human driving modes of an urban logistics vehicle taking into account driving risks and energy consumption, the method comprising:

[0007] According to the driving characteristic data of drivers in different driving scenarios, clustering algorithms are used to classify drivers, and driver style recognition models based on deep learning are used to identify driver style, and the mode switching process driving data of the current driver style category is obtained;

[0008] According to the driving data of the mode switching process of the current driver style category, the driving risk value during mode switching is calculated based on the risk field model, and the left interval of the mode switching decision corresponding to the driver style is obtained according to the driving risk value during mode switching; the future trajectories of the self-vehicle and surrounding vehicles are obtained based on the trajectory prediction model of the environmental attention network EA-Net, and the future driving risk change curve is obtained using the driving risk assessment method based on the risk field model, and the driving risk value of the collision position is used as the right interval of the mode switching decision;

[0009] By combining the future risk curve and the mode switching decision risk interval, the time interval for the driver to switch driving modes is obtained. The human driving mode speed prediction model based on the LSTM recurrent neural network is used to obtain the human driving mode speed change prediction sequence. According to the speed prediction and mode switching time interval constraints, a mode switching optimization model based on extreme energy consumption is established to solve the optimal switching timing.

[0010] Furthermore, the different driving scenarios include cruising driving scenario, following vehicle driving scenario, intersection passing scenario and intersection queuing scenario;

[0011] The driving characteristic data of the driver under different driving scenarios include accelerator pedal opening, steering wheel angle, steering wheel speed, vehicle longitudinal speed, lateral speed, longitudinal acceleration, lateral acceleration, vehicle roll angle and heading angle. The average, maximum and variance of the absolute values ​​of the driving characteristic data of the driver under different driving scenarios are used as driver style characteristic indicators.

[0012] Furthermore, the driver classification using a clustering algorithm includes:

[0013] Based on the driver style characteristic index, factor analysis was used to reduce the dimension and two factors were obtained, namely the horizontal factor and the vertical factor.

[0014] Based on the lateral factors and longitudinal factors, a clustering algorithm is used to classify drivers, and the resulting categories include three types, namely, radical type, ordinary type, and conservative type.

[0015] Furthermore, the driver style recognition model based on deep learning recognizes the driver style and obtains the mode switching process driving data of the current driver style category, including:

[0016] A learning sample data set is constructed based on classification features and category data, and a driver style recognition model based on deep learning is established. Based on the driver style recognition model, the processed driver feature index data is used as input to obtain the driver style recognition result. The calculation process is expressed as:

[0017]

[0018] Where X is the input vector containing the driver characteristic index data, X = {x1, x2}, x1 represents the driver characteristic lateral factor, x2 represents the driver characteristic longitudinal factor, W i and b i is the weight matrix and bias of the i-th layer of the multilayer perceptron, n is the number of layers of the multilayer perceptron, and Y is the driver style category output;

[0019] According to the clustering results, the driving data of the three driving styles in four driving scenarios are analyzed, including the position coordinates of the vehicle, the speed of the vehicle, the position coordinates of the surrounding vehicle, and the speed of the surrounding vehicle. The driving data set N of the mode switching process of the current driver style category is obtained by using the driver style recognition results. a , expressed as;

[0020]

[0021] in The driving data of the mode switching process of the j-th driver identified as having a driving style of type a, is the driving data of the i-th environment car, are the horizontal coordinate, vertical coordinate and speed of the environment vehicle i, respectively. I is the serial number set of the environment vehicles around the vehicle. is the driving data of the jth driver, They are respectively the horizontal coordinate, vertical coordinate and speed of the vehicle j.

[0022] Furthermore, the driving data of the mode switching process according to the current driver style category and the driving risk value when the mode is switched are calculated based on the risk field model, including:

[0023] Based on the influence of Z surrounding vehicles covering all directions around the vehicle on the vehicle, the field strength E of the driving risk field generated by the surrounding moving vehicles at the vehicle position is calculated. j , the calculation formula is:

[0024]

[0025] Among them, E V_ij is the surrounding vehicle i(x i ,y i ) in the vehicle j(x j ,yj )The risk field strength generated, (x i ,y i ), (x j ,y j ) is the coordinate of the center of mass of the vehicle, k1, k2, G are constants greater than 0, R i is the impact factor of road conditions at the surrounding vehicle i, M i is the equivalent mass of the surrounding vehicle i, r ij =(x j -x i ,y j -y i ), represents the distance vector between the surrounding vehicle i and the vehicle j, and its direction is the same as the field strength direction, v i is the speed of the surrounding vehicle i, θ i is the speed direction and distance vector r of the surrounding vehicle i ij The angle between the two, exp is an exponential function with the natural constant e as the base;

[0026] The force calculation formula in the field strength is used to calculate the force on the vehicle, and the force on the vehicle is used as the driving risk risk of the vehicle. j ; Wherein, the force calculation formula in the field strength is expressed as:

[0027] risk j =E j M j R j exp[-k2v j cos(θ j )](1+Dr j ) (4)

[0028] Among them, M j is the equivalent mass of the vehicle, R j is the influencing factor of the road conditions at the vehicle, v j is the vehicle speed, θ j is the direction of the vehicle's speed and the field strength E j The angle between the directions, D rj is the driver risk factor of the vehicle.

[0029] Furthermore, the left interval of the mode switching decision corresponding to the driver's style is obtained according to the driving risk value at the time of mode switching, including:

[0030] Based on the driving data of the mode switching process of the current driver style category, the driving risk assessment method based on the risk field model is used to calculate the driving risk value of each driver in the driving data set when switching the driving mode. The minimum value is taken as the driving risk value when the mode is switched for this type of driving style as the left endpoint of the driving risk interval for the mode switching decision. The calculation formula is:

[0031]

[0032] in, represents the left endpoint of the driving risk interval for the mode switching decision of a class driver, represents the driving risk value of the j-th driver mode switching of type a driver in the data set, and min represents the minimum function.

[0033] Furthermore, the trajectory prediction model based on the environment attention network EA-Net obtains the future trajectories of the vehicle and surrounding vehicles, and uses the driving risk assessment method based on the risk field model to obtain the future driving risk change curve, and uses the driving risk value of the collision position as the right interval of the mode switching decision, including:

[0034] A training data set is constructed based on the historical time domain information of the ego vehicle and surrounding vehicles in the experimental data, including lateral position, longitudinal position, lateral velocity, longitudinal velocity, lateral acceleration, longitudinal acceleration, and heading angle relative to the lane line. The interaction characteristics between vehicles are analyzed, and a trajectory prediction model based on the environmental attention network EA-Net is established. The trajectory prediction model based on the environmental attention network EA-Net is used to predict the trajectories of the ego vehicle and multiple surrounding vehicles in real time. The trajectory prediction input is the past T l The historical time domain information of the vehicle and surrounding vehicles is output as the future T p The predicted trajectory data of the duration; the trajectory prediction model based on the environmental attention network EA-Net is expressed as:

[0035]

[0036] in, Represents the output of the future trajectory of the vehicle to be predicted, T l is the length of the historical time domain, T g is the interval length between the historical time domain and the future time domain, T p For the future prediction time domain length, f (t) ={x (t) ,y (t)}, f (t) is the position coordinate at time t, x (t) is the horizontal coordinate of the position at time t, y (t) is the vertical coordinate of the position at time t; is the input feature matrix, which is constructed through the historical time domain information of the vehicle to be predicted and the surrounding vehicles. A is the space-time correlation matrix, according to W i With W j To establish the interaction effect at a certain moment, Indicates whether there is an interaction between the surrounding vehicles i and the vehicle j at a certain moment. If there is an interaction between the surrounding vehicles i and the vehicle j at a certain moment, then If there is no interaction between the surrounding vehicles i and the vehicle j at any time, then W i Represents the historical time domain information of the vehicle to be predicted and the surrounding vehicles, i=0,1,2,…,n; the characteristic information of the i-th vehicle at time t is expressed as x (t) ,y (t) , θ (t) Respectively represent the lateral position, longitudinal position, lateral velocity, longitudinal velocity, lateral acceleration, longitudinal acceleration and heading angle relative to the lane line at time t;

[0037] The future position coordinates and speeds of the ego vehicle and multiple surrounding vehicles are obtained based on the trajectory prediction results, and the driving risk assessment method based on the risk field model is used to calculate the future T of the ego vehicle. p The driving risk at each moment is used to obtain the future driving risk change curve, and the driving risk value of the collision position between the vehicle and the surrounding vehicles in the future trajectory is used as the right endpoint of the driving risk interval for the mode switching decision.

[0038] Furthermore, according to the combination of the future risk curve and the mode switching decision risk interval, the time interval for the driver to switch the driving mode is obtained, including:

[0039] Driving risk interval based on driver mode switching decision The driver can determine the recommended time interval for switching driving modes based on the future driving risk change curve. in, represents the left endpoint of the driving risk interval for the mode switching decision of a class driver, represents the right endpoint of the driving risk interval for the mode switching decision of a type of driver, Indicates the earliest recommended time to switch driving modes. Indicates the latest recommended time to switch driving modes.

[0040] Furthermore, the human driving mode speed prediction model based on the LSTM recurrent neural network obtains the human driving mode speed change prediction sequence.

[0041] According to the vehicle speed change characteristics before and after the driving mode switch for different driving styles in different driving scenarios, a human driving mode speed prediction model based on LSTM recurrent neural network is established. The model input is driving style and time length T. s The speed time series and driving risk time series of the intelligent driving mode are output with a time length of T e The human driving mode speed change prediction sequence, the human driving mode speed prediction model based on LSTM recurrent neural network is expressed as:

[0042] V hu =w*LSTM(V in ,R,S,h)+b (7)

[0043] in, represents the output predicted speed sequence, Respectively represent the future T e The predicted speed at different times within the time length, is the planned speed sequence of the vehicle in intelligent driving mode, Respectively represent the vehicle history T in the intelligent driving mode s The planning speed at different times within the time length, is the driving risk sequence, Respectively represent the history T s Driving risk at different times within a certain time period, S is the driver style, h is the hidden layer state, w, b are the weights and biases of the fully connected layer, and the time interval T s +T e ≥t f -t0, LSTM represents the long short-term memory network structure.

[0044] Furthermore, the mode switching optimization model based on extreme energy consumption is expressed as:

[0045]

[0046]

[0047] Among them, T is the decision variable switching time, v in (t) is the speed of the intelligent driving mode at time t, v hu (t) is the predicted speed of the human driving mode at time t, t is the time, v is the vehicle speed, v(t) is the vehicle speed at time t, a is the driver mode switching decision category, denergy is the instantaneous energy consumption calculation function obtained by fitting, and J is the energy consumption target value.

[0048] This application has at least the following beneficial effects:

[0049] (1) The present application determines the time interval for the driver's mode switching decision based on the analysis of driver characteristic data and driving risk prediction. This interval combines the earliest switching timing considering the driver's subjective style and the latest switching timing considering driving safety. It can more accurately characterize the time range for the driver to make mode switching decisions in the future while ensuring driving safety, thereby laying the foundation for optimizing the switching timing to provide the driver with more accurate decision-making suggestions.

[0050] (2) This application predicts the speed of the human driving mode based on the time interval of the driver's mode switching decision and the speed after the switching timing. By establishing a switching timing optimization model with energy consumption as the target, the optimal switching timing is obtained, and the driver's style and driving risk are comprehensively considered in the speed prediction, which can more accurately predict the speed changes of the human driving process.

[0051] (3) Compared with the existing mode switching timing recommendations that only consider driving safety, the present application comprehensively considers the balance between safety and energy consumption, and provides the driver with the optimal switching timing recommendations to achieve the lowest energy consumption while ensuring driving safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 A flowchart of a method for switching between intelligent driving and human driving modes for an urban logistics vehicle taking into account driving risks and energy consumption, provided in an embodiment of the present application.

[0053] Figure 2 A schematic diagram of a complex urban driving scenario provided in an embodiment of the present application.

[0054] Figure 3 A schematic diagram of the positions of eight surrounding vehicles around the vehicle provided in an embodiment of the present application.

[0055] Figure 4 A schematic diagram of the driving mode switching decision time interval generation provided in an embodiment of the present application.

[0056] Figure 5 A schematic diagram of human driving mode speed prediction after mode switching timing provided in an embodiment of the present application. DETAILED DESCRIPTION

[0057] The following describes the embodiments of the present application through specific examples, and those skilled in the art can easily understand other advantages and effects of the present application from the contents disclosed in this specification. The present application can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present application. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict.

[0058] The specific implementation of the present application is further described in detail below in conjunction with the drawings and examples.

[0059] The embodiment of the present application provides a method for switching between intelligent driving and human driving modes of urban logistics vehicles taking into account driving risks and energy consumption, such as Figure 1 As shown, it is a flow chart of the method. The method for switching between intelligent driving and human driving modes of urban logistics vehicles taking into account driving risks and energy consumption includes the following steps S1 to S3, which are described in detail as follows.

[0060] S1: Feature data extraction of mode switching process considering driver style analysis.

[0061] In this embodiment, the driving characteristic data of the driver in different driving scenarios is analyzed, the driver is classified using a clustering algorithm, the driver style is identified based on a deep learning driver style recognition model, and the mode switching process driving data of the current driver style category is obtained accordingly.

[0062] In some embodiments, different driving scenarios include cruising driving scenarios, following vehicle driving scenarios, intersection passing scenarios and intersection queuing scenarios; the driving characteristic data of the driver in different driving scenarios include accelerator pedal opening, steering wheel angle, steering wheel speed, vehicle longitudinal speed, lateral speed, longitudinal acceleration, lateral acceleration, vehicle roll angle and heading angle, and the average, maximum value and variance of the absolute values ​​of these data are calculated as driver style characteristic indicators.

[0063] like Figure 2 As shown in FIG. 1 , a schematic diagram of a complex urban driving scenario provided by an embodiment of the present application is provided. The method provided by the present embodiment can be applied to the following examples: Figure 2 In the scene shown. Figure 2 Four scenarios are shown, namely, a cruising driving scenario, a vehicle-following driving scenario, an intersection queuing scenario, and an intersection passing scenario.

[0064] In some embodiments, a clustering algorithm is used to classify drivers. Specifically, the obtained driver style characteristic indicators are reduced in dimension using factor analysis to eliminate the correlation between the characteristics, and two factors (a horizontal factor and a vertical factor) are obtained. The driver is then classified into three categories using a clustering algorithm based on the reduced dimension characteristic factor data: aggressive, ordinary, and conservative.

[0065] In some embodiments, the driver style recognition model based on deep learning recognizes the driver style, including:

[0066] A learning sample data set is constructed based on classification features and category data, and a driver style recognition model based on deep learning is established. The input is the horizontal factor and the vertical factor, and the output is the category, which is expressed as formula (1). Based on the driver style recognition model, the processed driver feature index data is taken as input, and the driver style recognition result is obtained as category a;

[0067]

[0068] Where X is the input vector containing the driver characteristic index data, X = {x1, x2}, x1 represents the driver characteristic lateral factor, x2 represents the driver characteristic longitudinal factor, W i and b i is the weight matrix and bias of the i-th layer of the multilayer perceptron, n is the number of layers of the multilayer perceptron, and Y is the driver style category output.

[0069] In some embodiments, the mode switching process driving data of the current driver style category is obtained by the following method:

[0070] According to the clustering results, the driving data of the three driving styles in four driving scenarios are analyzed, including the position coordinates of the vehicle, the speed of the vehicle, the position coordinates of the surrounding vehicle, and the speed of the surrounding vehicle. The driving data set N of the mode switching process of the current driver style category is obtained by using the driver style recognition results. a (Formula (2)).

[0071]

[0072] in The driving data of the mode switching process of the j-th driver identified as having a driving style of type a, is the driving data of the i-th environment car, are the horizontal coordinate, vertical coordinate and speed of the environment vehicle i, respectively. I is the serial number set of the environment vehicles around the vehicle. is the driving data of the jth driver, They are respectively the horizontal coordinate, vertical coordinate and speed of the vehicle j.

[0073] S2: Determination of the mode switching decision interval considering the driver's subjective style and the surrounding vehicle trajectory prediction.

[0074] In this embodiment, according to the driving data of the mode switching process of the current driver style category, the driving risk value during mode switching is calculated based on the risk field model to obtain the left interval of the mode switching decision corresponding to the driver style; the future trajectories of the vehicle and surrounding vehicles are obtained based on the environmental attention network EA-Net trajectory prediction model, and the future driving risk change curve is obtained using the driving risk assessment method based on the risk field model, and the driving risk value of the collision position is used as the right interval of the mode switching decision; in this way, the mode switching decision interval considering the driver's subjective style and the surrounding vehicle trajectory prediction is determined.

[0075] In some embodiments, Figure 3 As shown, it is a schematic diagram of the positions of eight environmental vehicles around the vehicle provided in an embodiment of the present application. Figure 3 In the figure, EV represents the ego vehicle, LF, L, LE, Re, RR, R and RF represent eight surrounding vehicles in all directions (eight directions) around the ego vehicle. The driving risk value during mode switching is calculated based on the risk field model according to the driving data of the mode switching process of the current driver style category, including:

[0076] Considering the impact of 8 surrounding vehicles covering all directions around the vehicle on the vehicle, calculate the field strength E of the driving risk field generated by the surrounding moving vehicles at the vehicle position j (Formula (3), where Z = 8), and then use the force calculation formula in the field strength to calculate the force on the vehicle, that is, the driving risk risk of the vehicle j (Formula (4));

[0077]

[0078] Among them, E V_ij is the surrounding vehicle i(x i ,y i ) in the vehicle j(x j ,y j ) generated by the risk field strength, (x i ,y i ), (x j ,y j ) is the coordinate of the center of mass of the vehicle, k1, k2, G are constants greater than 0, R i is the impact factor of road conditions at the surrounding vehicle i, M i is the equivalent mass of the surrounding vehicle i, r ij =(x j -x i ,y j -y i ), represents the distance vector between the surrounding vehicle i and the vehicle j, and its direction is the same as the field strength direction, v i is the speed of the surrounding vehicle i, θ iis the speed direction and distance vector r of the surrounding vehicle i ij The angle between the two angles is , and exp is an exponential function with the natural constant e as the base.

[0079] risk j =E j M j R j exp[-k2v j cos(θ j )](1+Dr j ) (4)

[0080] Among them, M j is the equivalent mass of the vehicle, R j is the influencing factor of the road conditions at the vehicle, v j is the vehicle speed, θ j is the direction of the vehicle's speed and the field strength E j The angle between the directions, D rj is the driver risk factor of the vehicle.

[0081] In some embodiments, the mode switching decision left interval corresponding to the driver's style is obtained by the following method:

[0082] For the driving data of the mode switching process of the current driver style category, the driving risk assessment method based on the risk field model is used to calculate the driving risk value of each driver in the driving data set when switching the driving mode. The minimum value is taken as the driving risk value of the mode switching of this type of driving style (Formula (5)), which is the left endpoint of the driving risk interval of the mode switching decision.

[0083]

[0084] in, represents the left endpoint of the driving risk interval for the mode switching decision of a class driver, represents the driving risk value of the j-th driver mode switching of type a driver in the data set, and min represents the minimum function.

[0085] In some embodiments, the specific implementation process of obtaining the future trajectories of the self-vehicle and surrounding vehicles based on the environmental attention network EA-Net trajectory prediction model is as follows:

[0086] The training data set is constructed based on the historical time domain information of the ego vehicle and surrounding vehicles in the experimental data, including lateral position, longitudinal position, lateral velocity, longitudinal velocity, lateral acceleration, longitudinal acceleration and heading angle relative to the lane line. The interaction characteristics between vehicles are analyzed and a trajectory prediction model based on the environmental attention network EA-Net is established (Equation (6)). The trajectory prediction model is used to predict the trajectories of the ego vehicle and the surrounding 8 vehicles in real time. The trajectory prediction input is the past Tl The historical time domain information of the vehicle and surrounding vehicles is output as the future T p Prediction trajectory data of duration;

[0087]

[0088] in, Represents the output of the future trajectory of the vehicle to be predicted, T l is the length of the historical time domain, T g is the interval length between the historical time domain and the future time domain, T p For the future prediction time domain length, f (t) ={x (t) ,y (t)}, f (t) is the position coordinate at time t, x (t) is the horizontal coordinate of the position at time t, y (t) is the vertical coordinate of the position at time t; is the input feature matrix, which is constructed through the historical time domain information of the vehicle to be predicted and the surrounding vehicles. A is the space-time correlation matrix, according to W i With W j To establish the interaction effect at a certain moment, Indicates whether there is an interaction between the surrounding vehicles i and the vehicle j at a certain moment. If there is an interaction between the surrounding vehicles i and the vehicle j at a certain moment, then If there is no interaction between the surrounding vehicles i and the vehicle j at any time, then W i Represents the historical time domain information of the vehicle to be predicted and the surrounding vehicles, i=0,1,2,…,n; the characteristic information of the i-th vehicle at time t is expressed as x (t) ,y (t) , θ (t) They represent the lateral position, longitudinal position, lateral velocity, longitudinal velocity, lateral acceleration, longitudinal acceleration and heading angle relative to the lane line at time t respectively.

[0089] In some embodiments, the mode switching decision right interval is determined by: obtaining the future position coordinates and speeds of the ego vehicle and the eight surrounding vehicles according to the trajectory prediction results, and calculating the future T of the ego vehicle using a driving risk assessment method based on a risk field model. p The driving risk at each moment is used to obtain the future driving risk change curve, and the driving risk value of the collision position between the self-vehicle and the surrounding vehicle in the future trajectory is used as the right end point of the driving risk interval for the mode switching decision.

[0090] S3: Optimization of mode switching decision timing based on driving risk and extreme energy consumption.

[0091] In this embodiment, the time interval for the driver to switch driving modes is obtained by combining the future risk curve and the mode switching decision risk interval. The human driving mode speed prediction model based on the LSTM recurrent neural network is used to obtain the human driving mode speed change prediction sequence. According to the speed prediction and mode switching time interval constraints, a mode switching optimization model based on extreme energy consumption is established to solve the optimal switching timing.

[0092] In some embodiments, Figure 4 As shown, it is a schematic diagram of generating a driving mode switching decision time interval provided by an embodiment of the present application. The time interval for the driver to switch the driving mode is obtained by the following method:

[0093] Driving risk interval based on driver mode switching decision The driver can determine the recommended time interval for switching driving modes based on the future driving risk change curve. in, represents the left endpoint of the driving risk interval for the mode switching decision of a class driver, represents the right endpoint of the driving risk interval for the mode switching decision of a type of driver, Indicates the earliest recommended time to switch driving modes. Indicates the latest recommended time to switch driving modes.

[0094] In some embodiments, Figure 5 As shown, it is a schematic diagram of the human driving mode speed prediction after the mode switching opportunity provided by the embodiment of the present application. The human driving mode speed change prediction sequence is obtained by the human driving mode speed prediction model based on the LSTM recurrent neural network, specifically including:

[0095] By analyzing the vehicle speed change characteristics before and after driving mode switching for different driving styles in different driving scenarios, a human driving mode speed prediction model based on LSTM recurrent neural network is established (Formula (7)). The model input is driving style and time length T s The speed time series and driving risk time series of the intelligent driving mode are output with a time length of T e The speed change prediction sequence of human driving mode;

[0096] V hu =w*LSTM(V in ,R,S,h)+b (7)

[0097] in, represents the output predicted speed sequence, Respectively represent the future T e The predicted speed at different times within the time length, is the planned speed sequence of the vehicle in intelligent driving mode, Respectively represent the vehicle history T in the intelligent driving mode s The planning speed at different times within the time length, is the driving risk sequence, Respectively represent the history T s Driving risk at different times within a certain time period, S is the driver style, h is the hidden layer state, w, b are the weights and biases of the fully connected layer, and the time interval T s +T e ≥t f -t0, LSTM represents the long short-term memory network structure.

[0098] In some embodiments, a mode switching optimization model based on extreme energy consumption is established according to speed prediction, mode switching time interval constraint and the energy consumption calculation model obtained by fitting (Formula (8)), and the model is solved to determine the optimal switching time T for the driver. opt .

[0099]

[0100]

[0101] Among them, T is the decision variable switching time, v in (t) is the speed of the intelligent driving mode at time t, v hu (t) is the predicted speed of the human driving mode at time t, t is the time, v is the vehicle speed, v(t) is the vehicle speed at time t, a is the driver mode switching decision category, denergy is the instantaneous energy consumption calculation function obtained by fitting, and J is the energy consumption target value.

[0102] The above implementation modes are only used to illustrate the present application, and are not intended to limit the present application. Ordinary technicians in the relevant technical field may make various changes and modifications without departing from the spirit and scope of the present application. Therefore, all equivalent technical solutions also belong to the scope of the present application, and the scope of patent protection of the present application shall be limited by the claims.

Claims

1. A method for switching between intelligent driving and human driving modes of urban logistics vehicles considering driving risks and energy consumption, characterized in that: The method comprises: According to the driving characteristic data of drivers in different driving scenarios, clustering algorithms are used to classify drivers, and driver style recognition models based on deep learning are used to identify driver style, and the mode switching process driving data of the current driver style category is obtained; According to the driving data of the mode switching process of the current driver style category, the driving risk value during the mode switching is calculated based on the risk field model, and the left interval of the mode switching decision corresponding to the driver style is obtained according to the driving risk value during the mode switching; Based on the trajectory prediction model of the environmental attention network EA-Net, the future trajectories of the self-vehicle and surrounding vehicles are obtained. The driving risk change curve of the future driving risk is obtained by using the driving risk assessment method based on the risk field model. The driving risk value of the collision position is used as the right interval of the mode switching decision. By combining the future risk curve and the mode switching decision risk interval, the time interval for the driver to switch driving modes is obtained. The human driving mode speed prediction model based on the LSTM recurrent neural network is used to obtain the human driving mode speed change prediction sequence. According to the speed prediction and mode switching time interval constraints, a mode switching optimization model based on extreme energy consumption is established to solve the optimal switching timing.

2. The method for switching between intelligent driving and human driving modes of urban logistics vehicles considering driving risks and energy consumption as claimed in claim 1, characterized in that: The different driving scenarios include cruising driving scenario, following vehicle driving scenario, intersection passing scenario and intersection queuing scenario; The driving characteristic data of the driver under different driving scenarios include accelerator pedal opening, steering wheel angle, steering wheel speed, vehicle longitudinal speed, lateral speed, longitudinal acceleration, lateral acceleration, vehicle roll angle and heading angle. The average, maximum and variance of the absolute values ​​of the driving characteristic data of the driver under different driving scenarios are used as driver style characteristic indicators.

3. The method for switching between intelligent driving and human driving modes of urban logistics vehicles considering driving risks and energy consumption as claimed in claim 2, characterized in that: The method of using a clustering algorithm to classify drivers includes: Based on the driver style characteristic index, factor analysis was used to reduce the dimension and two factors were obtained, namely the horizontal factor and the vertical factor. Based on the lateral factors and longitudinal factors, a clustering algorithm is used to classify drivers, and the resulting categories include three types, namely, radical type, ordinary type, and conservative type.

4. The method for switching between intelligent driving and human driving modes of urban logistics vehicles considering driving risks and energy consumption as claimed in claim 3 is characterized in that: The driver style recognition model based on deep learning recognizes the driver style and obtains the mode switching process driving data of the current driver style category, including: A learning sample data set is constructed based on classification features and category data, and a driver style recognition model based on deep learning is established. Based on the driver style recognition model, the processed driver feature index data is used as input to obtain the driver style recognition result. The recognition model is expressed as: Where X is the input vector containing the driver characteristic index data, X = {x1, x2}, x1 represents the driver characteristic lateral factor, x2 represents the driver characteristic longitudinal factor, W i and b i is the weight matrix and bias of the i-th layer of the multilayer perceptron, n is the number of layers of the multilayer perceptron, and Y is the driver style category output; According to the clustering results, the driving data of the three driving styles in four driving scenarios are analyzed, including the position coordinates of the vehicle, the speed of the vehicle, the position coordinates of the surrounding vehicle, and the speed of the surrounding vehicle. The driving data set N of the mode switching process of the current driver style category is obtained by using the driver style recognition results. a , expressed as; in The driving data of the mode switching process of the j-th driver identified as having a driving style of type a, is the driving data of the i-th environment car, are the horizontal coordinate, vertical coordinate and speed of the environment vehicle i, respectively. I is the serial number set of the environment vehicles around the vehicle. is the driving data of the jth driver, They are respectively the horizontal coordinate, vertical coordinate and speed of the vehicle j.

5. The method for switching between intelligent driving and human driving modes of urban logistics vehicles considering driving risks and energy consumption as claimed in claim 4, characterized in that: The method of calculating the driving risk value during mode switching based on the risk field model according to the mode switching process driving data of the current driver style category includes: Based on the influence of Z surrounding vehicles covering all directions around the vehicle on the vehicle, the field strength E of the driving risk field generated by the surrounding moving vehicles at the vehicle position is calculated. j , the calculation formula is: Among them, E V_ij is the surrounding vehicle i(x i ,y i ) in the vehicle j(x j ,y j ) generated by the risk field strength, (x i ,y i ), (x j ,y j ) is the coordinate of the center of mass of the vehicle, k1, k2, G are constants greater than 0, R i is the impact factor of road conditions at the surrounding vehicle i, M i is the equivalent mass of the surrounding vehicle i, r ij =(x j -x i ,y j -y i ), represents the distance vector between the surrounding vehicle i and the vehicle j, and its direction is the same as the field strength direction, v i is the speed of the surrounding vehicle i, θ i is the speed direction and distance vector r of the surrounding vehicle i ij The angle between the two angles, exp is an exponential function with the natural constant e as the base; The force calculation formula in the field strength is used to calculate the force on the vehicle, and the force on the vehicle is used as the driving risk risk of the vehicle. j ; Wherein, the force calculation formula in the field strength is expressed as: risk j =E j M j R j exp[-k2v j cos(θ j )]1+Dr j ) (4) Among them, M j is the equivalent mass of the vehicle, R j is the influencing factor of the road conditions at the vehicle, v j is the vehicle speed, θ j is the direction of the vehicle's speed and the field strength E j The angle between the directions, D rj is the driver risk factor of the vehicle.

6. The method for switching between intelligent driving and human driving modes of urban logistics vehicles considering driving risks and energy consumption as claimed in claim 5, characterized in that: The left interval of the mode switching decision corresponding to the driver's style is obtained according to the driving risk value when the mode is switched, including: Based on the driving data of the mode switching process of the current driver style category, the driving risk assessment method based on the risk field model is used to calculate the driving risk value of each driver in the driving data set when switching the driving mode. The minimum value is taken as the driving risk value when the mode is switched for this type of driving style as the left endpoint of the driving risk interval for the mode switching decision. The calculation formula is: in, represents the left endpoint of the driving risk interval for the mode switching decision of a class driver, represents the driving risk value of the j-th driver mode switching of type a driver in the data set, and min represents the minimum function.

7. The method for switching between intelligent driving and human driving modes of urban logistics vehicles considering driving risks and energy consumption as claimed in claim 6, characterized in that: The trajectory prediction model based on the environmental attention network EA-Net obtains the future trajectories of the self-vehicle and surrounding vehicles, and uses the driving risk assessment method based on the risk field model to obtain the future driving risk change curve, and uses the driving risk value of the collision position as the right interval of the mode switching decision, including: A training data set is constructed based on the historical time domain information of the ego vehicle and surrounding vehicles in the experimental data, including lateral position, longitudinal position, lateral velocity, longitudinal velocity, lateral acceleration, longitudinal acceleration, and heading angle relative to the lane line. The interaction characteristics between vehicles are analyzed, and a trajectory prediction model based on the environmental attention network EA-Net is established. The trajectory prediction model based on the environmental attention network EA-Net is used to predict the trajectories of the ego vehicle and multiple surrounding vehicles in real time. The trajectory prediction input is the past T l The historical time domain information of the vehicle and surrounding vehicles is output as the future T p The predicted trajectory data of the duration; the trajectory prediction model based on the environmental attention network EA-Net is expressed as: in, Represents the output of the future trajectory of the vehicle to be predicted, T l is the length of the historical time domain, T g is the interval length between the historical time domain and the future time domain, T p For the future prediction time domain length, f (t) ={x (t) ,y (t) }, f (t) is the position coordinate at time t, x (t) is the horizontal coordinate of the position at time t, y (t) is the vertical coordinate of the position at time t; is the input feature matrix, which is constructed through the historical time domain information of the vehicle to be predicted and the surrounding vehicles. A is the space-time correlation matrix, according to W i With W j To establish the interaction effect at a certain moment, Indicates whether there is an interaction between the surrounding vehicles i and the vehicle j at a certain moment. If there is an interaction between the surrounding vehicles i and the vehicle j at a certain moment, then If there is no interaction between the surrounding vehicles i and the vehicle j at any time, then W i Represents the historical time domain information of the vehicle to be predicted and the surrounding vehicles, The characteristic information of the i-th vehicle at time t is expressed as x (t) ,y (t) , θ (t) Respectively represent the lateral position, longitudinal position, lateral velocity, longitudinal velocity, lateral acceleration, longitudinal acceleration and heading angle relative to the lane line at time t; The future position coordinates and speeds of the ego vehicle and multiple surrounding vehicles are obtained based on the trajectory prediction results, and the driving risk assessment method based on the risk field model is used to calculate the future T of the ego vehicle. p The driving risk at each moment is used to obtain the future driving risk change curve, and the driving risk value of the collision position between the vehicle and the surrounding vehicles in the future trajectory is used as the right endpoint of the driving risk interval for the mode switching decision.

8. The method for switching between intelligent driving and human driving modes of urban logistics vehicles considering driving risks and energy consumption as claimed in claim 1, characterized in that: According to the future risk curve and the mode switching decision risk interval, the time interval for the driver to switch driving mode is obtained, including: Driving risk interval based on driver mode switching decision The driver can determine the recommended time interval for switching driving modes based on the future driving risk change curve. in, represents the left endpoint of the driving risk interval for the mode switching decision of a class driver, represents the right endpoint of the driving risk interval for the mode switching decision of a type of driver, Indicates the earliest recommended time to switch driving modes. Indicates the latest recommended time to switch driving modes.

9. The method for switching between intelligent driving and human driving modes of urban logistics vehicles considering driving risks and energy consumption as claimed in claim 8, characterized in that: The human driving mode speed prediction model based on the LSTM recurrent neural network obtains the human driving mode speed change prediction sequence, including: According to the vehicle speed change characteristics before and after the driving mode switch for different driving styles in different driving scenarios, a human driving mode speed prediction model based on LSTM recurrent neural network is established. The model input is driving style and time length T. s The speed time series and driving risk time series of the intelligent driving mode are output with a time length of T e The human driving mode speed change prediction sequence, the human driving mode speed prediction model based on LSTM recurrent neural network is expressed as: V hu =w*LSTM(V in ,R,S,h)+b (7) in, represents the output predicted speed sequence, Respectively represent the future T e The predicted speed at different times within the time length, is the planned speed sequence of the vehicle in intelligent driving mode, Respectively represent the vehicle history T in the intelligent driving mode s The planning speed at different times within the time length, is the driving risk sequence, Respectively represent the history T s Driving risk at different times within a certain time period, S is the driver style, h is the hidden layer state, w, b are the weights and biases of the fully connected layer, and the time interval T s +T e ≥t f -t0, LSTM represents the long short-term memory network structure.

10. The method for switching between intelligent driving and human driving modes of urban logistics vehicles considering driving risks and energy consumption as claimed in claim 9, characterized in that: The mode switching optimization model based on extreme energy consumption is expressed as: Among them, T is the decision variable switching time, v in (t) is the speed of the intelligent driving mode at time t, v hu (t) is the predicted speed of the human driving mode at time t, t is the time, v is the vehicle speed, v(t) is the vehicle speed at time t, a is the driver mode switching decision category, denergy is the instantaneous energy consumption calculation function obtained by fitting, and J is the energy consumption target value.

Images