Driving state recognition method and system

Abstract

The present application relates to the field of automobile automatic identification technology, and specifically to a driving state identification method and system. The method includes the following steps: determining the driving state of the vehicle in the current time period; in the current time period, when it is determined that the driving state of the vehicle is a state requiring driving identification, calculating the approximate entropy of the steering wheel angle in the current time period based on the steering wheel angle of the vehicle; sending the approximate entropy of the steering wheel angle in the current time period to a long short-term memory network LSTM model; the long short-term memory network LSTM model is used to predict whether the driving state in the subsequent time period is fatigued based on the approximate entropy of the steering wheel angle in the current time period. The system is used to execute the above method. The driving state identification method and system provided in the present application can solve the problem that the recognition accuracy in the related technology is not high and the ideal effect cannot be achieved.

Description

Technical Field

[0001] The present application relates to the technical field of automatic vehicle identification, and in particular to a driving status identification method and system. Background Art

[0002] With the deepening of the "new four modernizations" of automobiles, customers' demands for automobile intelligence and driving safety are also constantly increasing, and "driver fatigue driving behavior recognition" is gradually becoming an essential function to ensure driver driving safety.

[0003] The current driver fatigue recognition functions are mainly divided into three types:

[0004] One approach involves installing a camera in the cockpit to capture image data of the driver while driving, then using a convolutional neural network model to determine driver fatigue. This approach offers the advantage of providing direct indicators of driver fatigue, such as blinking, yawning, and wandering eyes, allowing for a more accurate assessment of driver fatigue. However, this approach requires the installation of additional camera hardware and communication devices, increasing vehicle hardware costs. Furthermore, the development of facial recognition algorithms significantly increases vehicle R&D costs. Furthermore, with the implementation of new data security laws and regulations, the collection of sensitive driver data carries certain legal risks.

[0005] Another method is to use invasive devices, such as head-mounted brainwave or bioelectric collectors to identify the driver's mental state. The advantage of this method is high accuracy, and the detection system is decoupled from the vehicle. The disadvantage is that the use of invasive devices may affect the driver's driving experience.

[0006] The remaining method is to determine whether the driver is in a fatigue state based on vehicle driving data. For example, Bosch believes that the driver's fatigue state can be determined based on the steering angle sensor signal and its related recognition algorithm. Summary of the Invention

[0007] The present application provides a driving status recognition method and system, which can solve the problem in related technologies of low recognition accuracy and inability to achieve ideal results.

[0008] In order to solve the technical problems described in the background technology, the first aspect of the present application provides a driving state recognition method, which is executed by a vehicle-side controller and includes the following steps:

[0009] Determine the driving status of the vehicle in the current time period;

[0010] In a current time period, when it is determined that the driving state of the vehicle is a state requiring driving identification, calculating an approximate entropy of the steering wheel angle in the current time period based on the steering wheel angle of the vehicle;

[0011] The approximate entropy of the steering wheel angle in the current time period is sent to a long short-term memory network LSTM model; the long short-term memory network LSTM model is used to predict whether the driving state in a subsequent time period is fatigued based on the approximate entropy of the steering wheel angle in the current time period.

[0012] Optionally, the step of sending the steering wheel angle approximate entropy in the current time period to a long short-term memory network (LSTM) model includes:

[0013] Sending the approximate entropy of the steering wheel angle in the current time period to a long short-term memory network (LSTM) model stored in the cloud via wireless communication;

[0014] The long short-term memory network (LSTM) model is obtained by pre-training the sample set on the cloud.

[0015] Optionally, the step of sending the steering wheel angle approximate entropy in the current time period to a long short-term memory network (LSTM) model includes:

[0016] The approximate entropy of the steering wheel angle in the current time period is sent to the long short-term memory network LSTM model stored in the vehicle-side controller;

[0017] The long short-term memory network LSTM model is pre-trained on the sample set by the cloud and sent to the vehicle-side controller via wireless communication.

[0018] Optionally, the step of determining the driving state of the vehicle in the current time period includes:

[0019] Get the vehicle speed in the current time period;

[0020] When the vehicle speed is greater than the vehicle speed threshold in the current time period, it is determined that the driving state of the vehicle is a driving identification-required state in the current time period.

[0021] Optionally, the step of determining the driving state of the vehicle in the current time period includes:

[0022] Get the vehicle speed in the current time period;

[0023] When the vehicle speed is less than the vehicle speed threshold in the current time period, it is determined that the driving state of the vehicle is a state that does not require driving identification in the current time period;

[0024] Stop calculating the steering wheel angle approximate entropy in the current time period.

[0025] In order to solve the technical problems described in the background technology, the second aspect of the present application provides a driving state recognition method, which is executed by the cloud and includes the following steps:

[0026] Obtaining an approximate entropy of a steering wheel angle in a current time period as described in the first aspect of the present application;

[0027] Inputting the steering wheel angle approximate entropy in the first time period into a long short-term memory network (LSTM) model to predict whether the driver is fatigued in the next time period;

[0028] The predicted result of whether the driving state in the subsequent time period is fatigued is sent to the vehicle-side controller.

[0029] Optionally, the driving state recognition method further includes the step of pre-training the sample set in the cloud to obtain the long short-term memory network LSTM model.

[0030] Optionally, the driving state recognition method further includes the steps of:

[0031] The cloud sends the long short-term memory network LSTM model to the vehicle-side controller via wireless communication.

[0032] In order to solve the technical problems described in the background technology, the third aspect of this application provides a driving state recognition system, which includes a vehicle-side controller for executing the driving state recognition method described in the first aspect of this application, and a cloud-end for executing the driving state recognition method described in the second aspect of this application.

[0033] Optionally, the vehicle-side controller includes:

[0034] a calculation unit configured to calculate an approximate entropy of the steering wheel angle in the current time period based on the steering wheel angle when the vehicle speed in the current time period is greater than a vehicle speed threshold;

[0035] The communication unit is used by the vehicle-side controller to establish a wireless communication connection with the cloud.

[0036] Optionally, the calculation unit includes an approximate entropy calculation module and a driving state calculation module;

[0037] The approximate entropy calculation module is configured to calculate an approximate entropy of the steering wheel angle in the current time period based on the steering wheel angle when the vehicle speed in the current time period is greater than a vehicle speed threshold;

[0038] The driving state calculation module stores a long short-term memory network LSTM model.

[0039] The technical solution of the present application includes at least the following advantages: the present application sends the approximate entropy of the steering wheel angle in the current time period to the long short-term memory network LSTM model through the vehicle-mounted controller. The long short-term memory network LSTM model is obtained by pre-training the sample set and stored in the cloud. The LSTM model is then used to predict whether the driving status in the subsequent time period is fatigued based on the approximate entropy of the steering wheel angle in the current time period, thereby facilitating management and updating, and there is no need to install additional camera equipment in the vehicle, saving the vehicle's hardware cost and the R&D cost of related algorithms, and can effectively identify the driver's fatigue status, thereby improving the cost-effectiveness of the solution. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the specific implementation methods of the present application or the technical solutions in the prior art, the following is a brief introduction to the drawings required for use in the specific implementation methods or the description of the prior art. Obviously, the drawings described below are some implementation methods of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0041] Figure 1 The following is a structural block diagram of a driving state recognition system provided by an embodiment of the present application;

[0042] Figure 2 A flowchart of a driving state identification method provided by an embodiment of the present application is shown;

[0043] Figure 2a shows a curve diagram A(t) of the change of the steering wheel angle in the current time period;

[0044] Figure 3 A structural block diagram of a driving state recognition system provided by another embodiment of the present application is shown;

[0045] Figure 4 A flow chart of a driving status identification method provided by another embodiment of the present application is shown. DETAILED DESCRIPTION

[0046] The following is a clear and complete description of the technical solutions in this application in conjunction with the accompanying drawings. Obviously, the embodiments described are part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making any creative efforts are within the scope of protection of this application.

[0047] In the description of this application, it should be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicating orientations or positional relationships, are based on the orientations or positional relationships shown in the accompanying drawings and are intended solely to facilitate the description of this application and simplify the description. They do not indicate or imply that the devices or components referred to must have a specific orientation, be constructed, or operate in a specific orientation. Therefore, they should not be construed as limitations on this application. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0048] In the description of this application, it should be noted that, unless otherwise expressly specified or limited, the terms "installed," "connected," and "connected" should be understood in a broad sense. For example, they can refer to fixed connections, detachable connections, or integral connections; they can refer to mechanical connections or electrical connections; they can refer to direct connections or indirect connections through an intermediate medium; they can refer to internal connections between two components; they can refer to wireless connections or wired connections. Those skilled in the art will understand the specific meanings of the above terms in this application based on the specific circumstances.

[0049] In addition, the technical features involved in the different embodiments of the present application described below can be combined with each other as long as they do not conflict with each other.

[0050] Figure 1 The structural block diagram of the driving state recognition system provided by an embodiment of the present application is shown. Figure 1 As can be seen in the figure, the driving status recognition system includes a vehicle-side controller 110 and a cloud 120 capable of wireless communication.

[0051] The vehicle-side controller 110 includes a calculation unit 111, which is used to calculate the approximate entropy of the steering wheel angle in the current time period based on the steering wheel angle when the vehicle speed in the current time period is greater than a vehicle speed threshold.

[0052] The vehicle-side controller 110 further includes a communication unit 112 , which is connected to the computing unit 111 . The vehicle-side controller 110 establishes a wireless communication connection with the cloud 120 via the communication unit 112 .

[0053] A long short-term memory network (LSTM) model is pre-trained and stored in the cloud 120 . The LSTM model is used to predict whether the driving state in a subsequent time period is fatigued based on the steering wheel angle approximate entropy in the current time period.

[0054] Figure 2 A flowchart of a driving state recognition method according to an embodiment of the present application is shown. Figure 2The driving state recognition method shown is based on Figure 1 The driving status recognition system shown, combined with Figure 1 and Figure 2 , it can be seen that the driving state recognition method includes the following steps S21 to S23:

[0055] Step S21: Determine the driving state of the vehicle in the current time period.

[0056] Determining the driving state of the vehicle in the current time period may include:

[0057] First, obtain the vehicle speed in the current time period.

[0058] Then, in the current time period, when the vehicle speed is greater than the vehicle speed threshold, it is determined that in the current time period, the driving state of the vehicle is a driving identification-required state.

[0059] Determining the driving state of the vehicle in the current time period may also include:

[0060] First, obtain the vehicle speed in the current time period.

[0061] Then, in the current time period, when the vehicle speed is less than the vehicle speed threshold, it is determined that in the current time period, the driving state of the vehicle is a state that does not require driving identification.

[0062] The calculation of the steering wheel angle approximate entropy in the current time period is stopped.

[0063] Among them, the vehicle speed and steering wheel angle of the vehicle in driving state can be obtained by the vehicle-side controller.

[0064] After obtaining the vehicle's speed in driving state, it is necessary to determine whether the speed is greater than the speed threshold. If it is greater than the speed threshold, it is determined that the driver's driving state needs to be identified to identify the driver's fatigue driving state at higher speeds, so as to facilitate timely reminders and adjustments.

[0065] Step S22: When the vehicle speed in the current time period is greater than the vehicle speed threshold, the vehicle-side controller calculates the approximate entropy of the steering wheel angle in the current time period based on the steering wheel angle.

[0066] Can be achieved through Figure 1 When the vehicle speed in the current time period is greater than the vehicle speed threshold, the calculation unit 111 calculates the steering wheel angle approximate entropy in the current time period based on the steering wheel angle.

[0067] The steering wheel angle approximate entropy is used to measure the degree of chaos of the steering wheel angle, and the degree of chaos of the steering wheel angle can quantify the driving state of the driver.

[0068] If the vehicle speed in the current time period is less than the vehicle speed threshold, the calculation of the steering wheel angle approximate entropy in the current time period is stopped, and the value calculated by the calculation unit 111 is cleared.

[0069] The steering wheel angle approximate entropy in the current time period may be calculated according to the following exemplary embodiments: Figure 2a The steering wheel angle in the current time period is shown in the curve diagram A(t). Figure 2a The horizontal axis is the time variable t, and the vertical axis is the steering wheel angle variable A.

[0070] First, the variation curve A(t) is sampled to obtain a steering wheel angle sequence {A(ti), i=0, 1, 2, 3...n}.

[0071] Then, using m as the selection window and one sequence as the sliding step, the steering wheel angle sequence is divided into k sequence groups, each of which contains m adjacent sequences. The kth sequence group is {A(t(n-m+1))…A(t(n-1)), A(tn)}, and the k-1th sequence group is {A(t(nm))…A(t(n-2)), A(t(n-1))}.

[0072] Then, the distance dxy between each sequence group and all k sequence groups is calculated. Alternatively, a difference operation can be performed between any two sequence groups, with the absolute value of the difference being the distance between the two sequence groups. The sequence group distance dxy between each sequence group and all k sequence groups is obtained as shown in the following table:

[0073] Sequence Group 1 Sequence Group 2 Sequence Group 3 … k-1th sequence group kth sequence group Sequence Group 1 d11 d21 d31 … d(k-1)1 dk1 Sequence Group 2 d12 d22 d32 … d(k-1)2 dk2 Sequence Group 3 d13 d23 d33 … d(k-1)3 dk3 … … … … … … … k-1th sequence group d1(k-1) D2(k-1) D3(k-1) … dk(k-1) kth sequence group d1k d2k d3k … d(k-1)k dkk

[0074] The sequence group distance of each row in the above table represents the sequence group distance between the sequence group and all k sequence groups.

[0075] A distance threshold is then set, and the proportion of all sequence groups in each row of the above table whose distances are greater than the distance threshold is calculated. The proportion of the sequence group distances between the first sequence group and all k sequence groups that are greater than the distance threshold is the first proportion C1, the proportion of the sequence group distances between the second sequence group and all k sequence groups that are greater than the distance threshold is the second proportion C2, and so on. The proportion of the sequence group distances between the kth sequence group and all k sequence groups that are greater than the distance threshold is the kth proportion Ck.

[0076] Then, the logarithmic mean θm of the first ratio C1 to the kth ratio Ck is calculated.

[0077] Repeat the above steps with m+1 as the selection window to obtain the logarithmic mean θm+1. The difference between the logarithmic mean θm and the logarithmic mean θm+1 is calculated as the steering wheel angle approximate entropy.

[0078] Step S23: The vehicle-side controller sends the approximate entropy of the steering wheel angle in the current time period to the long short-term memory network LSTM model stored in the cloud via wireless communication; the long short-term memory network LSTM model is used to predict whether the driving state in the subsequent time period is fatigued based on the approximate entropy of the steering wheel angle in the current time period.

[0079] Among them, the Long Short-Term Memory (LSTM) network is a time recursive neural network. LSTM has been widely used in many application scenarios. The LSTM network can be pre- Figure 1 The cloud 120 is shown as pre-trained.

[0080] The long short-term memory network LSTM model includes an input layer, a hidden layer, an output layer and a network optimization layer, wherein the hidden layer is an LSTM network.

[0081] The main steps for training the LSTM model include the following:

[0082] First, the input layer obtains the steering wheel angle sample set data of different driving states, and then preprocesses, extracts features and reduces the dimension of the sample set data as the input of the hidden layer.

[0083] Then the data of the input hidden layer is used to train the LSTM network.

[0084] Then the long short-term memory network LSTM model is continuously optimized through the network optimization layer.

[0085] Step S24: After the cloud receives the approximate entropy of the steering wheel angle in the current time period sent by the vehicle-side controller, the approximate entropy of the steering wheel angle in the previous time period is input into the long short-term memory network LSTM model, and the LSTM predicts whether the driving state in the subsequent time period is fatigued.

[0086] Step S25: The cloud sends the predicted result of whether the driving state in the subsequent time period is fatigued to the vehicle-side controller.

[0087] Figure 3 The structural block diagram of the driving state recognition system provided by another embodiment of the present application is shown. Figure 1 Based on the embodiment shown, Figure 1 The calculation unit 111 in the embodiment includes an approximate entropy calculation module 1111 and a driving state calculation module 1112. Figure 3 The other structures of the embodiment shown are the same as those of the embodiment shown in the figure and will not be described in detail here.

[0088] Figure 3The approximate entropy calculation module 1111 is used to calculate the approximate entropy of the steering wheel angle in the current time period based on the steering wheel angle when the vehicle speed in the current time period is greater than the vehicle speed threshold.

[0089] The driving state calculation module 1112 stores a long short-term memory network LSTM model.

[0090] The long short-term memory network LSTM model of the driving state calculation module 1112 can be composed of Figure 1 The cloud 120 shown is pre-trained and transmitted to the driving state calculation module 1112 for storage via the communication unit 112 .

[0091] Figure 4 A flowchart of a driving state identification method provided by another embodiment of the present application is shown. Figure 4 The driving state recognition method shown is based on Figure 3 The driving status recognition system shown, combined with Figure 3 and Figure 4 , it can be seen that the driving state recognition method includes the following steps S41 to S45:

[0092] Step S41: The vehicle-side controller obtains the vehicle speed and steering wheel angle when the vehicle is in driving state.

[0093] After obtaining the vehicle's speed in driving state, it is necessary to determine whether the speed is greater than the speed threshold. If it is greater than the speed threshold, it is determined that the driver's driving state needs to be identified to identify the driver's fatigue driving state at higher speeds, so as to facilitate timely reminders and adjustments.

[0094] Step S42: When the vehicle speed in the current time period is greater than the vehicle speed threshold, the steering wheel angle approximate entropy in the current time period is calculated based on the steering wheel angle.

[0095] Can be achieved through Figure 3 When the vehicle speed in the current time period is greater than the vehicle speed threshold, the approximate entropy calculation module 1111 calculates the steering wheel angle approximate entropy in the current time period based on the steering wheel angle.

[0096] The steering wheel angle approximate entropy is used to measure the degree of chaos of the steering wheel angle, and the degree of chaos of the steering wheel angle can quantify the driving state of the driver.

[0097] If the vehicle speed in the current time period is less than the vehicle speed threshold, the calculation of the steering wheel angle approximate entropy in the current time period is stopped, and the value calculated by the approximate entropy calculation module 1111 is cleared.

[0098] Step S43: Send the approximate entropy of the steering wheel angle in the current time period to the long short-term memory network LSTM model stored in the vehicle-side controller.

[0099] The approximate entropy of the steering wheel angle calculated by the approximate entropy calculation module 1111 can be sent to the long short-term memory network LSTM model of the driving state calculation module 1112. The long short-term memory network LSTM model predicts whether the driving state in the subsequent time period is fatigued based on the approximate entropy of the steering wheel angle in the current time period.

[0100] The long short-term memory network LSTM model is pre-trained by the cloud 120 and transmitted to the driving state calculation module 1112 for storage.

[0101] Step S44: After the driving state calculation module 1112 receives the approximate entropy of the steering wheel angle in the current time period sent by the vehicle-side controller, it inputs the approximate entropy of the steering wheel angle in the previous time period into the long short-term memory network LSTM model, and the LSTM predicts whether the driving state in the subsequent time period is fatigued.

[0102] Step S45: The driving status calculation module 1112 in the cloud sends the predicted result of whether the driving status in the subsequent time period is fatigued to the vehicle-side controller.

[0103] It can be seen from the above embodiments that the present application sends the approximate entropy of the steering wheel angle in the current time period to the long short-term memory network LSTM model through the vehicle-mounted controller. The long short-term memory network LSTM model is obtained by pre-training the sample set and stored in the cloud. The LSTM model is then used to predict whether the driving state in the subsequent time period is fatigued based on the approximate entropy of the steering wheel angle in the current time period, which facilitates management and updating, and does not require the installation of additional camera equipment in the car, saving the vehicle's hardware cost and the R&D cost of related algorithms, and can effectively identify the driver's fatigue state, thereby improving the cost-effectiveness of the solution.

[0104] Obviously, the above embodiments are merely examples for clarity of explanation and are not intended to limit the implementation methods. Those skilled in the art will appreciate that other variations or modifications can be made based on the above description. It is not necessary and impossible to enumerate all implementation methods here. Obvious variations or modifications arising therefrom remain within the scope of protection of this application.

Claims

1. A driving state recognition method, characterized in that: The driving state recognition method is executed by a vehicle-side controller and includes the following steps: Determine the driving status of the vehicle in the current time period; When it is determined that the driving state of the vehicle is a state requiring driving identification in the current time period, calculating an approximate entropy of a steering wheel angle in the current time period based on a steering wheel angle of the vehicle; Sending the steering wheel angle approximate entropy in the current time period to a long short-term memory network (LSTM) model; the long short-term memory network (LSTM) model is used to predict whether the driving state in a subsequent time period is fatigued based on the steering wheel angle approximate entropy in the current time period; Calculating the steering wheel angle approximate entropy in the current time period comprises the following steps: Get the steering wheel angle change curve A(t) in the current time period; Sampling the change curve A(t) to obtain a steering wheel angle sequence {A(ti), i=0, 1, 2, 3...n}; Taking m as the selection window and one sequence as the sliding step, the steering wheel angle sequence is divided into k sequence groups; Calculate the distance dxy between each sequence group and all k sequence groups; Set a distance threshold and calculate the proportion Cj of sequence groups whose distances between the j-th sequence group and all k sequence groups are greater than the distance threshold, where j = 1, 2, 3…k; Calculate the logarithmic mean θm of the first ratio C1 to the kth ratio Ck; Repeat the above steps with m+1 as the selection window to obtain the logarithmic mean value θm+1; The difference between the logarithmic mean value θm and the logarithmic mean value θm+1 is calculated as the steering wheel angle approximate entropy.

2. The driving state recognition method according to claim 1, wherein: The step of sending the steering wheel angle approximate entropy in the current time period to the long short-term memory network LSTM model includes: Sending the approximate entropy of the steering wheel angle in the current time period to a long short-term memory network (LSTM) model stored in the cloud via wireless communication; The long short-term memory network (LSTM) model is obtained by pre-training the sample set on the cloud.

3. The driving state recognition method according to claim 1, wherein: The step of sending the steering wheel angle approximate entropy in the current time period to the long short-term memory network LSTM model includes: Sending the steering wheel angle approximate entropy in the current time period to the long short-term memory network LSTM model stored in the vehicle-side controller; The long short-term memory network (LSTM) model is pre-trained on the sample set by the cloud and sent to the vehicle-side controller via wireless communication.

4. The driving state recognition method according to claim 1, wherein: The step of determining the driving state of the vehicle in the current time period includes: Get the vehicle speed in the current time period; When the vehicle speed is greater than the vehicle speed threshold in the current time period, it is determined that the driving state of the vehicle is a driving identification-required state in the current time period.

5. The driving state recognition method according to claim 1, wherein: The step of determining the driving state of the vehicle in the current time period includes: Get the vehicle speed in the current time period; When the vehicle speed is less than the vehicle speed threshold in the current time period, it is determined that the driving state of the vehicle is a state that does not require driving identification in the current time period; Stop calculating the steering wheel angle approximate entropy in the current time period.

6. A driving state recognition method, characterized in that: The driving state recognition method is executed by the cloud, and includes the following steps: Obtaining the steering wheel angle approximate entropy in the current time period as claimed in any one of claims 1 to 5; Inputting the steering wheel angle approximate entropy in the first time period into a long short-term memory network (LSTM) model to predict whether the driver is fatigued in the next time period; The predicted result of whether the driving state in the subsequent time period is fatigued is sent to the vehicle-side controller.

7. The driving state recognition method according to claim 6, wherein: The driving state recognition method also includes the step of pre-training the sample set in the cloud to obtain the long short-term memory network LSTM model.

8. The driving state recognition method according to claim 7, wherein: The driving state recognition method further comprises the steps of: The cloud sends the long short-term memory network LSTM model to the vehicle-side controller via wireless communication.

9. A driving status recognition system, characterized in that: The driving state recognition system includes a vehicle-side controller for executing the driving state recognition method as described in any one of claims 1 to 5, and a cloud-side for executing the driving state recognition method as described in any one of claims 6 to 8.

10. The driving state recognition system according to claim 9, characterized in that: The vehicle-side controller includes: a calculation unit configured to calculate an approximate entropy of the steering wheel angle in the current time period based on the steering wheel angle when the vehicle speed in the current time period is greater than a vehicle speed threshold; The communication unit is used by the vehicle-side controller to establish a wireless communication connection with the cloud.

11. The driving state recognition system according to claim 10, wherein: The calculation unit includes an approximate entropy calculation module and a driving state calculation module; The approximate entropy calculation module is configured to calculate an approximate entropy of a steering wheel angle in the current time period based on the steering wheel angle when the vehicle speed in the current time period is greater than a vehicle speed threshold; The driving state calculation module stores a long short-term memory network LSTM model.

Images