Electric vehicle motor control method and device and medium

Abstract

The invention relates to the technical field of electric vehicle power control, in particular to an electric vehicle motor control method and device and a medium, and the method comprises the steps: obtaining driving behavior data and road state data of a target vehicle; obtaining a driving behavior representation value F representing the aggressive degree of the driving operation of the target vehicle based on the driving behavior data, and obtaining a road state representation value RAI representing the abnormal degree of the driving road of the target vehicle based on the road state data; if F is greater than F0, dynamically adjusting the torque distribution proportion of the motor based on F; and if F is smaller than or equal to F0, the motor torque distribution proportion is dynamically adjusted based on RAI. According to the invention, through scene-divided torque control, accurate adaptation of the driving style and the road working condition is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of electric vehicle power control, in particular to an electric vehicle motor control method, device and medium. BACKGROUND

[0002] With the rapid development of new energy vehicle technology, the electric vehicle motor control performance has become a core factor affecting the vehicle power output, driving comfort and driving safety. As the core power source of electric vehicles, the output control strategy of the motor needs to cope with complex driving conditions, diversified driving behaviors and variable road environments to achieve a balance between power performance, comfort and economy. In the prior art, optimization of motor control has made certain progress. For example, patent CN111845373B discloses a motor control method, which obtains the current speed of the motor and the speed change rate within a preset sampling period, judges whether the speed fluctuates, calculates the compensation torque value and compensates the current output torque in real time if there is fluctuation, thereby suppressing the jitter problem caused by motor torque sudden change and mechanical transmission gap change during acceleration or braking conversion, and improving the driving comfort to a certain extent. This technology focuses on real-time suppression of speed fluctuation and has positive significance in solving the jitter problem in specific conditions.

[0003] However, the above-mentioned prior art still has obvious limitations: first, the control strategy is single, only compensates the torque for the jitter problem caused by motor speed fluctuation, does not consider the differences in driving behavior of drivers (such as different needs for power output in aggressive driving and smooth driving), and cannot dynamically adjust the motor output characteristics according to the driving style; second, it lacks perception and adaptation to road conditions, does not combine the characteristics of complex road scenes such as sudden acceleration, long uphill, high-speed cruising, congestion crawling and curves for targeted control, and it is difficult to optimize power distribution under different road conditions; third, it does not construct a comprehensive quantitative index of driving behavior and road state, and the control logic is only limited to the single goal of suppressing jitter, and cannot balance the power response demand in aggressive driving scenarios, the energy economy in smooth driving scenarios and the driving safety in complex road scenarios.

[0004] At the same time, in other existing motor control schemes, most of them rely on fixed parameters or single condition settings, neither quantitatively analyzing the driving aggressiveness to match the power output, nor realizing the dynamic distribution of motor torque through road abnormal state evaluation, resulting in that in complex driving scenarios and various road conditions, the motor control is difficult to balance the comprehensive needs of power performance, comfort and safety. Therefore, there is an urgent need for a technical solution that can integrate driving behavior perception and road state analysis to realize adaptive adjustment of motor output, to make up for the shortcomings of the prior art. SUMMARY

[0005] In view of the above technical problems, the technical scheme adopted by the present application is as follows:

[0006] According to the first aspect of the present application, a motor control method for an electric vehicle is provided, which comprises the following steps:

[0007] S100, obtaining sensor data of a target vehicle, the sensor data comprising driving behavior data and road state data; wherein the driving behavior data is data reflecting the operation characteristics of a driver, and the road state data is data reflecting the characteristics of a driving scene.

[0008] S200, obtaining a driving behavior representation value F representing the driving operation aggressiveness of the target vehicle based on the driving behavior data.

[0009] S300, obtaining a road state representation value RAI representing the road abnormality degree of the target vehicle based on the road state data.

[0010] S400, comparing the driving behavior representation value F with a preset threshold value F0, if F>F0, then dynamically adjusting the motor torque distribution ratio based on F, and if F≤F0, then dynamically adjusting the motor torque distribution ratio based on RAI.

[0011] According to the second aspect of the present application, an electronic device is provided, comprising a processor and a memory; the processor is used to execute the steps of the method according to the first aspect of the present application by calling the programs or instructions stored in the memory.

[0012] According to the third aspect of the present application, a computer readable storage medium is provided, which stores programs or instructions, and the programs or instructions make the computer execute the steps of the method according to the first aspect of the present application.

[0013] The present application has at least the following beneficial effects:

[0014] (1) Scene-based torque optimization, considering performance and energy efficiency:

[0015] Aggressive scene (F>F0): dynamically increasing the rear axle torque ratio (such as through a hyperbolic tangent function for rapid response), enhancing vehicle control flexibility and power response, and adapting to intense operations such as rapid acceleration, high-speed overtaking, etc.; Stable scene (F≤F0): adjusting torque distribution in combination with road abnormality degree, prioritizing driving stability and energy efficiency, and reducing energy consumption in scenes such as congestion crawling and long uphill driving.

[0016] (2) Multi-dimensional data fusion, improving control accuracy:

[0017] Integrate driving behavior (operation intensity) and road state (scene abnormality) multidimensional data, comprehensively depict actual driving conditions such as "aggressive driving + curve", and "smooth driving + long uphill, make the torque distribution strategy more accurate, and avoid one-sidedness of single dimension judgment.

[0018] (3) Complex working condition self-adaption, safety and experience upgrade:

[0019] Through the coordination of F and RAI, complex scenes such as sudden acceleration, long uphill, congestion, and curve are flexibly responded to, among which, the curve scene corrects the torque distribution through lateral acceleration to optimize the control stability; the long uphill and congestion scenes preferentially guarantee the power smoothness and energy efficiency, which can ensure the final improvement of the vehicle driving safety, driving experience, and energy utilization rate.

[0020] It should be understood that the content described in this part is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become apparent through the following description. BRIEF DESCRIPTION OF DRAWINGS

[0021] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiment description. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without creative labor.

[0022] Figure 1 The flowchart of the electric vehicle motor control method provided by the embodiments of the present application;

[0023] Figure 2 The flowchart for obtaining the driving behavior characteristic value. DETAILED DESCRIPTION

[0024] The technical solutions in the embodiments of the present application will be described clearly and completely in the following with reference to the drawings of the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0025] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used in the description of the application herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. As used herein, the term "and / or" includes any and all combinations of one or more of the associated listed items.

[0026] It is noted that some example embodiments are described as a process or method depicted as a flowchart. Although the process is described in a sequential order, some of the steps can be performed concurrently, in parallel, or simultaneously. Also, the order of the steps can be re-arranged. The process can be terminated when its operations are completed but can also have additional steps not included in the figure. The process can correspond to a method, function, procedure, subroutine, subprogram, etc.

[0027] The embodiment of the present application provides a motor control method for an electric vehicle, which comprises the following steps: Figure 1 As shown in the figure, the method comprises the following steps:

[0028] S100, acquiring sensor data of a target vehicle, wherein the sensor data comprises driving behavior data and road state data.

[0029] In the embodiment of the present application, the target vehicle is an electric vehicle, and specifically can be a dual-motor four-wheel drive vehicle. The sensor data can be continuously collected according to a preset time period (for example, 100 ms to 1 s, which can be dynamically adjusted according to the real-time requirement of data).

[0030] The driving behavior data is data reflecting the operation characteristics of a driver, and can comprise longitudinal / lateral acceleration, yaw angular velocity, accelerator pedal depth, brake pedal depth, frequency of accelerator operation per unit time, frequency of brake operation per unit time, steering wheel turning angle and steering speed. The road state data is data reflecting the characteristics of a driving scene, and can comprise sudden acceleration scene characteristic data, long uphill scene characteristic data, high-speed cruising scene characteristic data, congestion crawling scene characteristic data and curve maintaining scene characteristic data.

[0031] The sensor data is collected in real time by a vehicle-mounted sensor network on the target vehicle. The longitudinal / lateral acceleration can be collected by an inertial measurement unit (IMU), the accelerator pedal depth and brake pedal depth can be collected by a pedal position sensor (installed on the accelerator / brake pedal), the accelerator / brake operation frequency can be collected by a pedal position sensor + timestamp recording method, and the steering wheel angle and steering speed can be collected by a steering angle sensor (installed on the steering column). The sudden acceleration scene feature data is collected by an inertial measurement unit (IMU) + timestamp, and the collection logic is to determine the scene feature parameters (such as average acceleration and total duration) by combining longitudinal acceleration (more than 0.3g), acceleration duration and acceleration frequency per unit time. The long uphill scene feature data can be collected by a slope sensor / GPS + altitude data, and the collection logic is that the slope sensor directly outputs the road slope (%); or the real-time slope is calculated by collecting the altitude change and driving distance through GPS. The high-speed cruising scene feature data is collected by a wheel speed sensor / GPS, and the collection logic is to output the real-time vehicle speed (km / h) and combine the set cruising speed threshold (such as 120km / h) to calculate the speed fluctuation amplitude and sudden deceleration. The congestion crawling scene feature data is collected by a wheel speed sensor + pedal position sensor, and the collection logic is to determine the scene features by vehicle speed (such as less than 10km / h), start-stop frequency (the number of times that the speed changes from 0 to more than 5km / h per unit time) and idling time ratio. The curve maintaining scene feature data is collected by a GPS map + steering angle sensor + camera, and the collection logic is that the GPS map provides the recommended speed of the curve, the steering angle sensor and the camera output the lane deviation distance, and the IMU outputs the lateral acceleration to quantify the curve driving features.

[0032] S200, obtaining a driving behavior representation value F representing the driving operation aggressiveness of the target vehicle based on the driving behavior data.

[0033] S300, obtaining a road state representation value RAI representing the road abnormality degree of the target vehicle based on the road state data.

[0034] S400, comparing the driving behavior representation value F with a preset threshold F0, if F>F0, dynamically adjusting the motor torque distribution ratio based on F, and if F≤F0, dynamically adjusting the motor torque distribution ratio based on RAI.

[0035] Further, in the embodiment of the present application, as shown in Figure 2 F is obtained by the following steps:

[0036] S201, processing the driving behavior data by a first calculation method to obtain a driving behavior representation value F1, and processing the driving behavior data by a second calculation method to obtain a driving behavior representation value F2.

[0037] S202, calculate the absolute value of the difference between F1 and F2, denoted as △F.

[0038] S203, determine the final driving behavior characteristic value F based on the preset driving behavior characteristic value reference table and △F, wherein the reference table contains a plurality of preset difference intervals, each difference interval is mutually exclusive and continuously covers the possible value range of △F, and each difference interval corresponds to a unique value rule.

[0039] Further, the first calculation method satisfies the following conditions: F1 = w1 x AI1 acc +w2 x AI1 pedal +w3 x AI1 steer .

[0040] wherein AI1 acc is a vehicle dynamic response degree characteristic value determined based on a first vehicle dynamic response determination method, w1 is the weight of AI1 acc , AI1 pedal is a pedal operation degree characteristic value determined based on a first pedal operation determination method, w2 is the weight of AI1 pedal , AI1 steer is a steering operation degree characteristic value determined based on a first steering operation determination method, w3 is the weight of AI1 steer , and w1 + w2 + w3 = 1.

[0041] In the embodiment of the application, w1, w2 and w3 can be obtained based on methods such as analytic hierarchy process, grid method or gray relational analysis method. The analytic hierarchy process can include the following steps:

[0042] Step 1: Establish a hierarchical structure

[0043] Target layer: Determine w1, w2 and w3 to reflect the contribution to F1;

[0044] Criteria layer: AI1 acc , AI1 pedal , AI1 steer ;

[0045] Scheme layer: Specific influencing factors of each criteria layer (such as AI1 acc associated acceleration and angular velocity, AI1 pedal associated throttle / brake depth, etc.).

[0046] Step 2: Construct a judgment matrix

[0047] Based on expert experience or real vehicle data, compare the importance of each criteria layer (such as "vehicle dynamic response is more important than pedal operation" recorded as 3, equally important recorded as 1, etc.) to form a judgment matrix.

[0048] Step 3: Consistency check and weight calculation

[0049] Calculate the maximum eigenvalue of the judgment matrix and the corresponding eigenvector, and check the rationality of the matrix through the consistency index (CI) and the average random consistency index (RI) (CR = CI / RI < 0.1 is effective);

[0050] Normalize the eigenvector to get w1, w2, w3, and ensure that w1 + w2 + w3 = 1.

[0051] The grid method optimizes the weight combination through grid search, which can include the following steps:

[0052] Step 1: Set the weight range

[0053] Limit w1, w2, w3 ∈ [0, 1] and w1 + w2 + w3 = 1, and divide the grid according to the fixed step (such as 0.05) to generate all possible weight combinations (such as (0.5, 0.3, 0.2), (0.4, 0.3, 0.3), etc.).

[0054] Step 2: Evaluation index construction

[0055] Take the fitting degree of F1 and the actual driving aggressiveness (such as the "aggressive / stable" label marked by artificial labeling) in the real car test data as the evaluation index (such as the minimum mean square error).

[0056] Step 3: Optimal weight selection

[0057] Iterate through all weight combinations in the grid, and select the combination that makes the evaluation index optimal as w1, w2, w3.

[0058] The gray correlation analysis method determines the weight by analyzing the correlation between each dimension and driving aggressiveness, with the following steps:

[0059] Step 1: Determine the reference sequence and comparison sequence

[0060] Reference sequence: actual driving aggressiveness label (such as the aggressiveness value obtained by expert scoring);

[0061] Comparison sequence: the normalized value sequence of AI1 acc , AI1 pedal , AI1 steer .

[0062] Step 2: Calculate the gray correlation degree

[0063] Calculate the gray correlation coefficient between the comparison sequence and the reference sequence, and quantify the correlation strength between each dimension and the target (the higher the correlation degree, the greater the influence on F1).

[0064] Step 3: Weight normalization

[0065] The gray correlation degrees of each dimension are normalized to obtain w1, w2, and w3, which satisfy w1+w2+w3=1.

[0066] In a preferred embodiment of the application, the grid method is selected to obtain w1, w2, and w3. In a specific embodiment, w1=0.35, w2=0.2, and w3=0.35.

[0067] In the embodiment of the application, AI1 acc The aggressive degree of quantifying the longitudinal and lateral dynamic response of the vehicle is comprehensively considered according to the peak intensity of acceleration / angle speed and the frequency of aggressive operation, and the calculation formula is as follows: AI1 acc = k1 x (|a x | max / a max + |a y | max / a max + |ω z | max / ω max ) / 3 + k2 x min (1, N acc / N max ).

[0068] wherein |a x | max is the maximum value of the absolute value of the longitudinal acceleration in a set time window, with the unit of m / s 2 , and the positive direction is acceleration and the negative direction is deceleration. |a y | max is the maximum value of the absolute value of the lateral acceleration in a set time window, with the unit of m / s 2 , and the left turn is positive, reflecting the centrifugal force of turning. a max is the aggressive threshold of acceleration, which can be 0.6g≈5.88m / s 2 . |ω z | max is the maximum value of the absolute value of the yaw rate in a set time window, with the unit of rad / s, and the counterclockwise is positive, reflecting the rotation rate of the vehicle body. ω max is the aggressive threshold of yaw rate, which can be 0.8 rad / s, and is considered as a sharp turn when exceeding this value. N acc is the total number of times that the longitudinal acceleration, lateral acceleration, or yaw rate exceeds the corresponding threshold in a set time window, that is, the sum of the number of times that the longitudinal acceleration exceeds the normal acceleration threshold, the number of times that the lateral acceleration exceeds the normal turning threshold, and the number of times that the yaw rate exceeds the normal turning rate threshold in a set time window, wherein the normal acceleration threshold can be 0.3g, the normal turning threshold can be 0.2g, and the normal turning rate threshold can be 0.5 rad / s. maxThe maximum allowed number of times of dynamic response dimension aggressive operation, for example, can be 10 times. k1 and k2 are preset coefficients, k1+k2=1, in an illustrative embodiment, k1=0.7, k2=0.3.

[0069] In the embodiment of the application, the length of the set time window can be set based on actual needs, for example, can be several minutes, etc.

[0070] In the embodiment of the application, (|a x | max / a max +|a y | max / a max +|ω z | max / ω max ) is a peak intensity term, which quantifies the intensity of a single most aggressive dynamic response by peak normalization (divided by threshold) and averaging. min(1, N acc / N max ) is a frequency term, which is used to reflect the frequency of dynamic response in a short time by counting the frequency of aggressive events.

[0071] In the embodiment of the application, AI1 pedal is used to quantify the degree of aggressive operation of the accelerator / brake pedal, and the calculation formula is: AI1 pedal =c1×min(1, (N th / N th，max +N br / N br，max ) / 2)+c2×(β th，max / 100+β br，max / 100) / 2+c3×min(1, N switch / N switch，max ).

[0072] Wherein, N th is the number of times that the accelerator depth is greater than the preset accelerator depth threshold in the set time window, and the preset accelerator depth threshold can be 80%. N th，max is the maximum allowed number of times of accelerator aggressive operation, for example, can be 5 times. N br is the number of times that the brake depth is greater than the preset brake depth threshold in the set time window, and the preset brake depth threshold can be 60%. N br，max is the maximum allowed number of times of brake aggressive operation, for example, can be 3 times. β th，max is the maximum depth of the accelerator pedal in the set time window, and the unit is percentage, for example, 100% represents stepping to the bottom. β br，maxN is the maximum depth of the brake pedal in the set time window, in percentage. switch N is the number of times of rapid switching between the accelerator and the brake in the set time window, which can be specifically the number of risky operations (such as switching from the accelerator greater than 80% to the brake greater than 60%) in which the depth of both the accelerator and the brake exceeds the threshold value within 1 second in the window. switch，max N is the maximum allowed number of times of rapid switching between the accelerator and the brake in the set time window, for example, 3 times.c1, c2 and c3 are weight coefficients, c1+c2+c3=1, in an illustrative embodiment, c1=0.4, c2=0.3, and c3=0.3.

[0073] In the embodiment of the application, min(1, (N th / N th，max +N br / N br，max ) / 2) is the aggressive operation frequency term, which is used to reflect the frequent aggressiveness of pedal operation by the normalized accelerator / brake aggressive operation mean.(β th，max / 100+β br，max / 100) / 2 is the pedal depth term, which is used to quantify the intensity of single pedal operation by the maximum depth mean (the greater the depth, the higher the aggressiveness). min(1, N switch / N switch，max ) is the rapid switching term, which is used to specifically consider the high-risk pedal rapid switching behavior to strengthen the safety consideration.

[0074] In the embodiment of the application, AI1 steer is used to quantify the aggressiveness of the steering wheel operation, and the “steering speed limit”, “high rotation frequency” and “steering stability” are comprehensively calculated by the formula: AI1 steer =d1×|ω steer | max / ω steer ,max+d2×min(1, N steer / N steer ,max)+d3×min(1, σ δ / σ δmax )。

[0075] Wherein, |ω steer | max is the maximum value of the absolute value of the steering wheel rotation speed in the set time window, in degrees / second, reflecting the steering speed. ω steer ,max is the aggressive threshold of the steering wheel rotation speed, for example, 500 degrees / second, and exceeding this value is considered as rapid steering. |ω steer | max / ω steer,max is the steering speed limit term, used to quantify the speed aggression of a single steering operation by normalizing the peak speed. N steer This refers to the number of times the steering wheel speed exceeds a set speed threshold within a defined time window. The set speed threshold can be 200 degrees / second. steer `max` represents the number of times the steering wheel speed exceeds a set threshold within a defined time window, for example, 8 times. `min(1, N)` steer / N steer σmax) represents the high-speed frequency term, used to statistically analyze high-frequency sharp steering behavior, reflecting the frequent and aggressive nature of steering operations. δ σ represents the standard deviation of the steering wheel angle, expressed in degrees squared. It reflects the degree of steering fluctuation; a larger value indicates greater instability. δmax The radical threshold for the standard deviation of the steering wheel angle can be the square of 3000 degrees. min(1, σ) δ / σ δmax The ) is the steering stability term, used to quantify the smoothness of steering operations by normalizing the standard deviation (the greater the fluctuation, the higher the aggression).

[0076] d1, d2, and d3 are the speed limit weight, high-speed frequency weight, and stability weight, respectively, and d1 + d2 + d3 = 1. In one illustrative embodiment, d1 = 0.5, d2 = 0.3, and d3 = 0.2. Further, in this embodiment of the invention, the second calculation method uses a weighted fusion of normalized features from three dimensions: "vehicle dynamic response, pedal operation, and steering operation," and introduces a basic compensation term. The formula is: F2 = a1 × Norm(AI2) acc )+a2×Norm(AI2 pedal )+a3×Norm(AI2 steer )+a4×φ.

[0077] Among them, AI2 acc Let a1 be the vehicle dynamic response level characterization value obtained based on the second vehicle dynamic response determination method, and AI2 be the value of a1. acc The weights, AI2 pedal a2 is the pedal operation degree characterization value obtained based on the second pedal operation method, where a2 is AI2. pedal The weights, AI2 steer a3 is the steering operation degree characterization value obtained based on the second steering operation determination method, where a3 is AI2. steerThe weights are defined as follows: a4 is a preset coefficient, φ is a synchronization penalty term, and Norm(·) represents the normalization operation. In an illustrative embodiment, a1 = 0.4, a2 = 0.3, a1 = 0.2, a4 = 0.1, ensuring that F2 maintains the basic reference value when there is no aggressive operation. In this embodiment of the invention, if the instantaneous longitudinal acceleration is greater than or equal to 0.3g and the brake pedal depth is greater than or equal to 20%, it indicates that the simultaneous operation of rapid acceleration and braking poses a risk of power conflict, and φ = 0.2. If the instantaneous steering wheel speed is greater than or equal to 150° / s and the accelerator pedal depth is less than or equal to 10%, it indicates that there is insufficient power during sharp turns, posing a risk of sideslip, and φ = 0.1. In other cases, φ = 0.

[0078] In this embodiment of the invention, Norm(x) = (2 / (1+e) -k(x-x0) )-1. Where x is the original driving behavior representation value that needs to be normalized, specifically including: AI2 acc AI2 pedal and AI2 steer k is a preset curve steepness coefficient, which can be 2. x0 is a dynamic baseline value, a historical baseline value corresponding to x, used to quantify the reference baseline of current driving behavior. Its calculation formula is: x0 = 0.7 × x avg +0.3×x max x avg The rolling average of x within a set time window.

[0079] x(τ) is the instantaneous sampled value of x at time τ. max The 90th percentile of x for the target vehicle within a set historical statistical period, such as 3 months.

[0080] Furthermore, in this embodiment of the invention, AI2 acc The cumulative intensity of the dynamic response is quantified by continuous integration within a time window, focusing on the "persistent effects of longitudinal / lateral acceleration and yaw rate". The calculation formula is as follows:

[0081]

[0082] Where T is the duration of the set time window, and a x 2 (τ) represents the longitudinal acceleration at time τ, a y 2 (τ) represents the transverse acceleration at time τ, ω z (τ) represents the yaw rate at time τ, k x For the vertical weights, k y k represents the horizontal weights. ω For the yaw weight, k x +k y +kω = 1, t is the current time, τ is the integral variable, dτ represents the continuous integration operation on time τ. In an illustrative embodiment, k x = 0.5, k y = 0.3, k ω = 0.2.

[0083] AI2 acc The formula means that the integral result is divided by the window length T to obtain the average dynamic response strength in the time window, avoiding the one-sidedness of instantaneous peaks, and more in line with the cumulative effect of "continuous aggressiveness" in actual driving.

[0084] In the embodiment of the application, AI2 pedal The "aggressive operation frequency" and "pedal depth extreme value" are combined to quantify the aggressiveness of pedal operation, and the calculation formula is: AI2 pedal = h1 x N aggr / (T) 1 / 2 + h2 x (β th，max + β br，max ) / 2.

[0085] Wherein, N aggr is the sum of the number of times that the accelerator depth exceeds the corresponding threshold and the number of times that the brake depth exceeds the corresponding threshold in the set time window, which satisfies any one of the following conditions: the accelerator depth is greater than 80% and the change rate is greater than 50% / s; the brake depth is greater than 60% and the change rate is greater than 40% / s. h1 and h2 are weight coefficients, and h1+h2=1. In an illustrative embodiment, h1=0.7, h2=0.3.

[0086] In the embodiment of the application, AI2 steer Focus on "steering stability", "steering speed limit" and "high rotation speed duration" to quantify the aggressiveness of steering operation, and the calculation formula is: AI2 steer = max (σ δ / g1, ω δmax / g2) x (1+q x t over / T).

[0087] Wherein, g1 is a preset angle value, which can be 10° for example, and exceeding this value is considered as steering fluctuation aggressiveness. ω δmax is the maximum steering wheel rotation speed, specifically the maximum value of the absolute value of the steering wheel rotation speed in the set time window, that is, the maximum instantaneous value of the steering wheel rotation speed at all sampling times in the window, for example, the steering wheel rotation speed in a 30-second window reaches 450° / s, then ω δmax = 450° / s. g2 is a first preset rotation speed value, for example, 300 degrees per second, and exceeding this value is considered as rapid direction. t overFor a set time window, the cumulative time of the steering wheel speed greater than the second preset speed value, the second preset speed value can be 200 degrees per second. Q is a time proportion correction coefficient, which can be 0.5, used to enhance the influence of long-time high-speed steering (the risk of continuous sharp steering is higher).

[0088] In the embodiment of the application, AI2 acc The integral form is used to replace the discrete peak value statistics, which can more accurately capture the cumulative effect of "continuous dynamic response" and avoid the interference of instantaneous value fluctuation. AI2 pedal T 1 / 2 The influence of the balance window length on the frequency is balanced to ensure that the aggressive frequencies under different window lengths are comparable. AI2 steer By max() and the duration correction term, the "extreme risk point" in the steering operation is highlighted, and the sensitivity to dangerous steering behavior is improved. The introduction of a4 ensures that F2 still maintains a reasonable baseline value (such as 0.1) when there is no aggressive operation, avoiding the control strategy failure caused by zero value. Through the above quantitative and fusion of sub-indices, F2 can more comprehensively reflect the aggressiveness of the driving behavior, providing accurate scene judgment basis for the dynamic adjustment of the electric motor torque of the electric vehicle.

[0089] In the embodiment of the application, △F = |F1-F2|, that is, the absolute difference value of the first calculation mode (F1, focusing on real-time operation) and the second calculation mode (F2, focusing on cumulative characteristics), is used to quantify the divergence degree of the two evaluation models.

[0090] In the embodiment of the application, the plurality of preset difference intervals include a first difference interval to a fourth difference interval, the first difference interval is [0, p1], the second difference interval is (p1, p2], the third difference interval is (p2, p3], and the fourth difference interval is (p3, +∞); p1 to p3 are respectively a first set value to a third set value.

[0091] In one illustrative embodiment, p1 = 0.1, p2 = 0.3, and p3 = 0.5. By comparing the calculation deviation distribution of F1 and F2, p1 corresponds to "model slight divergence" (deviation ≤10%, proportion 60% normal scene), p2 corresponds to "moderate divergence" (10%<deviation ≤30%, proportion 30% scene), and p3 corresponds to "serious divergence" (deviation >30%, proportion 10% complex scene), ensuring that the interval division covers 99% of actual working conditions.

[0092] Further, the value rule comprises:

[0093] When △F belongs to the first difference interval, if the road state is a curve, F=F1, the curve scene depends on real-time dynamic response (such as instantaneous steering speed, lateral acceleration), F1 focuses on instantaneous operation, and is more accurate. If the road state is high-speed cruising or congestion crawling, F=F2, high-speed cruising (continuous stability) and congestion crawling (frequent start-stop) depend on cumulative characteristics (such as average speed fluctuation, start-stop frequency), F2 focuses on long-term statistics, and is more reliable. If the road state is rapid acceleration or long uphill, F=(F1+F2) / 2, rapid acceleration (instantaneous burst force) and long uphill (continuous load) need real-time + cumulative combination (such as peak value and duration of rapid acceleration, slope accumulation of long uphill), and the average of the two is more comprehensive.

[0094] When △F belongs to the second difference interval, F=∑ 3 i＝1 (E1 i ×AI1 i +E2 i ×Norm(AI2 i ))+f×a4.

[0095] Wherein, i=1, 2, 3, respectively corresponding to acc, pedal and steer, E1 i is the correction weight of AI1 i , E1 i =w i ×(1-|△AI i | / p2), E2 i is the correction weight of AI2 i , E1 i =a i ×(1-|△AI i | / p2), △AI i =AI1 i -Norm(AI2 i ), that is, the greater the divergence of the sub-item, the lower the correction weight, avoiding the influence of error amplification of the divergence dimension. f is the correction coefficient of a4. For example, f=0.8, f is used when the divergence increases, reducing the influence of the basic compensation, and avoiding the interference of the fixed compensation term with the dynamic correction.

[0096] When △F belongs to the third difference interval, F=S×(m1×F1+m2×F2)+(1-S)×F avg ; S is a data integrity score, m1 and m2 are preset coefficients, m1+m2=1, F avg is the average representation value of historical driving behavior of the same road section at the same period.

[0097] In the embodiment of the present application, S can be determined based on the effective frame proportion of sensors such as GPS, IMU, pedal sensor, S∈[0, 1], 1 represents full effectiveness, and 0.5 represents half effectiveness. Specifically, for the core sensors (IMU, pedal sensor, steering angle sensor, GPS, camera) of the target vehicle, the effective frame proportion of each sensor is counted in the same time window (such as 30 seconds) as F1 / F2, and the effective frame proportion is equal to the number of effective frames divided by the total number of collected frames.

[0098] The effective frame determination condition of the IMU is that the acceleration is less than 2g and the yaw rate is less than 5 rad / s. The effective frame determination condition of the pedal sensor is that the throttle and brake opening is less than 100% and the opening rate is less than 100% / s. The effective frame determination condition of the steering angle sensor is that the steering wheel angle is between -720° and +720° and the steering speed is less than 500° / s. The effective frame determination condition of the GPS is that the positioning accuracy is less than 2 and the vehicle speed is less than 250 km / h. The effective frame determination condition of the camera is that the lane line recognition confidence is greater than 60% and the target detection frame rate is greater than 10 fps.

[0099] S=∑ Z q＝1 (b q ×min(Sq, 1)), where Sq is the effective frame proportion of the qth sensor, q is 1 to Z, Z is the number of sensors, b q is the weight of the qth sensor. In an illustrative embodiment of the present application, the weights of the IMU, pedal sensor, steering angle sensor, GPS, and camera can be set to 0.35, 0.3, 0.2, 0.1, and 0.05, respectively.

[0100] In an illustrative embodiment, m1=0.6 and m2=0.4.

[0101] When △F belongs to the third difference interval, when S is high (data is reliable): prefer to rely on real-time models F1 and F2; when S is low (data is missing): prefer to rely on historical average Favg to avoid errors caused by real-time calculation errors.

[0102] When △F belongs to the fourth difference interval, if F1 belongs to [F min , F max ] and F2 does not belong to [F min , F max ], then F=F1, if F2 belongs to [F min , F max ] and F1 does not belong to [F min , F max ], then F=F2, if F1 and F2 both belong to [F min , Fmax If F1 and F2 both do not belong to [F min , F max ], then F=F avg , F min . max is the minimum value of the normal driving behavior characteristic value, and F min is the maximum value of the normal driving behavior characteristic value.

[0103] In the embodiment of the present application, F mamx and F mi may be calibrated through a million normal driving data: 1000 family cars are selected for non-aggressive driving data in urban, highway, mountain and other scenes (excluding driving behaviors corresponding to accidents and illegal records), the distribution characteristics of the driving behavior characteristic value F are counted, the lower limit of the 95% confidence interval is F mamx =0.2, and the upper limit is F 5 =0.6, covering most normal driving conditions.

[0104] Further, in the embodiment of the present application, the road state anomaly characteristic value (RAI) is a core index for comprehensively quantifying the abnormal degree of the road on which the vehicle travels, and is obtained by weighted fusion of abnormal characteristics of five typical scenes of sudden acceleration, long uphill, high-speed cruise, congestion creep and curve maintenance, and the formula is: RAI=∑ j＝1 n j ×B j .

[0105] Wherein, B j is the j road state anomaly characteristic values in the set time window, the first road state anomaly characteristic value to the fifth road state anomaly characteristic value, i.e., B1 to B5, are sudden acceleration anomaly characteristic value, long uphill anomaly characteristic value, high-speed cruise anomaly characteristic value, congestion creep anomaly characteristic value and curve maintenance anomaly characteristic value, the value of j is 1 to 5, n j is the weight of B j , and ∑ 5 j＝1 n j =1. In one illustrative embodiment, n1=n2=n3=n4=n5=0.2.

[0106] Wherein, B1 is used to quantify the abnormal degree of sudden acceleration behavior, and comprehensively quantifies three dimensions of “acceleration intensity, duration, frequency”, B1=w 11 ×Norm(a avg , a 01 , a 02 )+w 12 ×Norm(T total , t 01 , t 02 .)+w 13 ×Norm(f 01 ,f 02 ,f avg ), where a avg is the average acceleration, a N = 1 / N1 x∑ u11＝1 a u1 , a u1 is the peak instantaneous acceleration of the u1th harsh acceleration event within the set time window, u1 takes value from 1 to N1, N1 is the number of harsh acceleration events; a 01 is the normal acceleration threshold, a 02 is the harsh acceleration threshold, w 11 is the acceleration weight in harsh acceleration scenario, T total is the total harsh acceleration duration within the set time window, T total =∑ N1 u1＝1 t u1 , t u1 is the duration of the u1th harsh acceleration event within the set time window, t 01 is the normal acceleration duration threshold, t 02 is the harsh acceleration duration threshold, w 12 is the acceleration duration weight in harsh acceleration scenario, f is the harsh acceleration event frequency within the set time window, f = N / T, T is the time length of the set time window, f 01 is the normal event frequency threshold, f 02 is the harsh event frequency threshold, w 13 is the event frequency weight in harsh acceleration scenario.

[0107] In one illustrative embodiment, a 01 = 0.3g, a 02 = 0.6g, t 01 = 5s, t 02 = 015s, f 01 = 0.1 times / s, f 02 = 0.3 times / s, w 11 = 0.4, w 12 = 0.3, w 13 = 0.3.

[0108] B2 is used to quantify the abnormality degree of long uphill road section, which integrates the three dimensions of "slope size, duration, steep slope length", B2 = w 21 x Norm(I avg , I 01 , I 02 ) + w 22 x Norm(Tup, Tup 01 , Tup02 )+w 23 ×Norm(L total L 01 L 02 ); where I avg The average road gradient within a specified time window, expressed as a percentage. I(t) is the real-time road slope at time t, w 21 The average slope weight is denoted as Tup, which represents the total uphill time in a long uphill scenario. Tup = ∑ j1＝1 M (t j1 ), t j1 The uphill time corresponding to the j1th uphill event within a set time window, where j1 ranges from 1 to M, and M is the total number of uphill events, w 22 For uphill time weighting in long uphill scenarios, Tup 01 Tup is the threshold for the duration of a normal uphill climb. 02 The threshold for the duration of abnormal uphill slope; L total For the cumulative steep slope length, L total =∑ j1＝1 M (L j1 ), L j1 L represents the length of the steep slope section corresponding to the j1th uphill event within a defined time window. j1 Greater than or equal to I 01 The length of the road segment. L 01 L is the threshold for the cumulative length of a normal steep slope. 02 w is the threshold for the cumulative length of abnormally steep slopes. 23 The steep slope length weight is used in scenarios with long uphill slopes.

[0109] In one illustrative embodiment of the present invention, I 01 =3%, I 02 =8%, Tup 01 =60s, Tup 02 =300s, L 01 =1km, L 02 =5km, w 21 =0.4, w 22 =0.3, w 23 =0.3.

[0110] B3 is used to quantify the degree of abnormality in high-speed cruise scenarios, combining three dimensions: "speed fluctuation amplitude, sudden deceleration, and fluctuation frequency". B3 = w 31 ×Norm(△v avg , △v 01 , △v 02 )+w 32 ×Norm(davg , d 01 , d 02 )+w 33 x Norm (△vf, △vf 01 , △vf 02 ); wherein, △v avg is the average speed fluctuation amplitude, △v avg = 1 / N2 x ∑ N2 u2＝1 (△v u2 ), △v u2 is the u2th speed fluctuation amplitude in the set time window, u2 is 1 to N2, N2 is the speed fluctuation frequency in the set time window, △v 01 is the normal speed fluctuation threshold, △v 02 is the abnormal speed fluctuation threshold, w 31 is the speed fluctuation weight in the high-speed cruise scene; d avg is the average sudden deceleration, d avg = 1 / N2 x ∑ N2 u2＝1 d u2 , d u2 is the u2th sudden deceleration peak value in the set time window, the unit is m / s 2 . d 01 is the normal sudden deceleration threshold, d 02 is the abnormal sudden deceleration threshold, w 32 is the sudden deceleration weight in the high-speed cruise scene; △vf is the speed fluctuation frequency, △vf = N2 / T, △vf is the speed fluctuation frequency, △vf 01 is the normal speed fluctuation frequency threshold, △vf 02 is the abnormal speed fluctuation frequency threshold, w 33 is the speed fluctuation frequency weight in the high-speed cruise scene.

[0111] In an illustrative embodiment of the application, △v 01 = 5km / h, △v 02 = 15km / h, d 01 = 2m / s 2 , d 02 = 5m / s 2 , △vf 01 = 0.1 times / sec, △vf 02 = 0.3 times / sec, w 31 = 0.3, w 32 = 0.4, w 33 = 0.3.

[0112] B4 is used to quantify the abnormal degree of the congestion crawling scenario, and is a comprehensive three-dimensional of "start-stop frequency, idling proportion, average speed", B4 = w 41 × Norm (Sf avg , Sf 01 , Sf 02 ) + w 42 × Norm (r avg , r 01 , r 02 ) + w 43 × Norm (v avg , v 01 , v 02 ), wherein Sf avg is the average start-stop frequency, Sf avg = Sf / T, Sf is the number of start-stop times in a set time window, and start-stop refers to the speed from 0 to greater than 5 km / h or from greater than 5 km / h to 0. Sf 01 is the normal start-stop frequency threshold, Sf 02 is the abnormal start-stop frequency threshold, w 41 is the start-stop frequency weight in the congestion crawling scenario; r avg is the average idling proportion, r avg = (∑ N3 u3＝1 (r u3 )) / T, r u3 is the duration of the u3th idling event in a set time window, u3 is valued from 1 to N3, N3 is the total number of idling events in a set time window, r 01 is the normal idling proportion threshold, r 02 is the abnormal idling proportion threshold, w 42 is the idling event weight in the congestion crawling scenario; v avg is the average speed, v(t) is the real-time speed at time t, v 01 is the normal average speed threshold, v 02 is the abnormal average speed threshold, w 43 is the speed weight in the congestion crawling scenario.

[0113] In an illustrative embodiment of the present application, Sf 01 = 0.2 times / min, Sf 02 = 1 time / min, r 01 = 20%, r 02 = 60%, v 01 = 0.1 km / h, v 02 = 30 km / h, w 41 = 0.3, w 42 = 0.3, w 43 = 0.4.

[0114] B5 is used for quantifying the abnormal degree of curve driving, and is a comprehensive three-dimensional of "speed ratio, lane deviation distance, lateral acceleration", B5 = w 51 × Norm(k avg , k 01 , k 02 ) + w 52 × Norm(Dd avg , Dd 01 , Dd 02 ) + w 53 × Norm(a y-avg , a y-01 , a y-02 ).

[0115] wherein, k avg is an average speed ratio, k avg = 1 / N4 x ∑ N4 u4＝1 (k u4 ), k u4 is a ratio of an actual speed over a curve to a reference speed over a curve when the u4th curve driving occurs within a set time window, u4 is valued from 1 to N4, N4 is a number of curves within the set time window, k 01 is a normal speed ratio threshold, k 02 is an abnormal speed ratio threshold, w 51 is a speed ratio weight under a curve maintaining scenario; Dd avg is an average lane deviation distance, Dd avg = 1 / N4 x ∑ N4 u4＝1 (Dd u4 ), Dd u4 is a lane deviation distance when the u4th curve driving occurs within the set time window, Dd 01 is a normal lane deviation distance threshold, Dd 02 is an abnormal lane deviation distance threshold, a y-avg is an average lateral acceleration, a y-avg = 1 / N4 x ∑ N4 u4＝1 (a y-u4 ), a y-u4 is a lateral acceleration when the u4th curve driving occurs within the set time window, a y-01 is a normal lateral acceleration threshold, a y-02 is an abnormal lateral acceleration threshold.

[0116] In a schematic embodiment of the present application, k 01 = 1.1, k 02 = 1.5, Dd 01= 0.5m, Dd 02 = 1.5m, a y-01 = 0.4g, a y-02 = 0.8g, w 51 = 0.3, w 52 = 0.3, w 53 = 0.4.

[0117] wherein, Norm(x1, x 01 , x 02 ) = min(max((x-x 01 ) / (x 02 -x 01 ), 0), 1), wherein x1 is a raw physical parameter reflecting road state anomaly, including parameter values in B1 to B5, such as acceleration, speed, speed fluctuation, etc. x 01 is a normal threshold value corresponding to x1, and x 02 is an abnormal threshold value corresponding to x1. Through the above normalization processing, the raw parameters of B1 to B5 are converted into quantified values with clear risk semantics, providing reliable input for accurate calculation of road state anomaly degree.

[0118] The RAI output range is [0, 1], and the higher the value, the higher the road state anomaly degree, for example, RAI = 0.8 represents a relatively high overall road anomaly risk.

[0119] Further, in the embodiments of the present application, F0 can be set to 0.6. Through statistical analysis of 100 test vehicles, 100,000 kilometers of real road data, 95% of the smooth driving behavior representation value F is ≤0.6, and 90% of the aggressive driving behavior representation value F is >0.6, so F0 is set as the critical value between smooth and aggressive driving.

[0120] When F exceeds the preset threshold F0, the vehicle is in aggressive working conditions such as sudden acceleration and high-speed overtaking, and the torque distribution is centered on improving the handling response and power performance. Specifically, in S400, if F > F0, the motor torque distribution ratio is dynamically adjusted based on F, specifically including:

[0121] The rear axle torque distribution ratio of the target vehicle is set to f f = b1 + b2 x tanh(b3 x (F-b4)), and the front axle torque distribution ratio is 1-f f , b1, b2, b3 and b4 are all preset calibration parameters.

[0122] In the embodiment of the application, b1 ensures that the front and rear axle torques are basically divided equally when the degree of aggressiveness is low (F ≈ b4), and the basic stability is maintained, b2 is used to control the maximum change range of the rear axle ratio, the greater b3 is, the more sensitive the influence of F change on the rear axle ratio (when the degree of aggressiveness increases, the rear axle ratio rapidly increases), when F = b4, tanh(0) = 0, and the rear axle ratio is b1; when F > b4, the rear axle ratio increases with the increase of F. In an illustrative embodiment, b1 = 0.5, b2 = 0.3, b3 = 3.5, and b4 = 0.8.

[0123] When F ≤ F0, the vehicle is in a smooth working condition such as cruising, congestion, and gentle slope, and the torque distribution is focused on improving driving stability and energy efficiency, and is dynamically adjusted in combination with the road state abnormality RAI, specifically, in S400, if F ≤ F0, the motor torque distribution ratio is dynamically adjusted based on RAI, specifically including:

[0124] The rear axle torque distribution ratio of the target vehicle is set as f f = (1 / (1 + e z×(RAI-RAI0) )) + △f steer , the front axle torque distribution ratio is 1-f f , z is an exponential decay coefficient, which can be 2.5. RAI0 is a road state representation value threshold, which can be 0.5. RAI0 is based on the abnormality statistics of five typical road scenes (sudden acceleration, long uphill, etc.), when RAI < 0.5, the influence of road state on driving safety can be ignored (accident risk rate < 0.1%); when RAI > 0.5, the risk rate increases significantly (> 1%), so it is taken as the critical value. e is a natural constant, and △f steer is a curve correction term.

[0125] In the embodiment of the application, △f steer = C steer × sign(a lat ) × min(|a lat |, a lat,max ). Wherein, C steer is a curve correction gain, and the value is 0.1, that is, 0.1 of the ratio adjustment corresponds to every 1g lateral acceleration. a lat is the lateral acceleration, and a lat,max is the upper and lower lateral acceleration. The role of △f steer is to reduce the left axle torque ratio and increase the right axle torque ratio when turning left to suppress side slipping, and vice versa when turning right to optimize the curve grip.

[0126] Those skilled in the art should understand that the determination method of the total motor torque of the target vehicle belongs to the prior art, and the core logic is to match the driver's power demand and the real-time state of the vehicle, specifically as follows:

[0127] The total torque of the motor is mainly calculated according to the power request of the driver (such as the opening degree of the accelerator pedal, the target vehicle speed of the cruise control, etc.), in combination with the running constraint conditions of the vehicle (such as the current vehicle speed, the SOC of the battery, the motor temperature, the road adhesion, etc.). For example, in the active driving scene, the total torque is preliminarily determined through the opening degree of the accelerator pedal and a preset "pedal-torque mapping relationship", and then dynamically corrected according to the remaining battery capacity (such as limiting the output when the SOC is too low), the motor overheating protection threshold, the upper limit of the vehicle speed, etc. In the cruise scene, the total torque is adjusted based on the difference between the target vehicle speed and the actual vehicle speed through a proportional-integral-derivative (PID) control or a model predictive control algorithm, so as to maintain a stable vehicle speed and overcome the rolling resistance, air resistance, etc. The determination logic of the total torque described above has been widely used in the existing electric vehicle power control system, and belongs to the known technology in the field. The improvement of the present application does not involve the calculation method of the total torque of the motor itself, but focuses on the optimization of the distribution strategy after the total torque is determined: by introducing the driving behavior representation value F and the road state representation value RAI, the dynamic torque distribution rules in different scenes are constructed: when the driving is aggressive (F>F0), the front and rear axle torque ratio is adjusted based on F to enhance the control response; when the driving is stable (F≤F0), the distribution is optimized based on RAI to improve the stability and energy efficiency. That is, the innovation point of the present application lies in "how to distribute the total torque in different scenes", rather than "how to determine the size of the total torque", and the determination of the total torque is the basic condition for implementing the present application, which can be realized by using the existing technology. The present application also provides an electronic device, which comprises at least one processor and a memory in communication connection with the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are arranged to execute the method described in the embodiments of the present application.

[0128] The embodiments of the present application also provide a computer readable storage medium storing computer executable instructions, wherein the computer executable instructions are used to execute the method described in the embodiments of the present application.

[0129] It should be understood that the steps shown above can be reordered, added or deleted. For example, the steps described in the present application can be executed in parallel, in sequence or in different orders, as long as the desired results of the technical solutions disclosed in the present application can be achieved, which is not limited herein.

[0130] The above detailed description does not constitute a limitation on the protection scope of the present application. Those skilled in the art should understand that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent replacement and improvement made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A method of controlling an electric motor of an electric vehicle, characterized by, The method comprises the following steps: S100, acquiring sensor data of a target vehicle, the sensor data comprising driving behavior data and road state data; wherein the driving behavior data is data reflecting the operating characteristics of a driver, and the road state data is data reflecting the characteristics of a driving scene; S200, acquiring a driving behavior representation value F representing the degree of aggressiveness of driving operation of the target vehicle based on the driving behavior data; S300, acquiring a road state representation value RAI representing the degree of abnormality of the road on which the target vehicle is driving based on the road state data; S400, comparing the driving behavior representation value F with a preset threshold value F0, and if F>F0, then dynamically adjusting the motor torque distribution ratio based on F, and if F≤F0, then dynamically adjusting the motor torque distribution ratio based on RAI.

2. The method of claim 1, wherein, F is acquired by the following steps: S201, processing the driving behavior data using a first calculation method to obtain a driving behavior representation value F1, and processing the driving behavior data using a second calculation method to obtain a driving behavior representation value F2; S202, calculating the absolute value of the difference between F1 and F2, denoted as △F; S203, determining the final driving behavior representation value F based on a preset driving behavior representation value reference table and △F; wherein the preset driving behavior representation value reference table comprises a plurality of preset difference intervals, each difference interval is mutually exclusive and continuously covers the possible value range of △F, and each difference interval corresponds to a unique value rule.

3. The method of claim 2, wherein, The first calculation mode satisfies the following condition: F1 = w1 x AI1 acc + w2 x AI1 pedal + w3 x AI1 steer ; AI1 acc is a vehicle dynamic response degree representation value determined based on a first vehicle dynamic response determination manner, w1 is a weight of AI1 acc , w2 is a weight of AI1 pedal , and w3 is a weight of AI1 pedal ; AI1 steer is a steering operation degree representation value determined based on a first steering operation determination manner, w3 is a weight of AI1 steer , and w1+w2+w3=1. The second calculation mode satisfies the following condition: F2 = a1 x Norm(AI2 acc ) + a2 x Norm(AI2 pedal ) + a3 x Norm(AI2 steer ) + a4 x φ; wherein AI2 acc is a vehicle dynamic response degree representation value determined based on a second vehicle dynamic response determination manner, a1 is a weight of AI2 acc , AI2 pedal is a pedal operation degree representation value determined based on a second pedal operation manner, a2 is a weight of AI2 pedal , AI2 steer is a steering operation degree representation value determined based on a second steering operation determination manner, a3 is a weight of AI2 steer , a4 is a preset coefficient, φ is a synchronization penalty term, and Norm(·) represents a normalization operation.

4. The method of claim 3, wherein, AI1 acc = k1 x (|a x | max / a max + |a y | max / a max + |ω z | max / ω max ) / 3 + k2 x min(1, N acc / N max ); wherein |a x | max is the maximum value of the absolute value of the longitudinal acceleration within a set time window, |a y | max is the maximum value of the absolute value of the lateral acceleration within a set time window, a max is the aggressive threshold value of the acceleration, |ω z | max is the maximum value of the absolute value of the yaw rate within a set time window, ω max is the aggressive threshold value of the yaw rate; N acc is the total number of times that the longitudinal acceleration, the lateral acceleration or the yaw rate exceeds the corresponding threshold value within a set time window; N max is the maximum allowed number of times of the dynamic response dimension aggressive operation, k1 and k2 are preset coefficients, and k1 + k2 = 1. AI1 pedal = c1 x min(1, (N th / N th，max +N br / N br，max ) / 2) + c2 x ((β th，max / 100 + β br，max / 100) / 2) + c3 x min(1, N switch / N switch，max ); wherein N th is the number of times that the accelerator depth is greater than a preset accelerator depth threshold within a set time window, N th，max is the maximum allowed number of times of aggressive accelerator operation, N br is the number of times that the brake depth is greater than a preset brake depth threshold within the set time window, N br，max is the maximum allowed number of times of aggressive brake operation, β th，max is the maximum depth of the accelerator pedal within the set time window, β br，max is the maximum depth of the brake pedal within the set time window, N switch is the number of times of rapid switching between the accelerator and the brake within the set time window, N switch，max is the maximum allowed number of times of rapid switching between the accelerator and the brake within the set time window, and c1, c2, and c3 are weight coefficients, with c1 + c2 + c3 = 1. AI1 steer = d1 x |ω steer | max / ω steer ,max + d2 x min(1, N steer / N steer ,max) + d3 x min(1, σ δ / σ δmax ); wherein |ω steer | max is a maximum value of the absolute value of the steering wheel speed in a set time window, ω steer ,max is an aggressive threshold value of the steering wheel speed, N steer is a number of times that the steering wheel speed is greater than a set speed threshold value in a set time window, N steer ,max is an allowed number of times that the steering wheel speed is greater than a set speed threshold value in a set time window, σ δ is a standard deviation of the steering wheel angle, σ δmax is an aggressive threshold value of the standard deviation of the steering wheel angle, d1, d2, and d3 are a speed limit weight, a high speed frequency weight, and a stability weight, respectively, and d1+d2+d3=1.

5. The method of claim 4, wherein where T is the length of the set time window, a x 2 (τ) is the longitudinal acceleration at time τ, a y 2 (τ) is the lateral acceleration at time τ, ω z (τ) is the yaw rate at time τ, k x is the longitudinal weight, k y is the lateral weight, k ω is the yaw weight, k x +k y +k ω = 1, t is the current time, τ is the integral variable, and dτ represents a continuous integration operation with respect to time τ. AI2 pedal = h1 x N aggr / (T) 1 / 2 + h2 x (β th，max + β br，max ) / 2; wherein N aggr is the sum of the number of times the accelerator depth exceeds the corresponding threshold and the number of times the brake depth exceeds the corresponding threshold within the set time window, h1 and h2 are weight coefficients, and h1 + h2 = 1. AI2 steer = max(σ δ / g1, ω δmax / g2) x (1 + q x t over / T), wherein g1 is a preset steering angle value, ω δmax is a maximum steering wheel speed, g2 is a first preset speed value, t over is a cumulative time within a set time window in which the steering wheel speed is greater than a second preset speed value, and q is a time proportion correction coefficient.

6. The method of claim 3, wherein, the plurality of preset difference intervals comprise a first difference interval to a fourth difference interval, the first difference interval is [0, p1], the second difference interval is (p1, p2], the third difference interval is (p2, p3], and the fourth difference interval is (p3, +∞); p1 to p3 are respectively a first set value to a third set value; the value rule comprises: when △F belongs to the first difference interval, if the road state is a curve, F=F1, if the road state is high-speed cruising or congestion creeping, F=F2, and if the road state is sudden acceleration or long uphill, F=(F1+F2) / 2; When △F belongs to the second difference interval, F = ∑ 3 i＝1 (E1 i × AI1 i + E2 i × Norm(AI2 i )) + f × a4; wherein i = 1, 2, 3, respectively corresponding to acc, pedal and steer, E1 i is the correction weight of AI1 i , E1 i = w i × (1 - |△AI i | / p2), E2 i is the correction weight of AI2 i , E1 i = a i × (1 - |△AI i | / p2), △AI i = AI1 i - Norm(AI2 i ), f is the correction coefficient of a4. When the △F belongs to the third difference interval, F=S×(m1×F1+m2×F2)+(1-S)×F avg ; S is a data integrity score, m1 and m2 are preset coefficients, m1+m2=1, F avg is the average representation value of the historical driving behavior of the same section at the same time. When △F belongs to the fourth difference interval, if F1 belongs to [F min , F max ] and F2 does not belong to [F min , F max ], F=F1, if F2 belongs to [F min , F max ] and F1 does not belong to [F min , F max ], F=F2, if F1 and F2 both belong to [F min , F max ], F=(F1+F2) / 2, if F1 and F2 both do not belong to [F min , F max ], F=F avg , F min is the minimum value of the normal driving behavior characteristic value, and F max is the maximum value of the normal driving behavior characteristic value.

7. The method of claim 6, wherein, RAI satisfies the following condition: RAI =∑ 5 j＝1 n j ×B j ;B j is j road state anomaly characteristic values within a set time window, j is valued from 1 to 5, n j is the weight of B j , ∑ 5 j＝1 n j =1, B1 to B5 are respectively an abnormal acceleration characteristic value, a long uphill characteristic value, a high-speed cruise characteristic value, a congestion crawling characteristic value and a curve maintaining characteristic value.

8. The method of claim 1, wherein, in S400, if F>F0, the motor torque distribution ratio is dynamically adjusted based on F, specifically comprising: The rear axle torque distribution ratio of the target vehicle is set as f f = b1 + b2 x tanh(b3 x (F - b4)), and the front axle torque distribution ratio is 1 - f f , b1, b2, b3, and b4 are all preset calibration parameters in S400, if F≤F0, the motor torque distribution ratio is dynamically adjusted based on RAI, specifically comprising: The rear axle torque distribution ratio of the target vehicle is set as f f = (1 / (1+e z×(RAI-RAI0) ))+△f steer , the front axle torque distribution ratio is 1-f f , z is an exponential decay coefficient, RAI0 is a road state representation value threshold, e is a natural constant, and△f steer is a curve correction term.

9. An electronic device, comprising: comprising a processor and a memory; the processor is configured to execute the steps of the method according to any one of claims 1 to 8 by invoking programs or instructions stored in the memory.

10. A computer-readable storage medium, characterized in that, The computer readable storage medium is configured to store programs or instructions for enabling a computer to execute the steps of the method according to any one of claims 1 to 8.

Images