An adaptive guidance type assisted driving system considering driver skill difference

Abstract

The application relates to an adaptive guidance type auxiliary driving system considering driver skill difference, which comprises the following modules: a driving skill classification module, which obtains a vehicle stability margin based on a current vehicle state, and takes the vehicle stability margin and a current driver state as inputs of a driver skill classification model to obtain a corresponding driver skill classification result; a skill learning interval classification module, which is used for obtaining a vehicle stability margin and a distance between a vehicle and a lane line boundary, and adopts a skill learning interval classification model to obtain a skill learning interval classification result; and an adaptive guidance driving right distribution module, which is used for realizing driving right distribution control according to the driver skill classification result and the skill learning interval classification result, and generating an auxiliary driving torque acting on a vehicle steering system. Compared with the prior art, the application has the advantages of improving the use degree of an auxiliary system and improving the driver skill as soon as possible under safe conditions.

Description

TECHNICAL FIELD

[0001] The present application relates to an assisted driving system, in particular to an adaptive guidance type assisted driving system considering the difference in driver skills. BACKGROUND

[0002] The assisted driving system is one of the main means to reduce traffic accidents, so the research on the assisted driving system is concerned by the automobile industry and the Internet industry. In "A survey of personalization for advanced driver assistance systems", it is pointed out that the standardization of the assisted driving system will cause the discomfort of the driver, which makes the driver unwilling to turn on the assisted driving system. In order to improve the usage rate of the assisted system, the personalized assisted driving system is studied, and the allocation of driving right is the main research point.

[0003] Patent CN 107804315 B determines the allocation of driving right by the difference between the expected front wheel angle and the actual front wheel angle of the decision, that is, the greater the difference, the farther the vehicle deviates from the safety area, and the control authority of the machine should be increased. However, the difference in driving skills and driving styles of the driver is not considered, which will often lead to conflicts between the driver and the assisted system. Patent CN 108819951 B proposes a man-machine co-driving lateral driving right allocation method considering the driving skills of the driver, in which the driving skills are reflected by the number of lane line exceedances and braking intensity. However, this ignores the state of the driver, and in the process of normal driving of the vehicle on the road, it often happens that novice drivers cannot timely discover the sudden situation due to nervousness. In "Changes in collision rates among novice drivers during the first months of driving", it is also pointed out that the first 6 months after obtaining a driver's license is a period of time when the accident probability of novice drivers decreases greatly compared with 2 years later. Therefore, the assisted driving system based on visual guidance is proposed in patent CN 113989775 B, hoping to reduce the burden of novice drivers and improve the concentration of the driver through visual guidance. With the same starting point, patent CN 114771574 A proposes a method of allocating driving right online based on the neural muscular state of the driver, hoping to solve the driving errors caused by the nervousness of the driver.

[0004] The above existing patent technologies have made significant research on personalized assisted driving technology, and all start from assisting novice drivers, which leads to the dependence of the driver on the assisted system. Since the assisted system cannot cover all working conditions, when the driver needs to take over, a novice driver who relies too much on the assisted system cannot handle such dangerous working conditions. SUMMARY

[0005] The present application aims to overcome the above-mentioned defects of the prior art and provides an adaptive guidance type auxiliary driving system that considers driver skill differences and improves driver skills as soon as possible in a safe condition.

[0006] The object of the present application can be achieved by the following technical solutions:

[0007] An adaptive guidance type auxiliary driving system that considers driver skill differences, comprising:

[0008] A driving skill classification module configured to obtain a current vehicle state and a current driver state, calculate a vehicle stability margin based on the current vehicle state, and use the vehicle stability margin and the current driver state as inputs of a driver skill classification model to obtain a corresponding driver skill classification result, wherein the vehicle stability margin is calculated according to an offline determined vehicle stability boundary;

[0009] A skill learning interval classification module configured to obtain the vehicle stability margin and a distance between the vehicle and a lane line boundary, and use a skill learning interval classification model to obtain a skill learning interval classification result;

[0010] An adaptive guidance driving right allocation module configured to realize driving right allocation control according to the driver skill classification result and the skill learning interval classification result, and generate an auxiliary driving torque acting on a vehicle steering system.

[0011] Further, the determination process of the vehicle stability boundary specifically comprises:

[0012] Constructing a vehicle dynamics model;

[0013] Based on the vehicle dynamics model, obtaining a root locus graph of vehicle state changes;

[0014] Classifying and training state points in the root locus graph by using a support vector machine to obtain a classification hyperplane, and using the classification hyperplane as the vehicle stability boundary.

[0015] Further, the vehicle dynamics model uses a two-degree-of-freedom vehicle model and a Dugoff tire model.

[0016] Further, the root locus graph is obtained by iterating each group of states using a 3rd order Runge-Kutta formula.

[0017] Further, the skill learning interval classification model is expressed as:

[0018]

[0019] wherein C sis the skill learning interval category, 1 represents the learning interval, and 0 represents the non-learning interval; D sl is the distance from the current state to the lower boundary of the vehicle stability boundary; D sh is the distance from the current state to the upper boundary of the vehicle stability boundary; T tlc is the time required for the front wheels of the vehicle to cross the lane line boundary.

[0020] Further, the adaptive guiding driving right allocation module comprises a first driving right allocation unit, an adaptive learning rate adjustment unit, and a second driving right allocation unit.

[0021] The first driving right allocation unit generates a corresponding first driving right allocation coefficient based on the driver skill classification result.

[0022] The second driving right allocation unit generates a corresponding second driving right allocation coefficient based on the first driving right allocation coefficient, the skill learning interval classification result, and the guiding torque obtained by the adaptive learning rate adjustment unit, and generates an auxiliary driving torque based on the second driving right allocation coefficient.

[0023] The process of obtaining the guiding torque by the adaptive learning rate adjustment unit comprises:

[0024] The artificial potential field function coefficient is corrected according to the auxiliary consistency rate of the driver hand torque and the auxiliary driving torque, the auxiliary expected trajectory is adjusted, and the guiding torque is generated.

[0025] Further, the calculation formula of the first driving right allocation coefficient is represented as:

[0026]

[0027]

[0028] wherein, ξ 1o is the first driving right allocation coefficient of an experienced driver; ξ 1n is the first driving right allocation coefficient of a novice driver; D sl is the distance from the current state to the lower boundary of the vehicle stability boundary; D sh is the distance from the current state to the upper boundary of the vehicle stability boundary.

[0029] Further, the correction of the artificial potential field function coefficient is specifically that when the auxiliary consistency rate is greater than 80%, the artificial potential field function correction coefficient is reduced by 0.1.

[0030] Further, the calculation formula of the second driving right allocation coefficient is represented as:

[0031] ξ o = ξ 1o * ξ 2o

[0032] ξ n =ξ 1n *ξ 2n

[0033] wherein, ξ o is the secondary driving right allocation coefficient of experienced drivers; ξ n is the secondary driving right allocation coefficient of novice drivers; ξ 1o is the primary driving right allocation coefficient of experienced drivers; ξ 1n is the primary driving right allocation coefficient of novice drivers; ξ 2o is the learning driving right allocation coefficient of experienced drivers; ξ 2n is the learning driving right allocation coefficient of novice drivers.

[0034] Further, the calculation formula of the learning driving right allocation coefficient is represented as:

[0035] ξ 2o =1

[0036]

[0037] wherein, C s is the skill learning interval category, 1 represents the learning interval, and 0 represents the non-learning interval; D sl is the distance from the current state to the lower boundary of the vehicle stability boundary; D sh is the distance from the current state to the upper boundary of the vehicle stability boundary.

[0038] Compared with the prior art, the present application has the following beneficial effects:

[0039] 1. The present application realizes the allocation control of the auxiliary system driving right through the driving skill classification and the skill learning interval classification, improves the use degree of the auxiliary system in the novice drivers, and improves the driving skills of the novice drivers as soon as possible.

[0040] 2. The present application realizes the classification of the driving skills by taking the vehicle stability margin and the current driver state as the driving skill features, wherein the vehicle stability margin and the current driver state (such as physiological acquisition information) can have the same vehicle stability margin distribution of the same driver in the same driving scene of different complexity, the accurate driving skill classification can be realized through less daily driving scenes, the classification model training is convenient, and the classification accuracy is high.

[0041] 3、The application simultaneously considers driving skill classification and skill learning interval classification in driving right allocation control, and realizes adaptive learning rate adjustment based on auxiliary consistency rate, realizes the purpose of skill learning intensity adjustment for the difference of learning ability of different novice drivers and the difference of learning ability of the same novice driver in different skill stages, and further improves the reliability. BRIEF DESCRIPTION OF DRAWINGS

[0042] Figure 1 It is a system architecture schematic diagram of the application;

[0043] Figure 2 It is a working principle schematic diagram of the driving skill classification module of the application;

[0044] Figure 3 It is a vehicle stability boundary acquisition principle diagram based on the support vector machine of the application;

[0045] Figure 4 It is a working principle schematic diagram of the adaptive guidance driving right allocation module of the application. DETAILED DESCRIPTION

[0046] The application will be described in detail below in combination with the drawings and specific embodiments. The embodiment is implemented on the premise of the technical scheme of the application, and detailed implementation modes and specific operation processes are given, but the protection scope of the application is not limited to the following embodiments.

[0047] In "Learner Driver Experience and Teenagers' Crash Risk During the First Year of Independent Driving", it is pointed out that the learning of driving skills is related to personal characteristics, and excessive intervention and high-intensity intervention will lead to the closing of the auxiliary system by the driver. In view of the problem that the inconsistent demand for auxiliary driving intensity caused by the inconsistent driving skills of different drivers, and the problem that the driving skill cannot be improved and cannot cope with the dangerous working conditions not covered by the auxiliary system due to the excessive dependence of novice drivers on the auxiliary driving system, in order to cope with the difference of learning ability of different drivers and the difference of learning ability of the same driver in different skill stages, the application proposes an adaptive guidance type auxiliary driving system considering the difference of driver skills, the purpose is to improve the driving skill of the driver as soon as possible under the condition of safety.

[0048] As Figure 1As shown, the adaptive guidance type auxiliary driving system considering the difference of driver skills provided by the embodiment comprises a driving skill classification module, a skill learning interval classification module and an adaptive guidance driving right distribution module. The driving skill classification module is used to obtain the current vehicle state and the current driver state, calculate the vehicle stability margin based on the current vehicle state, and take the vehicle stability margin and the current driver state as the input of a driver skill classification model to obtain the corresponding driver skill classification result. The vehicle stability margin is used to map the situational awareness ability and vehicle control ability of the driver, and is calculated according to the off-line determined vehicle stability boundary. The current driver state is used to map the driving burden. The driving skill classification module is used to solve the inconsistency problem of the demand of novice drivers and experienced drivers for the auxiliary system, so as to improve the use rate of the auxiliary system. The skill learning interval classification module is used to obtain the vehicle stability margin and the distance of the vehicle from the lane boundary, and obtain the skill learning interval classification result by using a skill learning interval classification model, the purpose of which is to provide the space for driver skill learning under the premise of ensuring the safety of the vehicle; the adaptive guidance driving right distribution module is used to realize the driving right distribution control according to the driver skill classification result and the skill learning interval classification result, generate the auxiliary driving torque acting on the vehicle steering system, realize the change of the vehicle attitude through the change of the motor torque, and realize the purpose of ensuring the safety of the vehicle and guiding the driver skill learning. The above system can realize the adaptive learning intensity adjustment for the difference of the learning ability of different drivers and the difference of the learning ability of the same driver in different skill stages.

[0049] 1. Driving skill classification

[0050] The driving skill performance is reflected in three aspects of understanding of driving scene, control of vehicle and ability to handle unexpected tasks. The understanding of driving scene is embodied in the decision of driving behavior through situational awareness, such as speed and deceleration point when entering a curve. This is the main reason leading to driving accidents. However, it is impossible to exhaust all traffic scenes (including vehicle speed, curve radius, road surface condition, etc.) in an open environment. Similarly, the evaluation of the control ability of vehicle and the ability to handle unexpected tasks also depends on the scene, so it also faces the problem of inability to exhaust driving scenes. Therefore, the application proposes to take the vehicle stability margin distribution and driving burden in the driving process as the input of the driving skill classification model, and finally realize the classification of driver skill based on the Gaussian mixture-hidden Markov classification model. The architecture of the driving skill classification module is as shown in Figure 2 which has the advantage of realizing accurate driving skill classification through fewer daily driving scenes.

[0051] 1) Stability margin distribution acquisition

[0052] The stable margin distribution acquisition mainly includes four parts: vehicle dynamics construction, root locus graph acquisition, vehicle stability boundary learning, and vehicle stable margin acquisition, as shown in the structure of Figure 3 The vehicle dynamics construction adopts a two-degree-of-freedom vehicle model and a Dugoff tire model to construct a vehicle dynamics model. According to the computing platform capability, a simpler or more complex model can be used, which is not limited to this. The root locus graph is obtained by iterating the 3rd-order Runge-Kutta formula 100 times for each set of states. The states are composed of two parts: the first part is the input, which refers to the front wheel steering angle and the longitudinal vehicle speed; the second part is the initial state, which refers to the vehicle body mass center side slip angle and yaw rate. Secondly, the vehicle stability boundary learning adopts a support vector machine for learning. The specific method is to determine whether the state belongs to a stable point by judging whether the root locus under the state converges to a closed-loop pole. Finally, the vehicle stable margin is obtained by calculating the minimum Euclidean distance from the current state to the vehicle stable boundary under the current state.

[0053] 2) Driving burden acquisition

[0054] The physiological acquisition system is used to obtain the driver's heart rate, skin electrical response, and blood flow, etc. physiological information as the current driver state, and the driving task burden of the driver is evaluated according to these physiological information. This is the main indicator of the driver's response to unexpected situations.

[0055] 3) Gaussian mixture-hidden Markov classification model

[0056] The vehicle stable margin and driving burden obtained above are used as the input of the Gaussian mixture-hidden Markov classification model, and the label of the driver participating in the test is used as the recognition target. When the recognition error of the classification model is stable, the model training is stopped, and the important coefficients of the classification model are saved. The trained Gaussian mixture-hidden Markov classification model is used for actual classification.

[0057] 2, Skill learning interval classification

[0058] The use of the driving skill learning interval classification is to classify dangerous working conditions and safe working conditions, so as to ensure that novice drivers learn driving skills in safe working conditions. It is mainly classified by vehicle stability margin and lane line distance. The method of obtaining the vehicle stability margin is the same as that of obtaining the stability margin distribution in driving skill classification, which is not repeated here. It is pointed out in "Automotive Ergonomics" that drivers will use only 50% of the performance of the vehicle during normal driving, therefore, in the invention, the stable margin ratio is defined as greater than 25% of the stable boundary width, which is defined as a safe area (the stable boundary width is bidirectional symmetric, so the unilateral ratio is 50%), otherwise it is a dangerous area. The lane line distance is determined by the time (TLC) required for the vehicle front wheel to cross the lane line boundary, and the TLC threshold is recommended to be 0.5s in "Design and Experimental Research of Intelligent Vehicle Lane Departure Warning System". Therefore, the skill learning interval classification model designed in this embodiment is as follows:

[0059]

[0060]

[0061] In the formula: C s is the skill learning interval category, 1 represents the learning interval, and 0 represents the non-learning interval; D sl is the distance from the current state to the lower boundary of the vehicle stability boundary, specifically the minimum Euclidean distance; D sh is the distance from the current state to the upper boundary of the vehicle stability boundary; T tlc is the time required for the vehicle front wheel to cross the lane line boundary; D is the width between the lane lines; a and d are the distances from the vehicle center of mass to the front axle and the wheelbase respectively; L is the distance from the vehicle center of mass to the lane center line; θ is the vehicle relative heading angle; v y is the vehicle speed.

[0062] 3. Adaptive guidance of driving right distribution

[0063] The function of the adaptive guidance of driving right distribution module is to reasonably distribute the drivers according to the driver type and vehicle state, and to correct the driving skill learning rate according to the learning ability of the driver. It mainly consists of three parts: a primary driving right distribution unit, an adaptive learning rate adjustment unit and a secondary driving right distribution unit, as shown in the structure Figure 4 .

[0064] 1) Primary driving right distribution

[0065] The first level of driving right allocation is to allocate driving right according to the skill level of the driver, which is mainly used to cope with the different needs of different experienced drivers and novice drivers for different assistance intensity. Here, it is considered that experienced drivers can handle more than 90% of daily working conditions (which can be modified according to specific circumstances), so the only one for experienced drivers starts to work when the stable margin ratio is less than 5% of the stable boundary width, which is defined as a safe area, while the novice drivers are all working. Therefore, the first level of driving right allocation is as follows:

[0066]

[0067]

[0068] wherein: ξ 1o is the first level of driving right allocation coefficient for experienced drivers; ξ 1n is the first level of driving right allocation coefficient for novice drivers; D sl is the distance from the current state to the lower boundary of the vehicle stability boundary; D sh is the distance from the current state to the upper boundary of the vehicle stability boundary.

[0069] 2) Adaptive learning rate adjustment

[0070] The adaptive learning rate adjustment is to realize the adjustment of the skill learning intensity for the differences in learning ability of different novice drivers and the differences in learning ability of the same novice driver at different skill stages. It mainly consists of three parts: assistance consistency rate calculation, artificial potential field function modification, and trajectory tracking and turn following.

[0071] a) Assistance consistency rate calculation

[0072] The assistance consistency rate T cr is defined as the ratio of the time when the direction of the decision-making assistance driving torque is consistent with the direction of the driver's torque in one assistance process, and its calculation formula is as follows:

[0073] t cr = t hs / t ass (5)

[0074] wherein: t cr is the assistance consistency rate; t hs is the time when the direction of the decision-making assistance driving torque is consistent with the direction of the driver's torque in one assistance process; t ass is the total time of the secondary assistance system.

[0075] b) Artificial potential field function modification

[0076] This part is based on the auxiliary consistency rate to correct the path of the auxiliary system planning, which is realized by correcting the artificial potential field function. Experienced drivers have more accurate perception of distance and speed, and have the optimal obstacle avoidance trajectory and corner trajectory. Through large data learning, the repulsive force function of the experienced driver's object and road lane line can be obtained, which is used as the target repulsive force function. Novice drivers have inaccurate perception of distance and speed, and will take more conservative repulsive force function, called learning repulsive force function (i.e. the target repulsive force function is multiplied, and the initial value is 2). The present application determines the artificial potential field function correction coefficient through the auxiliary consistency rate. When the correction coefficient is 1, that is, the learning repulsive force function is consistent with the target repulsive force function, it can be considered that the driving skill learning is completed. The stage learning adjustment is realized by reducing the artificial potential field function correction coefficient by 0.1 when the auxiliary consistency rate is greater than 80%. Therefore, the learning repulsive force function calculation formula is as follows:

[0077]

[0078]

[0079] In the formula, T cr is the auxiliary consistency rate; U n_lane is the learning road repulsive force function; U o_lane is the target road repulsive force function, which is obtained by large data learning; k l is the road repulsive force function correction coefficient; U n_obs is the learning obstacle repulsive force function; U o_obs is the target obstacle repulsive force function, which is obtained by large data learning; k obs is the obstacle repulsive force function correction coefficient.

[0080] c) Trajectory tracking and corner following

[0081] The function of this part is to decide the driving behavior (steering angle and acceleration) required to track the expected trajectory and the guide torque required to follow the expected steering angle. The driving behavior can be obtained by model predictive control, and the guide torque can be obtained by three closed-loop position following control of the motor, and the methods are not limited to this.

[0082] 3) Secondary driving right allocation

[0083] The function of this part is divided into learning driving right allocation and auxiliary system torque decision. Learning driving right allocation is based on the skill learning part on the basis of primary driving right allocation, which realizes the connection of auxiliary driving function between skill learning and risk avoidance. The auxiliary system torque decision is the final output of the motor control torque of the adaptive guide type auxiliary driving system. It changes the vehicle attitude through the change of the motor torque to realize the purpose of ensuring the safety of the vehicle and guiding the skill learning of the driver.

[0084] a) Learning driving right distribution

[0085] Learning driving right distribution is the distribution of driving right according to vehicle stability margin and skill learning zone class, which aims to improve the learning space of different levels for novice drivers on the premise of ensuring safety according to the vehicle stability margin. Vehicle stability margin should be inversely proportional to the assistance intensity, and experienced drivers at this time hope to be assisted. Therefore, the learning driving right distribution is as follows:

[0086] ξ 2o = 1 (8)

[0087]

[0088] In the formula: ξ 2o is the learning driving right distribution coefficient of experienced drivers; ξ 2n is the learning driving right distribution coefficient of novice drivers; C s is the skill learning zone class, 1 represents the learning zone, and 0 represents the non-learning zone.

[0089] b) Assistance system torque decision

[0090] The secondary driving right distribution can be obtained by superimposing the learning driving right distribution and the primary driver right distribution. The secondary driving right distribution is the final permission ratio of the assistance system to control the vehicle state (1 for the assistance system to fully control, and 0 for the driver to fully control), which is realized by decision-making steering motor torque. Since the assistance intensity is proportional to the assistance driving torque, the calculation formula of the assistance driving torque is as follows:

[0091] T o_ass = ξ o * T sw = ξ 1o * ξ 2o * T sw (10)

[0092] T n_ass = ξ n * T sw = ξ 1n * ξ 2n * T sw (11)

[0093] In the formula: T o_ass is the experienced driver's assistance driving torque; ξ o is the experienced driver's secondary driving right distribution coefficient; T n_ass is the novice driver's assistance driving torque; ξ n is the novice driver's secondary driving right distribution coefficient; T swThe desired corner following torque is obtained from the three closed-loop position following control of the motor; ξ 1o The level 1 driving right allocation coefficient for experienced drivers; ξ 1n The level 1 driving right allocation coefficient for novice drivers; ξ 2o The learning driving right allocation coefficient for experienced drivers; ξ 2n The learning driving right allocation coefficient for novice drivers.

[0094] The overall working process of the adaptive guidance type auxiliary driving system includes the following steps:

[0095] 1) According to the vehicle to be equipped with the auxiliary system, change the parameters in the vehicle dynamic model, and then obtain the root locus diagram of the yaw rate and the mass center side slip angle by inputting the front wheel angle and the longitudinal vehicle speed through the 3rd order Runge-Kutta formula iteration.

[0096] 2) The roots in the root locus diagram are classified by using the support vector machine to obtain the classification hyperplane, that is, the vehicle stability boundary. Then, the vehicle stability margin is obtained in real time by calculating the minimum Euclidean distance from the current state to the vehicle stability boundary under the current state. At the same time, the physiological information such as the heart rate, skin electric response and blood flow of the driver is obtained through the physiological acquisition system.

[0097] 3) Set a 5s data observation window, and take the vehicle stability margin data and physiological data in the time window as the input of the Gaussian mixture-hidden Markov model to train the driver skill classification model. Finally, the driver skill classification model outputs the skill level of the current driver, and provides the basis for driving right allocation.

[0098] 4) The skill learning interval classification model classifies the vehicle state by the stability margin and the time (TLC) required for the vehicle front wheels to cross the lane line boundary. The stability margin accounts for more than 25% of the stability boundary width and the TLC is greater than 0.5s, which is defined as the safe area, otherwise it is the dangerous area. The skill learning interval classification model finally outputs the current vehicle state category, that is, the skill learning interval classification category, which provides the basis for driving right allocation.

[0099] 5) The auxiliary consistency rate calculation is the ratio of the time when the direction of the decision auxiliary driving torque is consistent with the direction of the driver torque, which provides the basis for the modification of the artificial potential field function.

[0100] 6) The artificial potential field function is modified by the auxiliary consistency rate calculated in step 5), and the hierarchical learning of driving skills is realized through the adjustment of the desired trajectory. Here, if the auxiliary consistency rate is greater than 80%, the artificial potential field function modification coefficient is reduced by 0.1, that is, the learning driving skill level is improved by one level (one level is divided into 10 levels).

[0101] 7) Obtain the driving behavior (steering wheel angle and acceleration) following the desired trajectory by model predictive control, and then obtain the guide torque required to follow the desired steering wheel angle by three closed-loop position following control of the motor.

[0102] 8) The distribution of driving right is made by the driver skill category obtained in step 3) and the current vehicle state category obtained in step 4). The first level driving right distribution is made according to the driver skill level, which only starts to work for experienced drivers when the stable margin ratio is less than 5% of the stable margin ratio boundary width, while it works for novice drivers. The learning driving right distribution is made according to the vehicle stability margin and the skill learning interval category. Only when the skill learning interval category is the safe interval, the auxiliary system has the driving right, and the vehicle stability margin should be inversely proportional to the auxiliary strength, i.e. the driving right is inversely proportional to the vehicle stability margin.

[0103] 9) The second level driving right distribution is obtained by multiplying the first level driving right distribution and the learning driving right distribution, realizing the connection of the skill learning and risk avoidance functions of the auxiliary driving function.

[0104] 10) The final auxiliary driving torque is obtained by multiplying the second level driving right distribution coefficient obtained in step 9) and the guide torque obtained in step 7), realizing the change of the vehicle attitude by the change of the motor torque to achieve the purpose of ensuring the safety of the vehicle and guiding the driver skill learning.

[0105] The above describes the preferred embodiments of the present application in detail. It should be understood that those skilled in the art can make many modifications and changes without creative labor according to the concept of the present application. Therefore, any technical solution obtained by logical analysis, reasoning or limited experiment based on the prior art according to the concept of the present application shall be within the protection scope determined by the claims.

Claims

1. An adaptive guidance-based driver assistance system that takes into account differences in driver skills, characterized in that, include: The driving skill classification module is used to obtain the current vehicle state and the current driver state, calculate the vehicle stability margin based on the current vehicle state, and use the vehicle stability margin and the current driver state as input to a driver skill classification model to obtain the corresponding driver skill classification result. The vehicle stability margin is calculated based on the vehicle stability boundary determined offline. The skill learning interval classification module is used to obtain the vehicle stability margin and the distance between the vehicle and the lane line boundary. The skill learning interval classification result is obtained by using a skill learning interval classification model. The adaptive guidance driving rights allocation module is used to control driving rights allocation based on the driver skill classification results and skill learning interval classification results, and generate auxiliary driving torques that act on the vehicle steering system. The process of determining the vehicle stability boundary specifically includes: Construct a vehicle dynamics model; Based on the vehicle dynamics model, the root trajectory diagram of the vehicle state change is obtained; The state points in the root trajectory graph are classified and trained by a support vector machine to obtain a classification hyperplane, which is then used as the vehicle stability boundary. The adaptive guided driving rights allocation module includes a primary driving rights allocation unit, an adaptive learning rate adjustment unit, and a secondary driving rights allocation unit, wherein... The Level 1 driving rights allocation unit generates a corresponding Level 1 driving rights allocation coefficient based on the driver skill classification results; The secondary driving authority allocation unit generates a corresponding secondary driving authority allocation coefficient based on the primary driving authority allocation coefficient, the skill learning interval classification result, and the guiding torque obtained by the adaptive learning rate adjustment unit, and generates an auxiliary driving torque based on the secondary driving authority allocation coefficient; The process by which the adaptive learning rate adjustment unit acquires the guiding torque includes: The coefficients of the artificial force field function are corrected based on the consistency rate between the driver's hand torque and the assisted driving torque, and the assisted desired trajectory is adjusted to generate the guiding torque.

2. The adaptive guidance-type driver assistance system considering differences in driver skills according to claim 1, characterized in that, The vehicle dynamics model adopts a two-degree-of-freedom vehicle model and a Dugoff tire model.

3. The adaptive guidance-type assisted driving system considering differences in driver skills according to claim 1, characterized in that, The root locus is obtained by iterating the third-order Runge-Kutta formula for each set of states.

4. The adaptive guidance-type driver assistance system considering differences in driver skills according to claim 1, characterized in that, The skill learning interval classification model is represented as follows: in, This is a skill learning interval category, where 1 represents a learning interval and 0 represents a non-learning interval. It is the distance from the current state to the lower boundary of the vehicle's stability boundary; It is the distance from the current state to the upper boundary of the vehicle's stability boundary; It is the time it takes for the front wheels of a vehicle to cross the lane line boundary.

5. The adaptive guidance-type driver assistance system considering differences in driver skills according to claim 1, characterized in that, The formula for calculating the first-level driving rights allocation coefficient is as follows: in, The first-level driving authority allocation coefficient for experienced drivers; The coefficient for allocating first-level driving rights to novice drivers; It is the distance from the current state to the lower boundary of the vehicle's stability boundary; It is the distance from the current state to the upper boundary of the vehicle's stability boundary.

6. The adaptive guidance-type driver assistance system considering differences in driver skills according to claim 1, characterized in that, Specifically, the correction coefficient of the artificial power field function is as follows: if the auxiliary consistency rate is greater than 80%, the correction coefficient of the artificial power field function is reduced by 0.

1.

7. The adaptive guidance-type driver assistance system considering differences in driver skills according to claim 1, characterized in that, The formula for calculating the secondary driving rights allocation coefficient is as follows: in, The coefficient for allocating secondary driving rights to experienced drivers; The coefficient for allocating secondary driving rights to novice drivers; The first-level driving authority allocation coefficient for experienced drivers; The coefficient for allocating first-level driving rights to novice drivers; The coefficient for allocating learning driving rights to experienced drivers; A coefficient is assigned to novice drivers to grant them the right to learn to drive.

8. The adaptive guidance-type driver assistance system considering differences in driver skills according to claim 7, characterized in that, The formula for calculating the learning driving rights allocation coefficient is as follows: in, This is a skill learning interval category, where 1 represents a learning interval and 0 represents a non-learning interval. It is the distance from the current state to the lower boundary of the vehicle's stability boundary; It is the distance from the current state to the upper boundary of the vehicle's stability boundary.

Images