New energy automobile emergency braking method based on behavior habits of driver

Abstract

The invention discloses a new energy automobile emergency braking method based on driver behavior habits. The new energy automobile emergency braking method comprises the steps that emergency braking test data of different drivers are extracted; s2, performing noise reduction optimization processing on the emergency braking test data of different drivers extracted in the step S1 to obtain data after noise reduction optimization processing; and performing clustering classification on the data after the noise reduction optimization processing through an improved DBSCAN clustering algorithm to obtain emergency braking test data of different classification driver behavior habits, the improved DBSCAN clustering algorithm uses an optimal parameter combination obtained by a self-adaptive parameter selection strategy, an improved clustering distance calculation formula and an evaluation function to perform clustering classification on the data subjected to noise reduction optimization processing; and according to the emergency braking test data of the different classification driver behavior habits, an emergency braking strategy model conforming to the different classification driver behavior habits is established, and emergency braking is achieved. The emergency braking requirements of different drivers can be met, and the driving safety is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent vehicles, and particularly to an emergency braking method for new energy vehicles based on driver behavior habits. Background Art In recent years, with the continuous progress of autonomous driving technology and advanced driver assistance systems (ADAS), the emergency braking system (EBS), as an important part of improving road safety, has received extensive attention. The emergency braking system can quickly apply braking force when the vehicle faces potential collision risks, thereby avoiding or reducing the occurrence of accidents.

[0002] However, traditional emergency braking systems mainly rely on real-time monitoring of the vehicle's surrounding environment, such as sensors like forward radars and cameras, to judge the braking timing, and often cannot fully consider the driver's behavior habits and individual differences. Recent studies have shown that the driver's behavior habits have a significant impact on vehicle control and response. For example, different drivers may have different reaction speeds and operation methods in similar situations, which makes traditional emergency braking strategies may not be applicable to different drivers. Summary of the Invention

[0003] The purpose of the present invention is to provide an emergency braking method for new energy vehicles based on driver behavior habits to meet the emergency braking needs of different drivers and improve driving safety.

[0004] An emergency braking method for new energy vehicles based on driver behavior habits includes: Step S1, extracting emergency braking test data of different drivers; Step S2, performing noise reduction and optimization processing on the emergency braking test data of different drivers extracted in Step S1 to obtain the data after noise reduction and optimization processing; Step S3, performing clustering classification on the data after noise reduction and optimization processing through an improved DBSCAN clustering algorithm to obtain emergency braking test data of different classified driver behavior habits, wherein the improved DBSCAN clustering algorithm performs clustering classification on the data after noise reduction and optimization processing by using the optimal parameter combination obtained by the adaptive parameter selection strategy, as well as the improved clustering distance calculation formula and evaluation function; Step S4, establishing an emergency braking strategy model that conforms to different classified driver behavior habits according to the emergency braking test data of different classified driver behavior habits to achieve emergency braking.

[0005] According to the emergency braking method for new energy vehicles based on driver behavior habits provided by the present invention, the following beneficial effects are achieved: (1)The present invention improves the traditional DBSCAN clustering method. The improved DBSCAN clustering algorithm performs clustering and classification on the data after noise reduction and optimization processing by using the optimal parameter combination obtained by the adaptive parameter selection strategy, as well as the improved clustering distance calculation formula and evaluation function, effectively improving the clustering efficiency and clustering effect of the algorithm. (2)Based on the emergency braking test data of different classified driver behavior habits, the present invention establishes an emergency braking strategy model that conforms to different classified driver behavior habits. Based on the reaction time of the driver during emergency braking, the distance between the host vehicle and the vehicle ahead during emergency braking, and the driver influence coefficient of different classified driver behavior habits, a personalized warning time model is established. Then, based on the early braking time, vehicle braking deceleration, driver influence coefficient, and personalized warning time model, an emergency braking distance model is established, which can effectively improve the braking stability and provide braking strategies adapted to different driver behavior habits.

[0006] (3)The present invention solves the problem of poor adaptive ability of the traditional vehicle emergency braking model, can achieve corresponding emergency braking based on different driving habits of different drivers, improves the safety and stability of new energy vehicles during the braking process, and also improves the comfort of drivers. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] Figure 1 is a schematic flow chart of the emergency braking method for new energy vehicles based on driver behavior habits of the present invention; Figure 2 is a result graph of the first simulation test; Figure 3 is a result graph of the second simulation test. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0008] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Apparently, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0009] Please refer to Figure 1 , the embodiments of the present invention provide an emergency braking method for new energy vehicles based on driver behavior habits, including steps S1 to S4: Step S1, extract the emergency braking test data of different drivers.

[0010] Among them, the emergency braking test data of different drivers is derived from the driving data generated during the daily driving of the drivers.

[0011] Step S2, perform noise reduction and optimization processing on the emergency braking test data of different drivers extracted in Step S1 to obtain the data after noise reduction and optimization processing.

[0012] Specifically, for the emergency braking test data of different drivers extracted in Step S1, use the smoothing method to remove the mutation part and retain the trend information of the data for noise reduction and optimization processing to obtain the data after noise reduction and optimization processing.

[0013] Step S3, perform clustering classification on the data after noise reduction and optimization processing through an improved DBSCAN clustering algorithm to obtain the emergency braking test data of different classified driver behavior habits, where the improved DBSCAN clustering algorithm performs clustering classification on the data after noise reduction and optimization processing using the optimal parameter combination obtained by the adaptive parameter selection strategy, as well as the improved clustering distance calculation formula and evaluation function.

[0014] Among them, Step S3 specifically includes: Step S301, randomly select a parameter combination, where the parameter combination includes the minimum clustering distance threshold and the number of core centers; Step S302, based on the randomly selected parameter combination, calculate the distance between each clustering data and the core center through the improved clustering distance calculation formula; Step S303, construct an evaluation function related to the clustering distance, the minimum silhouette coefficient, the number of algorithm iterations, and the algorithm running time; Step S304, through the evaluation function constructed in Step S303, select the optimal parameter combination and perform cross-validation to obtain the optimal parameter combination; Step S305, use the optimal parameter combination to perform clustering division on the test data.

[0015] Specifically, in Step S302, calculate the distance between each clustering data and the core center through the improved clustering distance calculation formula, and the expression is: ;

[0016] ; Among them, L is the clustering distance, L 1 is the geometric distance from the clustering data to the core center, L 2 is the class distance of the clustering data, k 1 and k 2 are the weighting coefficients, l k is thek The characteristic weight of a nuclear center D d is the d th clustering data M k is the k th nuclear center m is the total number of clustering data n is the total number of nuclear centers; P is the distance parameter. When the distance between the clustering data and the nuclear center is less than or equal to 1, P is equal to 0. When the distance between the clustering data and the nuclear center is greater than 1, P is equal to 1.

[0017] In step S303, the expression of the evaluation function is: ; Among them, E represents the evaluation function C min represents the minimum silhouette coefficient num represents the number of algorithm iterations T d represents the running time of the algorithm , , , are weight coefficients.

[0018] Table 1 compares the clustering data of the traditional DBSCAN clustering method and the improved DBSCAN clustering algorithm proposed by the present invention: Table 1

[0019] As can be seen from Table 1, for the improved DBSCAN clustering algorithm proposed by the present method, the four indicators of clustering distance, minimum silhouette coefficient, number of algorithm iterations, and running time of the algorithm all perform more excellently, and the clustering method of the improved DBSCAN clustering algorithm proposed by the present invention has a better effect.

[0020] In this embodiment, after clustering, four types of drivers are obtained, specifically: Steady drivers, who have the behavior habits of steady drivers. This type of driver has the longest braking distance, the smallest deceleration, and the longest braking time; Relatively steady drivers, who have the behavior habits of relatively steady drivers. This type of driver has a braking distance and time slightly shorter than that of steady drivers, and a deceleration slightly larger than that of steady drivers; Standard drivers, who have the behavior habits of standard drivers. This type of driver has a braking distance and time slightly shorter than that of relatively steady drivers, and a deceleration slightly larger than that of relatively steady drivers; Aggressive drivers, who have aggressive driving behavior habits. This type of driver will complete braking in the shortest distance and the fastest time, with the maximum deceleration.

[0021] Step S4: Based on the emergency braking test data of different classified driver behavior habits, establish an emergency braking strategy model that conforms to different classified driver behavior habits to achieve emergency braking.

[0022] Among them, step S4 specifically includes: Step S401: Obtain the reaction time, pre-braking time, vehicle braking deceleration, and the distance between the own vehicle and the vehicle in front during emergency braking in the emergency braking test data of different classified driver behavior habits, and obtain the driver influence coefficient of different classified driver behavior habits based on the distance between the own vehicle and the vehicle in front during braking. It should be noted that in the specific implementation, when determining the driver influence coefficient, the design principle is that the more aggressive the driving style, the smaller the distance between the own vehicle and the vehicle in front during emergency braking, and the greater the driver influence coefficient.

[0023] Step S402: Based on the reaction time of the driver during emergency braking, the distance between the own vehicle and the vehicle in front during emergency braking, and the driver influence coefficient of different classified driver behavior habits, establish a personalized warning time model. Step S403: Based on the pre-braking time, vehicle braking deceleration, driver influence coefficient, and personalized warning time model, establish an emergency braking distance model. Step S404: Based on the emergency braking distance model, establish an emergency braking strategy model that conforms to different classified driver behavior habits to achieve emergency braking.

[0024] Specifically, in step S402, the expression of the established personalized warning time model is: ; ; Among them, T w is the warning time, f j is the driver influence coefficient corresponding to the j th type of driver behavior habit, is the relative acceleration between the own vehicle and the vehicle in front, t j is the reaction time of the driver corresponding to the j th type of driver behavior habit during emergency braking, V j is the driver reaction speed corresponding to the j th type of driver behavior habit, D jis the distance between the host vehicle and the preceding vehicle when the driver corresponding to the j th driver behavior habit performs an emergency braking, V h is the driving speed of the host vehicle, is the acceleration of the host vehicle. In this embodiment, j takes a value of 1 or 2 or 3 or 4. When j = 1, it corresponds to the driving behavior habit of a steady driver. When j = 2, it corresponds to the driving behavior habit of a relatively steady driver. When j = 3, it corresponds to the driving behavior habit of a standard driver. When j = 4, it corresponds to the driving behavior habit of an aggressive driver.

[0025] In step S403, the expression of the emergency braking distance model is: ; ; ; wherein, D safe represents the emergency braking distance, T j is the pre-braking time corresponding to the j th driver behavior habit, is the vehicle braking deceleration corresponding to the j th driver behavior habit, V f is the driving speed of the preceding vehicle, D 1 represents the driver warning distance.

[0026] The following conducts a simulation test on the present invention: The first simulation test The simulation environment of the simulation experiment is Matlab2020a. The initial speed of the host vehicle is 100 km / h, and the target is that the host vehicle stops at a speed of 0. There are obvious differences in braking performance among drivers with different classified driver behavior habits.

[0027] The results of the simulation test are as Figure 2 shown. As can be seen from Figure 2, the braking distance of steady drivers is the longest, the deceleration is the smallest, and the braking time is the longest; the relatively steady drivers are the second, with their braking distance and time shorter than those of the steady type and the deceleration slightly larger; the braking distance and time of standard drivers are shorter, and the deceleration is larger; aggressive drivers complete braking in the shortest distance and the fastest time, with the largest deceleration. The above content conforms to the classification of driver behavior habits.

[0028] The second simulation test The simulation environment for the simulation experiment is Matlab2020a. The initial speed of the host vehicle is 70 km / h, the speed of the leading vehicle is 40 km / h, and the distance from the leading vehicle is 100 meters. The results of the simulation test are as Figure 3 shown. As can be seen from Figure 3, when faced with a situation that requires braking, the steady driver reacts and starts braking earliest. Relatively speaking, the aggressive driver starts braking latest. The performance of the relatively steady driver and the standard driver is between that of the steady driver and the aggressive driver, and the braking start time and effect gradually fall between the two. Generally speaking, there are significant differences in the reaction time and effect of braking among different types of drivers, and the trend conforms to the classification of driver behavior habits.

[0029] In summary, according to the new energy vehicle emergency braking method based on driver behavior habits provided by the present invention, the following beneficial effects are achieved: (1) The present invention improves the traditional DBSCAN clustering method. The improved DBSCAN clustering algorithm uses the optimal parameter combination obtained by the adaptive parameter selection strategy, as well as the improved clustering distance calculation formula and evaluation function to perform clustering classification on the data after noise reduction and optimization processing, effectively improving the clustering efficiency and clustering effect of the algorithm; (2) According to the emergency braking test data of different classified driver behavior habits, the present invention establishes an emergency braking strategy model that conforms to different classified driver behavior habits. Based on the reaction time of the driver during emergency braking, the distance between the host vehicle and the leading vehicle during emergency braking, and the driver influence coefficient of different classified driver behavior habits, a personalized warning time model is established. Then, based on the early braking time, vehicle braking deceleration, driver influence coefficient, and personalized warning time model, an emergency braking distance model is established, which can effectively improve the braking stability and provide braking strategies adapted to different driver behavior habits.

[0030] (3) The present invention solves the problem of poor adaptive ability of the traditional vehicle emergency braking model. It can realize corresponding emergency braking based on different driving habits of different drivers, improve the safety and stability of new energy vehicles during the braking process, and at the same time improve the comfort of the driver.

[0031] In the description of this specification, the description with reference to terms such as "one embodiment", "some embodiments", "example", "specific example", or "some examples" means that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0032] Although embodiments of the present invention have been shown and described, those of ordinary skill in the art can understand that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the claims and their equivalents.

Claims

1. A new energy vehicle emergency braking method based on driver behavior habits, characterized in that: include: Step S1, extracting emergency braking test data of different drivers; Step S2, performing noise reduction optimization processing on the emergency braking test data of different drivers extracted in step S1 to obtain noise reduction optimized data; Step S3, clustering and classifying the noise reduction optimized data by using the improved DBSCAN clustering algorithm to obtain emergency braking test data of different classifications of driver behavior habits, wherein the improved DBSCAN clustering algorithm clusters and classifies the noise reduction optimized data by using the optimal parameter combination obtained by the adaptive parameter selection strategy, and the improved clustering distance calculation formula and evaluation function; Step S4, based on the emergency braking test data of the driver behavior habits of different categories, an emergency braking strategy model that conforms to the driver behavior habits of different categories is established to implement emergency braking.

2. The new energy vehicle emergency braking method based on driver behavior habits according to claim 1 is characterized in that: Step S3 specifically includes: Step S301, randomly selecting a parameter combination, the parameter combination including a minimum cluster distance threshold and a number of core centers; Step S302, based on the randomly selected parameter combination, the distance between each cluster data and the core center is calculated by using an improved cluster distance calculation formula; Step S303, constructing an evaluation function related to cluster distance, minimum silhouette coefficient, number of algorithm iterations and algorithm running time; Step S304, selecting the optimal parameter combination through the evaluation function constructed in step S303, and performing cross-validation to obtain the optimal parameter combination; Step S305: cluster the test data using the optimal parameter combination.

3. The new energy vehicle emergency braking method based on driver behavior habits according to claim 2 is characterized in that: In step S302, the distance between each cluster data and the core center is calculated by using an improved cluster distance calculation formula, which is expressed as: ； ； in, L is the clustering distance, L 1 is the geometric distance from the cluster data to the core center, L 2 is the class distance of clustered data, k 1 and k 2 is the weighting coefficient, l k It is k The feature weight of the kernel center, D d It is d Cluster data, M k It is k The core center, m is the total number of clustered data, n is the total number of nuclear centers; P is the distance parameter. When the distance between the cluster data and the core center is less than or equal to 1, P is equal to 0, when the distance between the cluster data and the core center is greater than 1, P is equal to 1; In step S303, the expression of the evaluation function is: ； in, E represents the evaluation function, C min represents the minimum silhouette coefficient, num represents the number of algorithm iterations, T d represents the algorithm running time, , , , is the weight coefficient.

4. The new energy vehicle emergency braking method based on driver behavior habits according to claim 1 is characterized in that: Step S4 specifically includes: Step S401, obtaining the driver's reaction time during emergency braking, advance braking time, vehicle braking deceleration, and the distance between the vehicle and the vehicle in front during emergency braking from the emergency braking test data of the driver's behavior habits of different categories, and obtaining the driver influence coefficient of the driver's behavior habits of different categories according to the distance between the vehicle and the vehicle in front during braking; Step S402, establishing a personalized warning time model based on the driver's reaction time during emergency braking, the distance between the vehicle and the preceding vehicle during emergency braking, and the driver influence coefficients of different driver behavior habits; Step S403, establishing an emergency braking distance model based on the advance braking time, the vehicle braking deceleration, the driver influence coefficient and the personalized warning time model; Step S404: based on the emergency braking distance model, an emergency braking strategy model that meets the behavior habits of drivers of different categories is established to implement emergency braking.

5. The new energy vehicle emergency braking method based on driver behavior habits according to claim 4 is characterized in that: In step S402, the expression of the established personalized warning time model is: ； ； in, T w It's warning time. f j It corresponds to j The driver influence coefficient of the driver behavior habits, is the relative acceleration between the ego vehicle and the preceding vehicle, t j It corresponds to j The reaction time of drivers with different driving behavior habits during emergency braking, V j It corresponds to j The driver's reaction speed of the driver's behavior habits, D j It corresponds to j The distance between the driver's vehicle and the vehicle ahead during emergency braking for the driver with this type of driver behavior habit is V h is the vehicle's speed, is the vehicle acceleration.

6. The new energy vehicle emergency braking method based on driver behavior habits according to claim 5 is characterized in that: In step S403, the expression of the emergency braking distance model is: ； ； ； in, D safe Indicates the emergency braking distance. T j It corresponds to j The advance braking time of the driver's behavior habits, It corresponds to j The vehicle braking deceleration of the driver's behavior habits, V f is the speed of the preceding vehicle, D 1 indicates the driver’s warning distance.

Images