Vehicle emergency brake prediction method and device, storage medium and equipment

Abstract

The invention discloses a vehicle emergency brake prediction method and device, a storage medium and equipment. The method comprises the steps of firstly obtaining driving data of a target vehicle; wherein the driving data comprises vehicle condition information of the target vehicle, environment data information of the target vehicle, traffic road condition information and road history information of a city where the target vehicle is located, then carrying out standardization processing on the driving data of the target vehicle, and extracting a target feature vector corresponding to the processed driving data; and inputting the target feature vector into a pre-constructed vehicle emergency brake prediction model, and predicting to obtain a prediction result of emergency brake of the target vehicle. And when the prediction result shows that the target vehicle is about to be braked suddenly, the emergency brake early warning prompt information is sent to the vehicle-mounted passengers of the target vehicle, so that the passengers can make psychological expectations in advance, the dizziness of the passengers can be reduced, and the riding experience and riding safety of the passengers are improved.

Description

Technical Field

[0001] The present application relates to the field of vehicle technology, and in particular to a vehicle emergency braking prediction method, device, storage medium and equipment. Background Art

[0002] With the improvement of people's living standards and the rapid development of social economy, the use rate of cars has gradually increased, and people's requirements for the riding experience have become higher and higher. However, traditional cars mainly focus on vehicle driving safety and performance, but pay relatively little attention to the passengers on board.

[0003] For example, in the current driving process of a vehicle, other passengers in the vehicle, except the driver, often do not know when the driver will make an emergency brake operation, which may cause them to feel motion sickness due to the driver's emergency brake operation. This feeling of motion sickness is caused by the passenger's body not anticipating the upcoming movement changes, which causes the brain to be confused about the body's movement perception, thereby causing motion sickness symptoms. In addition, the driver's emergency brake operation may also cause other passengers in the vehicle to rush forward due to inertia, which not only brings sudden fright and psychological pressure to the passengers, but may also cause the passengers to hit some hard objects inside the vehicle, causing injuries or even falls, thereby affecting the passengers' riding experience and riding safety. Summary of the invention

[0004] The main purpose of the embodiments of the present application is to provide a vehicle emergency braking prediction method, device and equipment, which can predict the driver's upcoming emergency braking operation in advance and send corresponding prompt information to the passengers on the vehicle so that the passengers can prepare psychological expectations in advance, thereby reducing the occurrence of dizziness and improving the passengers' riding experience and riding safety.

[0005] The present application embodiment provides a vehicle emergency braking prediction method, comprising:

[0006] Acquire the driving data of the target vehicle; the driving data includes the vehicle condition information of the target vehicle, the environmental data information of the target vehicle, the traffic condition information and the road history information of the city where the target vehicle is located;

[0007] Standardizing the driving data of the target vehicle and extracting a target feature vector corresponding to the processed driving data;

[0008] Inputting the target feature vector into a pre-built vehicle emergency braking prediction model to predict the emergency braking of the target vehicle;

[0009] When the prediction result of the sudden braking of the target vehicle indicates that the target vehicle is about to make an sudden braking, an emergency braking warning prompt message is sent to passengers on board the target vehicle.

[0010] In a possible implementation, the standardizing the driving data of the target vehicle and extracting the target feature vector corresponding to the processed driving data includes:

[0011] The driving data of the target vehicle is normalized to obtain normalized driving data; and the normalized driving data are spliced ​​to form a target feature vector corresponding to the processed driving data.

[0012] In a possible implementation, the inputting the target feature vector into a pre-built vehicle emergency braking prediction model to predict the emergency braking of the target vehicle includes:

[0013] Inputting the target feature vector into a pre-built vehicle emergency braking prediction model, and calculating the distance between the target feature vector and the average feature vector;

[0014] According to the value of the distance, a prediction result of the sudden braking of the target vehicle is determined.

[0015] In a possible implementation, calculating the distance between the target feature vector and the average feature vector includes:

[0016] The chi-square distance between the target feature vector and the average feature vector is calculated.

[0017] In a possible implementation, calculating the chi-square distance between the target feature vector and the average feature vector includes:

[0018] The information gain of each type of data contained in the driving data is calculated, and the chi-square distance between the target feature vector and the average feature vector is optimized by using the information gain to obtain a more accurate calculation result.

[0019] In a possible implementation, the vehicle emergency braking prediction model is constructed as follows:

[0020] Acquire driving data of a sample vehicle; the driving data of the sample vehicle includes vehicle condition information of the sample vehicle, environmental data information of the sample vehicle, traffic condition information, and road history information of the city where the sample vehicle is located;

[0021] Standardizing the driving data of the sample vehicles and extracting sample feature vectors corresponding to the processed sample driving data;

[0022] Calculate the average of all sample feature vectors as the average feature vector;

[0023] The initial vehicle emergency braking prediction model is trained according to the sample feature vector, the average feature vector and the prediction label of emergency braking corresponding to the driving data of the sample vehicle to generate the vehicle emergency braking prediction model.

[0024] In one possible implementation, the initial vehicle emergency braking prediction model is a machine learning model that performs binary classification prediction on data.

[0025] In a possible implementation, the method further includes:

[0026] Acquire the driving data of the verification vehicle; the driving data of the verification vehicle includes the vehicle condition information of the verification vehicle, the environmental data information of the verification vehicle, the traffic condition information and the road history information of the city where the verification vehicle is located;

[0027] Standardizing the driving data of the verification vehicle and extracting a verification feature vector corresponding to the processed verification driving data;

[0028] Inputting the verification feature vector feature into the vehicle emergency braking prediction model to obtain the prediction result of emergency braking corresponding to the driving data of the verification vehicle;

[0029] When the prediction result of sudden braking corresponding to the driving data of the verification vehicle is inconsistent with the sudden braking marking result corresponding to the driving data of the verification vehicle, the driving data of the verification vehicle is re-used as the driving data of the sample vehicle to update the vehicle sudden braking prediction model.

[0030] The embodiment of the present application also provides a vehicle emergency braking prediction device, the device comprising:

[0031] A first acquisition unit is used to acquire driving data of a target vehicle; the driving data includes vehicle condition information of the target vehicle, environmental data information of the target vehicle, traffic condition information, and road history information of the city where the target vehicle is located;

[0032] A first extraction unit, used for performing standardization processing on the driving data of the target vehicle and extracting a target feature vector corresponding to the processed driving data;

[0033] A prediction unit, used for inputting the target feature vector into a pre-built vehicle emergency braking prediction model to predict the emergency braking of the target vehicle;

[0034] The early warning unit is used to send an early warning prompt message of an emergency brake to passengers of the target vehicle when the prediction result of the emergency brake of the target vehicle indicates that the target vehicle is about to brake suddenly.

[0035] In a possible implementation, the first extraction unit is specifically configured to:

[0036] The driving data of the target vehicle is normalized to obtain normalized driving data; and the normalized driving data are spliced ​​to form a target feature vector corresponding to the processed driving data.

[0037] In a possible implementation, the prediction unit includes:

[0038] A calculation subunit, used for inputting the target feature vector into a pre-built vehicle emergency braking prediction model, and calculating the distance between the target feature vector and the average feature vector;

[0039] The prediction subunit is used to determine the prediction result of the sudden braking of the target vehicle according to the value of the distance.

[0040] In a possible implementation, the computing subunit is specifically used for:

[0041] The chi-square distance between the target feature vector and the average feature vector is calculated.

[0042] In a possible implementation, the computing subunit is specifically used for:

[0043] The information gain of each type of data contained in the driving data is calculated, and the chi-square distance between the target feature vector and the average feature vector is optimized by using the information gain to obtain a more accurate calculation result.

[0044] In a possible implementation, the device further includes:

[0045] A second acquisition unit is used to acquire driving data of a sample vehicle; the driving data of the sample vehicle includes vehicle condition information of the sample vehicle, environmental data information of the sample vehicle, traffic condition information, and road history information of the city where the sample vehicle is located;

[0046] A second extraction unit, used for performing standardization processing on the driving data of the sample vehicle and extracting a sample feature vector corresponding to the processed sample driving data;

[0047] A calculation unit, used for calculating the average value of all sample feature vectors as an average feature vector;

[0048] A training unit is used to train an initial vehicle emergency braking prediction model according to the sample feature vector, the average feature vector and the prediction label of emergency braking corresponding to the driving data of the sample vehicle to generate the vehicle emergency braking prediction model.

[0049] In one possible implementation, the initial vehicle emergency braking prediction model is a machine learning model that performs binary classification prediction on data.

[0050] In a possible implementation, the device further includes:

[0051] A third acquisition unit is used to acquire the driving data of the verification vehicle; the driving data of the verification vehicle includes the vehicle condition information of the verification vehicle, the environmental data information of the verification vehicle, the traffic condition information and the road history information of the city where the verification vehicle is located;

[0052] A third extraction unit, used for performing standardization processing on the driving data of the verification vehicle and extracting a verification feature vector corresponding to the processed verification driving data;

[0053] An obtaining unit, used for inputting the verification feature vector feature into the vehicle emergency braking prediction model to obtain a prediction result of emergency braking corresponding to the driving data of the verification vehicle;

[0054] An updating unit is used to update the vehicle emergency braking prediction model by reusing the driving data of the verification vehicle as the driving data of the sample vehicle when the prediction result of emergency braking corresponding to the driving data of the verification vehicle is inconsistent with the emergency braking marking result corresponding to the driving data of the verification vehicle.

[0055] The embodiment of the present application also provides a vehicle emergency braking prediction device, including: a processor, a memory, and a system bus;

[0056] The processor and the memory are connected via the system bus;

[0057] The memory is used to store one or more programs, and the one or more programs include instructions. When the instructions are executed by the processor, the processor executes any one of the implementation methods of the above-mentioned vehicle emergency braking prediction method.

[0058] An embodiment of the present application also provides a computer-readable storage medium, in which instructions are stored. When the instructions are executed on a terminal device, the terminal device executes any one of the implementation methods of the above-mentioned vehicle emergency braking prediction method.

[0059] An embodiment of the present application also provides a computer program product. When the computer program product is run on a terminal device, the terminal device executes any one of the implementations of the above-mentioned vehicle emergency braking prediction method.

[0060] The embodiment of the present application provides a method, device, storage medium and equipment for predicting sudden braking of a vehicle. First, the driving data of the target vehicle is obtained; wherein the driving data includes the vehicle condition information of the target vehicle, the environmental data information of the target vehicle, the traffic road condition information and the road history information of the city where the target vehicle is located. Then, the driving data of the target vehicle is standardized and the target feature vector corresponding to the processed driving data is extracted. Then, the target feature vector is input into a pre-built vehicle sudden braking prediction model to predict the prediction result of sudden braking of the target vehicle. Furthermore, when the prediction result of sudden braking of the target vehicle indicates that the target vehicle is about to brake suddenly, an emergency braking warning prompt message is sent to the passengers on board the target vehicle.

[0061] It can be seen that since the present application can use the pre-built vehicle emergency braking prediction model to process the collected driving data of the target vehicle during the driving of the target vehicle, so as to predict in advance the emergency braking operation that the driver of the target vehicle is about to perform, and send corresponding early warning prompt information to the passengers on board, so that the passengers can prepare psychological expectations in advance, thereby reducing the occurrence of dizziness and improving the passengers' riding experience and safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0063] Figure 1 A schematic diagram of a process flow of a vehicle emergency braking prediction method provided in an embodiment of the present application;

[0064] Figure 2 A schematic diagram of the composition of a vehicle emergency braking prediction device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0065] With the popularization of automobiles, people have higher and higher requirements for vehicle intelligence, comfort and safety. At present, although the in-vehicle intelligent system can provide some basic safety tips for in-vehicle users, such as speeding warnings, fatigue driving warnings, etc., most systems are not specially designed and optimized for the problem of passenger motion sickness. For example, during the driving of the vehicle, for other passengers in the vehicle except the driver, they often do not know when the driver will make an emergency brake operation, which may cause them to feel motion sickness due to the driver's emergency brake. This feeling of motion sickness is due to the passenger's body not anticipating the upcoming movement changes, causing the brain to confuse the body's movement perception, thereby causing motion sickness symptoms. In addition, the driver's emergency brake operation may also cause other passengers in the vehicle to rush forward due to inertia, which not only brings sudden fright and psychological pressure to the passengers, but also may cause the passengers to hit some hard objects inside the vehicle, causing injuries or even falls, thereby affecting the passengers' riding experience and riding safety.

[0066] Moreover, some existing prediction systems mainly rely on the vehicle's speed and acceleration for prediction, and often ignore other important factors such as the vehicle's braking time and braking amplitude. Therefore, the accuracy and practicality of their predictions need to be improved.

[0067] In order to solve the above defects, the present application provides a method for predicting vehicle emergency braking, which first obtains the driving data of the target vehicle; wherein the driving data includes the vehicle condition information of the target vehicle, the environmental data information of the target vehicle, the traffic road condition information and the road history information of the city where the target vehicle is located, and then the driving data of the target vehicle is standardized, and the target feature vector corresponding to the processed driving data is extracted; then the target feature vector is input into the pre-built vehicle emergency braking prediction model to predict the prediction result of the target vehicle emergency braking. Furthermore, when the prediction result of the target vehicle emergency braking indicates that the target vehicle is about to emergency brake, an emergency braking warning prompt message is sent to the passengers on board the target vehicle.

[0068] It can be seen that since the present application can use the pre-built vehicle emergency braking prediction model to process the collected driving data of the target vehicle during the driving of the target vehicle, so as to predict in advance the emergency braking operation that the driver of the target vehicle is about to perform, and send corresponding early warning prompt information to the passengers on board, so that the passengers can prepare psychological expectations in advance, thereby reducing the occurrence of dizziness and improving the passengers' riding experience and safety.

[0069] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0070] First embodiment

[0071] See also Figure 1 , is a flow chart of a vehicle emergency braking prediction method provided in this embodiment, the method comprising the following steps:

[0072] S101: Acquire driving data of a target vehicle; wherein the driving data includes vehicle condition information of the target vehicle, environmental data information of the target vehicle, traffic condition information, and road history information of the city where the target vehicle is located.

[0073] In this embodiment, any vehicle for which emergency braking prediction is performed by this embodiment is defined as a target vehicle. In order to improve the emergency braking prediction effect of the target vehicle and improve the riding experience and riding safety of the passengers on the vehicle, firstly, it is necessary to obtain the driving data of the target vehicle during the driving process by using various sensors pre-installed on the target vehicle, so as to perform the subsequent step S102.

[0074] Among them, the driving data of the target vehicle may include but is not limited to the vehicle condition information of the target vehicle, the environmental data information of the target vehicle, the traffic road condition information, and the road history information of the city where the target vehicle is located. Among them, the vehicle condition information of the target vehicle may include but is not limited to the driving speed, acceleration, braking amplitude, etc. The environmental data information of the target vehicle may include but is not limited to the current time, rainfall, temperature, humidity, etc. Traffic road condition information may include but is not limited to real-time information such as the distance between the target vehicle and the vehicle in front, the distance from the intersection / exit, and the degree of traffic congestion. The road history information of the city where the target vehicle is located may include but is not limited to the longitude and latitude distance of the historical high-frequency emergency brakes of the current city.

[0075] Furthermore, it should be noted that the present application does not limit the method for obtaining the driving data of the target vehicle, which can usually be obtained (when the driver releases the accelerator or in real time) using various sensors pre-installed on the target vehicle and other vehicle-mounted systems. For example, weather information can be obtained through the weather application of the target vehicle's vehicle-mounted system, while data such as the degree of traffic congestion and the distance to the intersection / exit can be obtained through the target vehicle's vehicle-mounted map application. The distance between the target vehicle and the vehicle in front can be measured by the target vehicle's front radar. The braking amplitude can be read directly from the target vehicle's brake sensor.

[0076] In addition, it should be noted that the present application does not limit the specific structure, quantity and installation location of each sensor on the target vehicle. For example, the target vehicle can be pre-installed with a vehicle speed sensor (VSS), an acceleration sensor and other sensors to obtain various driving data such as the target vehicle's driving speed and acceleration, and the number of each sensor can be one or more.

[0077] S102: Standardizing the driving data of the target vehicle and extracting a target feature vector corresponding to the processed driving data.

[0078] In this embodiment, after obtaining the target vehicle's vehicle condition information, the target vehicle's environmental data information, traffic condition information, and the city's road history information and other driving data through step S101, since the units of each type of data contained in these driving data are different, there may be different scales. In order to facilitate the subsequent improvement of the prediction effect of the sudden braking of the target vehicle, after obtaining these driving data, each type of data contained in these driving data can be further standardized using existing or future data standardization methods, so that each type of data is on the same scale. For example, each type of data contained in the driving data can be normalized to obtain normalized driving data.

[0079] For example, when the target vehicle's speed is in kilometers per hour, 0 can be converted to 0, 150 to 1, and when it exceeds 150, it can also be converted to 1. Other numbers are normalized proportionally, and the speed can be normalized to a number between 0 and 1. When the target vehicle's acceleration is in meters per second, 2When the target vehicle is at a certain time, 0 can be converted to 0, 10 can be converted to 1, and when it exceeds 10, it can also be converted to 1. Other numbers are normalized in proportion, that is, the acceleration can be normalized to a number between 0 and 1. Since the longitude and latitude are only used to calculate the distance from the longitude and latitude of high-frequency emergency brakes, they are not used as characteristic values. Specifically, through the previous data collection, the longitude and latitude of the TOP N (the value is not limited, and the value can be 100, etc.) emergency brake points of each city, district and county are recorded. The longitude and latitude are accurate to 10 meters. Each data record the distance from the high-frequency brake point at the current moment. When the unit is meter, 0 can be converted to 0, 1000 can be converted to 1, and when it exceeds 1000, it can also be converted to 1. When the unit of the distance between the target vehicle and the preceding vehicle and the distance between the intersection / exit is meter, 0 can be converted to 0, 200 can be converted to 1, and when it exceeds 200, it can also be converted to 1. For time, 0 o'clock can be converted to 0, 24 o'clock can be converted to 1, and the value can be 0 to a number infinitely close to 1. If the unit of rainfall data is millimeters per hour, 0 can be converted to 0, 20 can be converted to 1, and when it exceeds 20, it can also be converted to 1. When the temperature unit is ℃, -20 can be converted to 0, and below -20 can also be converted to 0, 50 can be converted to 1, and above 50 can also be converted to 1. For humidity, 0% can be converted to 0, and 100% can be converted to 1. Since the level of traffic congestion varies from place to place, 4 to 5 equally divided numbers between 0 and 1 can be taken according to the level, such as 0, 0.25, 0.5, 0.75, and 1. For the braking amplitude, the driver's failure to brake can be regarded as 0, and the rapid depressing of the brake pedal to the bottom can be regarded as 1. The intensity and braking amplitude can be converted to a number between 0 and 1.

[0080] Furthermore, after the driving data of the target vehicle is standardized (such as normalized) to obtain the processed driving data, the existing or future data feature extraction methods can be further used to extract the feature vector corresponding to the processed driving data (defined here as the target feature vector). For example, the normalized driving data can be spliced ​​to form the target feature vector corresponding to the processed driving data, etc., to execute the subsequent step S103.

[0081] For example, if the target vehicle's driving data contains the following types of data: the current time is 12 o'clock, the driving speed is 60 kilometers per hour, and the acceleration is 0 meters per second 2 , the high-frequency emergency braking latitude and longitude distance is 500 meters, the distance between the target vehicle and the front vehicle is 10 meters, the intersection / exit distance is more than 200 meters, the rainfall is 0 mm / hour, the temperature is 26℃, the humidity is 70%, the traffic congestion is basically smooth, and the braking amplitude is 50%. Then it is normalized and the normalized driving data is spliced ​​to form the target feature vector: [0.5 (current time: 12 o'clock), 0.4 (driving speed: 60 kilometers / hour), 0 (acceleration: 0 meters / second 2), 0.5 (high-frequency emergency braking latitude and longitude distance: 500 meters), 0.05 (distance between the target vehicle and the front vehicle: 10 meters), 1 (intersection / exit distance: more than 200 meters), 0 (rainfall: 0 mm / hour), 0.657 (temperature: 26℃), 0.7 (humidity: 70%), 0.25 (traffic congestion level: basically unobstructed), 0.5 (braking amplitude: 50%)], that is, [0.5, 0.4, 0, 0.5, 0.05, 1, 0, 0.657, 0.7, 0.25, 0.5].

[0082] It should be noted that the embodiments of the present application do not limit the method for extracting the target feature vector, nor do they limit the specific extraction process. An appropriate extraction method may be selected according to actual conditions, and corresponding feature extraction operations may be performed.

[0083] S103: Inputting the target feature vector into a pre-built vehicle emergency braking prediction model to obtain a prediction result of emergency braking of the target vehicle.

[0084] In this embodiment, after extracting the target feature vector corresponding to the processed driving data through step S102, in order to improve the efficiency and accuracy of emergency braking prediction for the target vehicle, the obtained target feature vector can be further input into a pre-built vehicle emergency braking prediction model to predict the emergency braking result of the target vehicle, so as to execute the subsequent step S104.

[0085] It should be noted that the specific composition and specific detection process of the vehicle emergency braking prediction model can be set according to actual conditions, and this embodiment does not limit this. The model training method used in this embodiment will be described in detail later, and will not be repeated here.

[0086] Specifically, an optional implementation manner is that the implementation process of step S103 may include the following steps AB:

[0087] Step A: Input the target feature vector into the pre-built vehicle emergency braking prediction model and calculate the distance between the target feature vector and the average feature vector.

[0088] It should be noted that when constructing a vehicle emergency braking prediction model, in order to improve the prediction accuracy of the model, this embodiment will construct a feature vector corresponding to each sample driving data after obtaining a large number of historical driving data of vehicles in braking scenarios as sample driving data (here it is defined as a sample feature vector, and the specific construction method can refer to the construction method of the target feature vector in the above step S1032), and then calculate the average value of all sample feature vectors as the average feature vector, which is used as an indirect measurement basis for predicting whether the target vehicle will brake suddenly next.

[0089] Specifically, after determining the target feature vector, the target feature vector can first be input into a pre-built vehicle emergency braking prediction model, and the distance between the target feature vector and the average feature vector can be calculated using an existing or future vector distance calculation formula, such as calculating the chi-square distance (Chi-square Measure) or Euclidean distance between the target feature vector and the average feature vector, as the distance between the target feature vector and the average feature vector. It should be noted that the chi-square distance between the target feature vector and the average feature vector will be introduced later as the distance between the target feature vector and the average feature vector.

[0090] Among them, the calculation method of the chi-square measure is to use the contingency table analysis method to obtain a chi-square statistic to measure the difference between two vectors. The larger the chi-square statistic, the more significant the choice of vector has on the value of the variable, that is, the greater the difference between the two vectors. The specific calculation formula is as follows:

[0091]

[0092] Where n represents the dimensions of x and y; x i represents the value of the i-th dimension of vector x; E(x i ) represents the expected value (average value) of vector x at the i-th dimension; y i represents the value of the i-th dimension of vector y; E(y i ) represents the expected value (average value) of vector y at the i-th dimension.

[0093] Based on this, the vector y in the above formula (1) can be replaced by the average eigenvector, that is, y i =E(y i )=E(x i ), the above formula (1) can be converted into the following formula (2), which is used to express the calculation formula of the chi-square distance between the target feature vector and the average feature vector:

[0094]

[0095] Where d(x,y) represents the chi-square distance between the target feature vector and the average feature vector; x i represents the value of the i-th dimension of the target feature vector x; y i represents the value of the i-th dimension of the average feature vector y; n represents the dimension of the target feature vector x and the average feature vector y.

[0096] It should be noted that the chi-square distance between the target feature vector and the average feature vector is used to characterize the distance between the target feature vector and the average feature vector. The smaller the distance, the closer the target feature vector is to the average feature vector, that is, the greater the possibility of the sighted vehicle braking. However, when the chi-square distance formula is used to calculate the distance between the target feature vector and the average feature vector, the weights of each dimension are the same, and in fact, the determining factors when the driver of each vehicle brakes may be different, that is, the influence of different types of data contained in the vehicle's driving data on the model prediction results is also different. Therefore, after calculating the information gain of each type of data contained in the driving data, this embodiment uses the information gain to optimize the chi-square distance between the target feature vector and the average feature vector, that is, to optimize the above formula (2). Specifically, before the accumulation in formula (2), each dimension is multiplied by the corresponding weight (that is, the information gain of the type of data) to obtain a more accurate calculation result.

[0097] The information gain of each type of data contained in the driving data is actually for each type of data. That is to say, for a certain type of data, the difference in the overall information entropy when it is included and when it is not included is the amount of information that this type of data brings to the whole. For a type of data (represented by T), the calculation formula for its information gain (represented by IG(T)) is as follows:

[0098] IG(T)=H(C)-H(C|T) (3)

[0099] Among them, H(C) represents the entropy of the category (in this application, the categories are only two categories: sudden braking and non-sudden braking); H(C|T) represents the conditional entropy of category C when the feature is T.

[0100] Assume that category C has m different values ​​C1, C2, ..., C m , then the calculation formula of H(C) is as follows:

[0101]

[0102] Among them, P(C j ) means C j Probability of occurrence.

[0103] According to the conditional probability formula, the calculation formula of H(C|T) is as follows:

[0104]

[0105] Among them, P(t) represents the probability of T appearing; represents the probability that T does not appear; P(C j |t) indicates that when T type data appears, C jProbability of occurrence; Indicates that C does not appear when T type data j Probability of occurrence.

[0106] On this basis, by substituting the above formulas (4) and (5) into the above formula (3), the calculation formula of the information gain IG(T) corresponding to the T type data can be obtained as follows:

[0107]

[0108] It can be understood that in this embodiment, since there are only two categories, namely, emergency braking and non-emergency braking, that is, the value of m is 2, there are only two types, emergency braking C1 and non-emergency braking C2. Accordingly, P(C1|t) and P(C2|t) respectively represent the probability of the driver braking suddenly and not braking suddenly when considering the T type data. They represent the probability of the driver braking suddenly and not braking suddenly when the T type data is not considered.

[0109] It should be noted that the present application defines P(C1), P(C2), P(C1|t), P(C2|t), The value of is not limited, and can be calculated using existing vehicle operation data according to actual conditions and experience. An optional implementation method is that the method adopted in this embodiment is to obtain it by using the driving data statistics of a large number of sample vehicles obtained when building a vehicle emergency braking prediction model.

[0110] In this way, after calculating the information gain of each type of data contained in the driving data, the information gain of each type of data can be further used to optimize the chi-square distance between the target feature vector and the average feature vector. That is, substitute the above formula (6) into the above formula (2) to obtain the calculation formula of the optimized chi-square distance between the target feature vector and the average feature vector as follows:

[0111]

[0112] Among them, IG(x i ) represents the value x of the i-th dimension of the target feature vector x i The information gain of the corresponding type of data. d(x,y)1 represents the optimized chi-square distance between the target feature vector and the average feature vector.

[0113] Step B: Determine the prediction result of the target vehicle's sudden braking based on the chi-square distance between the target feature vector and the average feature vector.

[0114] In this implementation, after determining the (optimized) chi-square distance between the target feature vector and the average feature vector in step A, the prediction result of the target vehicle's sudden braking is further determined based on the value of the chi-square distance. Specifically, the smaller the value of the chi-square distance, the greater the possibility of the target vehicle's sudden braking. For the vehicle sudden braking prediction model, it can output a value between 0 and 1 to characterize the predicted result of the target vehicle's sudden braking. The closer the model output prediction result is to 0, the smaller the possibility of the target vehicle's sudden braking; conversely, the closer it is to 1, the greater the possibility of the target vehicle's sudden braking.

[0115] S104: When the prediction result of the target vehicle's sudden braking indicates that the target vehicle is about to make an sudden braking, an emergency braking warning message is sent to passengers in the target vehicle.

[0116] In this embodiment, after the prediction result of the target vehicle's emergency braking is obtained through step S103, when the prediction result indicates that the target vehicle is about to brake suddenly, if the model outputs a prediction result that is greater than a preset threshold (the specific value is not limited and can be set according to actual conditions and experience values, such as 0.5) and is close to 1 (such as 0.8), this indicates that there is a high possibility of an emergency braking, indicating that corresponding measures need to be taken as soon as possible to remind the passengers on the vehicle that an emergency braking is about to occur, so that the passengers can prepare psychological expectations in advance, thereby reducing the occurrence of dizziness and improving the passengers' riding experience and riding safety.

[0117] Specifically, the present application does not limit the form and specific content of the emergency brake warning information sent to the passengers on the vehicle, and can be set according to the actual situation. For example, the emergency brake warning voice can be played to the passengers on the target vehicle in the form of voice broadcast, such as "The vehicle is about to brake suddenly, please pay attention to protection"; and / or, the emergency brake warning text can be displayed to the passengers on the target vehicle in the form of text on the large screen on the vehicle, such as "The vehicle is about to brake suddenly, please pay attention to protection"; and / or, the alarm pre-installed on the vehicle can be used to send a flashing light alarm signal for warning, etc.

[0118] Next, this embodiment will introduce the construction process of the vehicle emergency braking prediction model used in this application, which may specifically include the following steps (1)-(4):

[0119] Step (1): Obtain driving data of a sample vehicle; wherein the driving data of the sample vehicle includes vehicle condition information of the sample vehicle, environmental data information of the sample vehicle, traffic condition information, and road history information of the city where the sample vehicle is located.

[0120] In this embodiment, in order to build a vehicle emergency braking prediction model, a lot of preparatory work needs to be done in advance. First, it is necessary to collect and obtain various vehicle driving data recorded every time the driver releases the accelerator (such as the vehicle's condition information, the vehicle's environmental data information, traffic conditions information, and the city's road history information, etc.) and the maximum braking amplitude in the next 5 seconds, and use these driving data of the vehicle obtained each time as a sample vehicle's driving data (including the sample vehicle's condition information, the sample vehicle's environmental data information, traffic conditions information, and the city's road history information, etc.), and use the corresponding braking amplitude as a prediction label to train the vehicle emergency braking prediction model.

[0121] Step (2): Standardize the driving data of the sample vehicles and extract the sample feature vector corresponding to the processed sample driving data.

[0122] In this embodiment, after the driving data of the sample vehicle is obtained through step (1), it cannot be directly used to train and generate a vehicle emergency braking prediction model. Instead, the driving data of the sample vehicle needs to be standardized (such as normalized) and the sample feature vector corresponding to the processed sample driving data is extracted. The specific implementation process can refer to the processing process of the driving data of the target vehicle in the above step S102, and the sample feature vector corresponding to each processed (such as normalized) sample driving data can be obtained.

[0123] Step (3): Calculate the average of all sample feature vectors as the average feature vector.

[0124] In this embodiment, after obtaining the sample feature vector corresponding to each processed (such as normalized) sample driving data through step (2), the average value of the sample feature vectors under all braking scenarios can be further calculated as the average feature vector. It should be noted here that although the sample feature vectors of all historical driving data may have been calculated, the calculation of the average feature vector will only use the vector situation under the braking scenario as the standard. When taking the distance threshold, the feature vector of each data will still be used to calculate the distance with the average feature vector. Finally, by calculating the distance between the feature vector of each data and the average feature vector, the possibility of the driver braking in this time period is determined.

[0125] Step (4): Train the initial vehicle emergency braking prediction model according to the sample feature vector, the average feature vector and the prediction label of emergency braking corresponding to the driving data of the sample vehicle to generate a vehicle emergency braking prediction model.

[0126] In this embodiment, after obtaining the sample feature vector, the average feature vector and the prediction label of the sudden braking corresponding to the driving data of the sample vehicle, further, when performing model training, a sample feature vector of the driving data of the sample vehicle can be extracted from the training data in turn as the model input, and the corresponding prediction result of the sudden braking can be used as the output. Multiple rounds of model training are performed, and the prediction results obtained in each round of training are compared with the corresponding prediction labels, and the model parameters of the initial vehicle emergency braking prediction model are updated according to the difference between the two until the preset conditions are met, such as the value of the loss constraint function is very small and basically unchanged, then the update of the model parameters is stopped, the training of the vehicle emergency braking prediction model is completed, and a trained vehicle emergency braking prediction model is generated.

[0127] Among them, the present application does not limit the specific composition structure of the initial vehicle emergency braking prediction model, and any existing or future machine learning model that performs binary classification prediction on data can be used as the initial vehicle emergency braking prediction model, such as Convolutional Neural Network (CNN), Recurrent Neural Network (RNN), etc.

[0128] On this basis, after the vehicle emergency braking prediction model is generated by training according to the driving data of the sample vehicle, the generated vehicle emergency braking prediction model can be further verified using the driving data of the verification vehicle. The specific verification process may include the following steps ①-④:

[0129] Step ①: Obtain the driving data of the verification vehicle; wherein the driving data of the verification vehicle includes the vehicle condition information of the verification vehicle, the environmental data information of the verification vehicle, the traffic condition information and the road history information of the city where the verification vehicle is located.

[0130] In this embodiment, in order to verify the vehicle emergency braking prediction model, it is first necessary to obtain the driving data of the verification vehicle. For example, within 5 seconds after the driver of vehicle A releases the accelerator, the vehicle driving data obtained by various sensors pre-installed on vehicle A and other vehicle-mounted systems can be used as the driving data of the verification vehicle, including but not limited to the vehicle condition information of the verification vehicle, the environmental data information of the verification vehicle, the traffic road condition information and the road history information of the city where the verification vehicle is located. Among them, the driving data of the verification vehicle refers to the vehicle driving data that can be used to verify the vehicle emergency braking prediction model. After obtaining the driving data of these verification vehicles and the emergency braking prediction label corresponding to each verification driving data, the subsequent step ② can be continued.

[0131] Step ②: Standardize the driving data of the verification vehicle and extract the verification feature vector corresponding to the processed verification driving data.

[0132] In this embodiment, after the driving data of the verification vehicle is obtained through step ①, it cannot be directly used to verify the vehicle emergency braking prediction model. Instead, the driving data of the verification vehicle needs to be standardized (such as normalized) and the verification feature vector corresponding to the processed verification driving data is extracted. The specific implementation process can refer to the processing process of the driving data of the target vehicle in the above step S102, and the verification feature vector corresponding to the processed (such as normalized) verification driving data can be obtained.

[0133] Step ③: Input the verification feature vector features into the vehicle emergency braking prediction model to obtain the prediction result of emergency braking corresponding to the driving data of the verification vehicle.

[0134] After extracting the verification feature vector corresponding to the processed verification driving data through step ②, the verification feature vector feature can be further input into the vehicle emergency braking prediction model to obtain the prediction result of emergency braking corresponding to the verification vehicle driving data, so as to execute the subsequent step ④.

[0135] Step ④: When the prediction result of emergency braking corresponding to the driving data of the verification vehicle is inconsistent with the emergency braking marking result corresponding to the driving data of the verification vehicle, the driving data of the verification vehicle is used again as the driving data of the sample vehicle to update the vehicle emergency braking prediction model.

[0136] After obtaining the prediction result of emergency braking corresponding to the driving data of the verification vehicle through step ③, if the prediction result is inconsistent with the actual emergency braking marking result corresponding to the verification driving data (such as the pre-marked emergency braking prediction label, the label that emergency braking may occur or the label that emergency braking will not occur), the driving data of the verification vehicle can be used again as the driving data of the sample vehicle, and the parameters of the vehicle emergency braking prediction model can be updated to further improve the prediction precision and accuracy of the vehicle emergency braking prediction model.

[0137] It should be noted that in order to further improve the prediction effect of the model, the latest vehicle driving data can be used regularly to optimize the model and eliminate factors with low information gain, so that the model is more in line with the driving habits of more drivers.

[0138] In this way, by executing the above steps S101-S104, the impending emergency braking event of the target vehicle can be predicted in advance, and the passengers of the target vehicle can be given a corresponding emergency braking warning, so that the passengers can prepare for the psychological anticipation in advance and reduce the occurrence of dizziness.

[0139] In summary, the present embodiment provides a method for predicting sudden braking of a vehicle, which first obtains the driving data of the target vehicle; wherein the driving data includes the vehicle condition information of the target vehicle, the environmental data information of the target vehicle, the traffic condition information, and the road history information of the city where the target vehicle is located, and then the driving data of the target vehicle is standardized, and the target feature vector corresponding to the processed driving data is extracted; then the target feature vector is input into the pre-built vehicle sudden braking prediction model to predict the prediction result of sudden braking of the target vehicle. Furthermore, when the prediction result of sudden braking of the target vehicle indicates that the target vehicle is about to brake suddenly, an emergency braking warning prompt message is sent to the passengers on board the target vehicle.

[0140] It can be seen that since the present application can use the pre-built vehicle emergency braking prediction model to process the collected driving data of the target vehicle during the driving of the target vehicle, so as to predict in advance the emergency braking operation that the driver of the target vehicle is about to perform, and send corresponding early warning prompt information to the passengers on board, so that the passengers can prepare psychological expectations in advance, thereby reducing the occurrence of dizziness and improving the passengers' riding experience and safety.

[0141] Second embodiment

[0142] This embodiment will introduce a vehicle emergency braking prediction device. For related content, please refer to the above method embodiment.

[0143] See also Figure 2 , is a schematic diagram of the composition of a vehicle emergency braking prediction device provided in this embodiment, the device 200 includes:

[0144] The first acquisition unit 201 is used to acquire the driving data of the target vehicle; the driving data includes the vehicle condition information of the target vehicle, the environmental data information of the target vehicle, the traffic condition information and the road history information of the city where the target vehicle is located;

[0145] The first extraction unit 202 is used to perform standardization processing on the driving data of the target vehicle and extract a target feature vector corresponding to the processed driving data;

[0146] The prediction unit 203 is used to input the target feature vector into a pre-built vehicle emergency braking prediction model to predict the emergency braking of the target vehicle;

[0147] The warning unit 204 is used to send an emergency braking warning prompt message to the passengers of the target vehicle when the prediction result of the emergency braking of the target vehicle indicates that the target vehicle is about to brake suddenly.

[0148] In an implementation of this embodiment, the first extraction unit 202 is specifically configured to:

[0149] The driving data of the target vehicle is normalized to obtain normalized driving data; and the normalized driving data are spliced ​​to form a target feature vector corresponding to the processed driving data.

[0150] In an implementation of this embodiment, the prediction unit 203 includes:

[0151] A calculation subunit, used for inputting the target feature vector into a pre-built vehicle emergency braking prediction model, and calculating the distance between the target feature vector and the average feature vector;

[0152] The prediction subunit is used to determine the prediction result of the sudden braking of the target vehicle according to the value of the distance.

[0153] In an implementation of this embodiment, the computing subunit is specifically configured to:

[0154] The chi-square distance between the target feature vector and the average feature vector is calculated.

[0155] In an implementation of this embodiment, the computing subunit is specifically configured to:

[0156] The information gain of each type of data contained in the driving data is calculated, and the chi-square distance between the target feature vector and the average feature vector is optimized by using the information gain to obtain a more accurate calculation result.

[0157] In an implementation of this embodiment, the device further includes:

[0158] A second acquisition unit is used to acquire driving data of a sample vehicle; the driving data of the sample vehicle includes vehicle condition information of the sample vehicle, environmental data information of the sample vehicle, traffic condition information, and road history information of the city where the sample vehicle is located;

[0159] A second extraction unit, used for performing standardization processing on the driving data of the sample vehicle and extracting a sample feature vector corresponding to the processed sample driving data;

[0160] A calculation unit, used for calculating the average value of all sample feature vectors as an average feature vector;

[0161] A training unit is used to train an initial vehicle emergency braking prediction model according to the sample feature vector, the average feature vector and the prediction label of emergency braking corresponding to the driving data of the sample vehicle to generate the vehicle emergency braking prediction model.

[0162] In one implementation of this embodiment, the initial vehicle emergency braking prediction model is a machine learning model that performs binary classification prediction on data.

[0163] In an implementation of this embodiment, the device further includes:

[0164] A third acquisition unit is used to acquire the driving data of the verification vehicle; the driving data of the verification vehicle includes the vehicle condition information of the verification vehicle, the environmental data information of the verification vehicle, the traffic condition information and the road history information of the city where the verification vehicle is located;

[0165] A third extraction unit, used for performing standardization processing on the driving data of the verification vehicle and extracting a verification feature vector corresponding to the processed verification driving data;

[0166] An obtaining unit, used for inputting the verification feature vector feature into the vehicle emergency braking prediction model to obtain a prediction result of emergency braking corresponding to the driving data of the verification vehicle;

[0167] An updating unit is used to update the vehicle emergency braking prediction model by reusing the driving data of the verification vehicle as the driving data of the sample vehicle when the prediction result of emergency braking corresponding to the driving data of the verification vehicle is inconsistent with the emergency braking marking result corresponding to the driving data of the verification vehicle.

[0168] Furthermore, an embodiment of the present application also provides a vehicle emergency braking prediction device, including: a processor, a memory, and a system bus;

[0169] The processor and the memory are connected via the system bus;

[0170] The memory is used to store one or more programs, and the one or more programs include instructions. When the instructions are executed by the processor, the processor executes any one of the implementation methods of the above-mentioned vehicle emergency braking prediction method.

[0171] Furthermore, an embodiment of the present application also provides a computer-readable storage medium, in which instructions are stored. When the instructions are executed on a terminal device, the terminal device executes any one of the implementation methods of the above-mentioned vehicle emergency braking prediction method.

[0172] Furthermore, an embodiment of the present application also provides a computer program product, which, when executed on a terminal device, enables the terminal device to execute any one of the implementation methods of the above-mentioned vehicle emergency braking prediction method.

[0173] It can be known from the description of the above implementation mode that those skilled in the art can clearly understand that all or part of the steps in the above-mentioned embodiment method can be implemented by means of software plus a necessary general hardware platform. Based on such an understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product can be stored in a storage medium such as ROM / RAM, a disk, an optical disk, etc., including several instructions for enabling a computer device (which can be a personal computer, a server, or a network communication device such as a media gateway, etc.) to execute the methods described in the various embodiments of the present application or certain parts of the embodiments.

[0174] It should be noted that the various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments, and the same or similar parts between the various embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part description.

[0175] It should also be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.

[0176] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A vehicle emergency braking prediction method, characterized in that: include: Acquire the driving data of the target vehicle; the driving data includes the vehicle condition information of the target vehicle, the environmental data information of the target vehicle, the traffic condition information and the road history information of the city where the target vehicle is located; Standardizing the driving data of the target vehicle and extracting a target feature vector corresponding to the processed driving data; Inputting the target feature vector into a pre-built vehicle emergency braking prediction model to predict the emergency braking of the target vehicle; When the prediction result of the sudden braking of the target vehicle indicates that the target vehicle is about to make an sudden braking, an emergency braking warning prompt message is sent to passengers on board the target vehicle.

2. The method according to claim 1, characterized in that The step of normalizing the driving data of the target vehicle and extracting a target feature vector corresponding to the processed driving data includes: The driving data of the target vehicle is normalized to obtain normalized driving data; and the normalized driving data are spliced ​​to form a target feature vector corresponding to the processed driving data.

3. The method according to claim 1, characterized in that The step of inputting the target feature vector into a pre-built vehicle emergency braking prediction model to predict the emergency braking of the target vehicle includes: Inputting the target feature vector into a pre-built vehicle emergency braking prediction model, and calculating the distance between the target feature vector and the average feature vector; According to the value of the distance, a prediction result of the sudden braking of the target vehicle is determined.

4. The method according to claim 3, characterized in that The calculating the distance between the target feature vector and the average feature vector comprises: The chi-square distance between the target feature vector and the average feature vector is calculated.

5. The method according to claim 4, characterized in that The calculating the chi-square distance between the target feature vector and the average feature vector comprises: The information gain of each type of data contained in the driving data is calculated, and the chi-square distance between the target feature vector and the average feature vector is optimized by using the information gain to obtain a more accurate calculation result.

6. The method according to claim 3, characterized in that The vehicle emergency braking prediction model is constructed as follows: Acquire driving data of a sample vehicle; the driving data of the sample vehicle includes vehicle condition information of the sample vehicle, environmental data information of the sample vehicle, traffic condition information, and road history information of the city where the sample vehicle is located; Standardizing the driving data of the sample vehicles and extracting sample feature vectors corresponding to the processed sample driving data; Calculate the average of all sample feature vectors as the average feature vector; The initial vehicle emergency braking prediction model is trained according to the sample feature vector, the average feature vector and the prediction label of emergency braking corresponding to the driving data of the sample vehicle to generate the vehicle emergency braking prediction model.

7. The method according to claim 6, characterized in that The initial vehicle emergency braking prediction model is a machine learning model that performs binary classification prediction on data.

8. The method according to claim 6, characterized in that The method further comprises: Acquire the driving data of the verification vehicle; the driving data of the verification vehicle includes the vehicle condition information of the verification vehicle, the environmental data information of the verification vehicle, the traffic condition information and the road history information of the city where the verification vehicle is located; Standardizing the driving data of the verification vehicle and extracting a verification feature vector corresponding to the processed verification driving data; Inputting the verification feature vector feature into the vehicle emergency braking prediction model to obtain the prediction result of emergency braking corresponding to the driving data of the verification vehicle; When the prediction result of sudden braking corresponding to the driving data of the verification vehicle is inconsistent with the sudden braking marking result corresponding to the driving data of the verification vehicle, the driving data of the verification vehicle is re-used as the driving data of the sample vehicle to update the vehicle sudden braking prediction model.

9. A vehicle emergency braking prediction device, characterized in that: include: A first acquisition unit is used to acquire driving data of a target vehicle; the driving data includes vehicle condition information of the target vehicle, environmental data information of the target vehicle, traffic condition information, and road history information of the city where the target vehicle is located; A first extraction unit, used for performing standardization processing on the driving data of the target vehicle and extracting a target feature vector corresponding to the processed driving data; A prediction unit, used for inputting the target feature vector into a pre-built vehicle emergency braking prediction model to predict the emergency braking of the target vehicle; The early warning unit is used to send an early warning prompt message of an emergency brake to passengers of the target vehicle when the prediction result of the emergency brake of the target vehicle indicates that the target vehicle is about to brake suddenly.

10. A vehicle emergency braking prediction device, characterized in that: include: Processor, memory, system bus; The processor and the memory are connected via the system bus; The memory is used to store one or more programs, wherein the one or more programs include instructions, and when the instructions are executed by the processor, the processor executes the method according to any one of claims 1 to 8.

11. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores instructions, and when the instructions are executed on a terminal device, the terminal device executes the method according to any one of claims 1 to 8.

Images