Vehicle collision early warning method and device based on gradient lifting algorithm

Abstract

The embodiment of the invention discloses a vehicle collision early warning method and device based on a gradient lifting algorithm. The method comprises the steps of obtaining a historical data set corresponding to vehicle driving through an Internet of Vehicles cloud platform, constructing a vehicle collision estimation model corresponding to a historical training set based on a gradient lifting algorithm, and determining a collision estimation probability corresponding to real-time driving data according to the vehicle collision estimation model, and finally, determining a warning mode corresponding to the collision estimation probability based on the target driving style, and performing warning according to the warning mode. Multi-dimensional vehicle related data are fused through the Internet of Vehicles cloud platform, and the vehicle collision prediction model constructed based on the gradient lifting algorithm is constructed for collision probability prediction, so that the final early warning result is more accurate, the real-time requirement of real-time driving is met, and the driving experience of a target user is improved.

Description

Technical Field

[0001] One or more embodiments of this specification relate to the technical field of vehicle collision prediction, and in particular, to a vehicle collision warning method and device based on a gradient boosting algorithm. Background Art

[0002] With the continuous increase in the number of automobiles, traffic accidents occur frequently, posing a serious threat to people's lives and property safety. Therefore, a vehicle collision warning method is needed to minimize the occurrence of traffic accidents as much as possible. Currently, traditional vehicle collision warning methods mainly rely on the vehicle's own sensors, such as radar, cameras, etc. However, the detection range of these sensors is limited, and they are easily affected by factors such as weather and road conditions, resulting in inaccurate warning effects and failure to meet the real-time requirements, thus reducing the driving experience. Summary of the Invention

[0003] Embodiments of this specification provide a vehicle collision warning method and device based on a gradient boosting algorithm, and its technical solutions are as follows: In a first aspect, embodiments of this specification provide a vehicle collision warning method based on a gradient boosting algorithm, and the method includes: Obtaining a historical data set corresponding to vehicle driving based on a vehicle networking cloud platform, where the historical data set is composed of each driving feature data and collision result data corresponding to each driving feature data; Constructing a vehicle collision prediction model corresponding to the historical training set based on the gradient boosting algorithm; Determining a collision prediction probability corresponding to real-time driving data according to the vehicle collision prediction model; Determining a warning mode corresponding to the collision prediction probability based on a target driving style, and giving a warning according to the warning mode.

[0004] In a second aspect, a vehicle collision warning device based on a gradient boosting algorithm is provided, and the device includes: An obtaining module, configured to obtain a historical data set corresponding to vehicle driving based on a vehicle networking cloud platform, where the historical data set is composed of each driving feature data and collision result data corresponding to each driving feature data; A constructing module, configured to construct a vehicle collision prediction model corresponding to the historical training set based on the gradient boosting algorithm; An estimating module, configured to determine a collision prediction probability corresponding to real-time driving data according to the vehicle collision prediction model; A determining module, configured to determine a warning mode corresponding to the collision prediction probability based on a target driving style, and give a warning according to the warning mode.

[0005] In a third aspect, an electronic device is provided, including a device processor and a memory; The device processor is connected to the memory; The memory is used to store executable program code; The device processor runs a program corresponding to the executable program code by reading the executable program code stored in the memory, so as to execute the steps of the method provided in the first aspect or any possible implementation manner of the first aspect.

[0006] In a fourth aspect, a computer-readable storage medium is provided, on which a computer program is stored. Instructions are stored in the computer-readable storage medium. When the instructions run on a computer or a device processor, the computer or the device processor is made to execute the method provided in the first aspect or any possible implementation manner of the first aspect.

[0007] The beneficial effects brought by the technical solutions provided in some embodiments of this specification at least include: In one or more embodiments of this specification, a historical data set corresponding to vehicle driving is obtained through a vehicle networking cloud platform, and then a vehicle collision prediction model corresponding to the historical training set is constructed based on the gradient boosting algorithm. Then, the collision prediction probability corresponding to the real-time driving data is determined according to the vehicle collision prediction model. Finally, a warning mode corresponding to the collision prediction probability is determined based on the target driving style, and a warning is given according to the warning mode. By integrating multi-dimensional vehicle-related data through the vehicle networking cloud platform and constructing a vehicle collision prediction model based on the gradient boosting algorithm for collision probability prediction, the final warning result is more accurate, and the real-time requirement of real-time driving is met, improving the driving experience of the target user. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0009] Figure 1 It is a flowchart of a vehicle collision warning method based on the gradient boosting algorithm provided by an embodiment of this specification; Figure 2 It is a schematic structural diagram of a vehicle collision warning device based on the gradient boosting algorithm provided by an embodiment of this specification; Figure 3 It is a schematic structural diagram of an electronic device provided by an embodiment of this specification. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0010] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application.

[0011] In this specification, the terms "first", "second", "third", etc. in the description and claims and the above accompanying drawings are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but optionally further includes steps or units not listed, or optionally further includes other steps or units inherent to these processes, methods, products or devices.

[0012] The following description provides examples and does not limit the scope, applicability or examples set forth in the claims. Changes can be made to the functions and arrangements of the described elements without departing from the scope of the content of this specification. Each example can appropriately omit, substitute or add various processes or components. For example, the described method can be executed in a different order than the described order, and various steps can be added, omitted or combined. In addition, the features described in some examples can be combined into other examples.

[0013] Please refer to Figure 1 , Figure 1 which shows the overall flowchart of a vehicle collision warning method based on the gradient boosting algorithm provided by the embodiments of this specification.

[0014] As Figure 1 shown, the vehicle collision warning method based on the gradient boosting algorithm can at least include the following steps: Step 101, obtain the historical data set corresponding to the vehicle driving based on the vehicle networking cloud platform.

[0015] Wherein, the historical data set is composed of each driving feature data and the collision result data corresponding to each driving feature data.

[0016] In the embodiments of this specification, the vehicle collision warning method based on the gradient boosting algorithm involved in the present invention can be applied to a vehicle collision warning system, and the vehicle collision warning system can include a vehicle networking cloud platform, various vehicle sensors, and an in-vehicle processor. Specifically, after each vehicle sensor transmits the collected driving data to the in-vehicle processor, it is then uploaded to the vehicle networking cloud platform through the in-vehicle processor. When it is necessary to use the vehicle collision warning system to warn the real-time humidifying vehicle, the in-vehicle processor obtains the historical data set including vehicle-to-vehicle information from the vehicle networking cloud platform, constructs a vehicle collision prediction model, and then inputs the real-time driving data collected by each vehicle sensor into the vehicle collision prediction model, and gives an alarm according to the model output result.

[0017] Specifically, since the vehicle networking technology can enable information sharing and communication between vehicles, enabling vehicles to obtain more comprehensive road traffic information, thereby improving the accuracy and timeliness of collision warning, all data such as the vehicle's speed, acceleration, distance from the vehicle in front, and driving direction of the vehicle can be obtained first through various vehicle sensors, such as accelerometers, gyroscopes, GPS modules, OBD-II interfaces, etc. Then, all the data is transmitted to the vehicle networking cloud platform through 4G / 5G or V2X communication. When the in-vehicle processor receives the vehicle collision warning instruction, the historical data set corresponding to the vehicle's driving is obtained through the vehicle networking cloud platform. Specifically, the driving characteristic data corresponding to each acquisition moment within multiple acquisition time periods of the vehicle can be determined first from the vehicle networking cloud platform, and the collision result data indicating whether the vehicle has collided with surrounding vehicles in each driving characteristic data state can be determined from accident reporting systems such as the accident records of insurance companies and traffic management departments. The collision result data can be divided into collided and not collided. Further, the driving characteristic data and the collision result data are associated and paired to obtain the historical data set.

[0018] In an implementable manner, obtaining the historical data set corresponding to the vehicle's driving based on the vehicle networking cloud platform includes: Obtaining at least two groups of historical record data corresponding to the vehicle's driving based on the vehicle networking cloud platform; Determining the driving characteristic data corresponding to each of the historical record data according to the feature extraction method; Integrating the driving characteristic data and the collision result data corresponding to each of the driving characteristic data to obtain the historical data set corresponding to the vehicle's driving.

[0019] In the embodiments of this specification, when obtaining the historical data set corresponding to the vehicle's driving based on the vehicle networking cloud platform, in order to reduce the subsequent algorithm complexity while ensuring data quality and improve the real-time performance of vehicle collision warning, multiple groups of historical record data corresponding to the driving of each vehicle within a fixed time range, such as the past six months, in various driving scenarios can be obtained first based on the vehicle networking cloud platform to enhance the generalization ability of the data set and avoid the deviation of a single data source. Then, since the historical record data includes not only the characteristic data related to the driving state but also non-driving characteristic data such as the vehicle's driving direction and vehicle size, the driving characteristic data that can be used for the subsequent collision prediction model can be extracted from each historical record data according to the feature extraction method. Finally, the driving characteristic data and the collision result data corresponding to each of the driving characteristic data are integrated and paired to obtain the historical data set corresponding to the vehicle's driving.

[0020] In an implementable manner, after obtaining at least two groups of historical record data corresponding to the vehicle's driving based on the vehicle networking cloud platform, it further includes: Normalize each piece of the historical record data to obtain each piece of standard data; Perform attribute reduction processing on each piece of the standard data to obtain each piece of merged data; Determining the driving feature data corresponding to each piece of the historical record data according to the feature extraction method includes: Determine the driving feature data corresponding to each piece of the merged data according to the feature extraction method.

[0021] In the embodiments of this specification, in order to ensure the data quality of the obtained historical record data and the effectiveness of subsequent analysis, after obtaining each piece of historical record data, abnormal data and duplicate data in each piece of historical record data can be removed through data cleaning first, and then missing data records can be deleted, and missing values can be filled with statistical values (such as mean, median). Then, since data with different dimensions may cause the subsequent model to assign too high weights to certain features during the training process, thereby affecting the performance of the model, it is necessary to normalize each piece of historical record data to convert historical record data with different dimensions to the same scale to obtain each piece of standard data for subsequent algorithm processing. Optionally, the Min-Max Scaling normalization algorithm can be used to linearly scale each piece of historical record data to a specific interval, usually [0, 1] or [-1, 1]. Further, since the high-dimensional data corresponding to each piece of standard data may cause problems such as increased computational complexity and model overfitting, attribute reduction processing can be performed on each piece of standard data to reduce the data dimension by creating new attribute dimensions through attribute merging or directly deleting irrelevant attributes to obtain each piece of merged data. Subsequently, the driving feature data corresponding to each piece of merged data can be directly determined according to the feature extraction method.

[0022] Step 102, construct a vehicle collision prediction model corresponding to the historical training set based on the gradient boosting algorithm.

[0023] In the embodiments of this specification, after obtaining the historical data set corresponding to vehicle driving through the vehicle networking cloud platform, in order to perform vehicle collision warning on the real-time driving state, it is necessary to construct a vehicle collision prediction model according to the historical training set. Specifically, an initial prediction model can be pre-constructed using the XGBoost or CatBoost gradient boosting algorithm, and then the obtained historical training set is used to train and validate its initial prediction model to obtain a trained vehicle collision prediction model for subsequent real-time vehicle collision warning.

[0024] Among them, after deploying the trained model with satisfactory performance to the actual vehicle-mounted processor, it is necessary to continuously monitor the performance of the prediction model and perform retraining or updating regularly according to new data to maintain the accuracy and timeliness of the prediction model.

[0025] In an implementable manner, constructing the vehicle collision prediction model corresponding to the historical training set based on the gradient boosting algorithm includes: Determining the dynamic regularization coefficient in the objective function based on the gradient boosting algorithm, and constructing an initial vehicle collision prediction model according to the dynamic regularization coefficient; Training the initial vehicle collision prediction model according to the historical training set to obtain a vehicle collision prediction model.

[0026] In the embodiments of this specification, when using the XGBoost gradient boosting algorithm to construct the vehicle collision prediction model corresponding to the historical training set, in order to flexibly adjust the complexity limit of the model according to the data situation or training progress during the model training process to prevent overfitting, the logistic loss function can be first selected as the objective function, and the dynamic regularization coefficient in the objective function can be determined according to the adjustment of the tree structure parameters, and an initial vehicle collision prediction model can be constructed through the dynamic regularization coefficient. Specifically, the objective loss function in the initial vehicle collision prediction model is: where n is the number of samples, is the true label of the i-th sample, is the predicted value of the i-th sample, is the loss function of a single sample (such as mean squared error, cross entropy), K is the total number of decision trees (number of iterations), is the model of the k-th decision tree, is the dynamic regularization coefficient of the k-th tree.

[0027] The dynamic regularization coefficient of the k-th tree The calculation formula of is: where T is the number of leaf nodes of the k-th tree, is the weight vector of the leaf node, is the penalty coefficient of the number of leaf nodes, is the second-layer regularization coefficient, controlling the sum of squares of the leaf weights.

[0028] Then, iteratively train the constructed initial vehicle collision prediction model according to the historical training set to obtain a vehicle collision prediction model that meets the prediction requirements.

[0029] In an implementable manner, training the initial vehicle collision prediction model according to the historical training set to obtain a vehicle collision prediction model includes: Dividing the historical training set into a model training set and a model optimization set based on the time sequence; Train the initial vehicle collision prediction model according to the model training set to obtain a trained vehicle collision prediction model; Perform lightweight compression optimization on the trained vehicle collision prediction model according to the model optimization set to obtain a vehicle collision prediction model.

[0030] In the embodiments of this specification, when using a historical training set to train an initial vehicle collision prediction model, in order to improve the generalization ability of the prediction model, all records in the historical training set can be sorted according to timestamps based on the time sequence first, and the sorted data set can be divided into 80% and 20% proportions. The first 80% is used as the model training set, and the latter 20% is used as the model optimization set. Then, train the initial vehicle collision prediction model through the model training set, and adjust the parameters inside the gradient boosting model to make the prediction result of the model on the training set as close as possible to the true label. Further, in order to reduce the computational complexity of the vehicle collision prediction model and improve the real-time performance of the real-time prediction model, the leaf nodes or branches in the model that contribute little or zero to the prediction can be determined and removed according to the model optimization set first, and then the floating-point weights or node values used in the model can be converted into a lower-precision data type to complete the lightweight compression optimization of the prediction model and obtain a vehicle collision prediction model.

[0031] Step 103: Determine the collision prediction probability corresponding to the real-time driving data according to the vehicle collision prediction model.

[0032] In the embodiments of this specification, after constructing a vehicle collision prediction model, only the real-time driving data in the real-time driving state obtained needs to be used as the model input and transmitted to the end of the vehicle collision prediction model, and the corresponding collision prediction probability in this real-time driving state can be obtained. Specifically, the value range of the collision prediction probability is [0, 1].

[0033] Step 104: Determine the warning mode corresponding to the collision prediction probability based on the target driving style, and give a warning according to the warning mode.

[0034] In the embodiments of this specification, after determining the collision prediction probability corresponding to the real-time driving data through the vehicle collision prediction model, it is necessary to further evaluate according to the collision prediction probability to determine the corresponding warning mode for warning. Among them, since different drivers have different driving styles, even for the same collision prediction probability, the acceptable warning modes are different. Therefore, it is necessary to first determine the target driving style corresponding to the real-time driving state at this time through the driving instruction issued by the target terminal, and then determine different warning criteria or warning databases through the target driving style, input the obtained collision prediction probability into them, and the corresponding warning mode can be obtained. Finally, warning is carried out according to the warning mode. Optionally, when determining the warning mode, the corresponding collision risk threshold can also be determined first through the target driving style, and then the corresponding warning mode is determined according to the comparison result between the collision risk threshold and the collision prediction probability.

[0035] As an example, the warning mode may include sound, display, and vibration warning modes. A certain collision risk threshold is 0.7. When the determined collision risk threshold is greater than 0.7, the corresponding different warning modes can be triggered according to the range exceeding 0.7.

[0036] In one implementable manner, determining the warning mode corresponding to the collision prediction probability based on the target driving style includes: Determining the collision risk threshold corresponding to the target driving style based on a preset style-threshold database; Performing a difference calculation between the collision prediction probability and the collision risk threshold, and determining the warning mode based on the difference calculation result.

[0037] In the embodiments of this specification, a style-threshold database can be constructed first through a large number of historical data records and driver questionnaires. In the preset style-threshold database, only by inputting a driving style can a corresponding collision risk threshold be output. Therefore, when determining the warning mode corresponding to the collision prediction probability, the collision risk threshold corresponding to the target driving style can be determined through the preset style-threshold database, and then the collision prediction probability obtained through the vehicle collision prediction model is subjected to a difference calculation with the collision risk threshold. When the difference calculation result indicates that the collision prediction probability is less than the collision risk threshold, no warning is given. When the difference calculation result indicates that the collision prediction probability is greater than the collision risk threshold, the corresponding warning mode is determined according to the magnitude of the difference corresponding to the difference calculation result.

[0038] Optionally, when determining the warning mode according to the difference calculation result, the magnitude of the difference calculation result can be mapped and bound to the warning mode one by one, or the corresponding warning mode can be determined through the difference interval rule.

[0039] In one implementable manner, determining the warning mode based on the difference calculation result includes: Determining the warning interval corresponding to the difference calculation result based on the difference interval rule; When the warning interval is characterized as the first interval, determining the warning mode as display warning; When the warning interval is characterized as the second interval, determining the warning mode as speaker warning; When the warning interval is characterized as the third interval, determining the warning mode as steering wheel warning.

[0040] In the embodiments of this specification, determining the warning mode based on the difference calculation result can first determine the warning interval corresponding to the difference calculation result through the difference interval rule. As an example, the difference interval rule can be that when 0 ≤ difference calculation result ≤ 0.1, the corresponding warning interval is the first interval; when 0.1 < difference calculation result ≤ 0.2, the corresponding warning interval is the second interval; when 0.2 < difference calculation result ≤ 1, the corresponding warning interval is the third interval. Then, when the determined warning interval is characterized as the first interval, determine that the warning mode at this time is display warning, that is, display warning on the main driver display screen. When the determined warning interval is characterized as the second interval, determine that the warning mode at this time is speaker warning, that is, while displaying warning on the main driver display screen, voice warning is also carried out through the in-vehicle audio. When the determined warning interval is characterized as the third interval, determine that the warning mode at this time is steering wheel warning, that is, while displaying warning on the main driver display screen and voice warning is carried out through the in-vehicle audio, the highest-level warning is also carried out through the vibration of the steering wheel.

[0041] The above describes specific embodiments of this specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims can be performed in a different order than in the embodiments and still achieve the desired results. Additionally, the processes depicted in the drawings do not necessarily require the specific order or sequential order shown to achieve the desired results. In certain implementations, multitasking and parallel processing are also possible or may be advantageous.

[0042] Next, please refer to Figure 2 , Figure 2 which shows a schematic structural diagram of a vehicle collision warning device based on the gradient boosting algorithm provided by the embodiments of this specification. It should be noted that Figure 2 the shown vehicle collision warning device based on the gradient boosting algorithm is used to execute the method of the embodiments of this application Figure 1 shown. For the sake of convenience of description, only the parts related to the embodiments of this application are shown. For the specific technical details not disclosed, please refer to the embodiments Figure 1 shown in this application.

[0043] As Figure 2 shown, the vehicle collision warning device based on the gradient boosting algorithm may at least include: An acquisition module 201, configured to obtain a historical data set corresponding to vehicle driving based on a vehicle networking cloud platform, where the historical data set is composed of each driving feature data and the collision result data corresponding to each driving feature data; A construction module 202, configured to construct a vehicle collision prediction model corresponding to the historical training set based on the gradient boosting algorithm; A prediction module 203, configured to determine a collision prediction probability corresponding to real-time driving data according to the vehicle collision prediction model; A determination module 204, configured to determine a warning mode corresponding to the collision prediction probability based on a target driving style, and perform a warning according to the warning mode.

[0044] In an implementable manner, the acquisition module 201 is specifically configured to: Obtain at least two groups of historical record data corresponding to vehicle driving based on a vehicle networking cloud platform; Determine the driving feature data corresponding to each historical record data according to a feature extraction method; Integrate the driving feature data and the collision result data corresponding to each driving feature data to obtain a historical data set corresponding to vehicle driving.

[0045] In an implementable manner, the acquisition module 201 is specifically further configured to: Perform normalization processing on each historical record data to obtain each standard data; Perform attribute reduction processing on each standard data to obtain each merged data; The determining the driving feature data corresponding to each historical record data according to a feature extraction method includes: Determining the driving feature data corresponding to each merged data according to a feature extraction method.

[0046] In an implementable manner, the construction module 202 is specifically configured to: Determine a dynamic regularization coefficient in an objective function based on the gradient boosting algorithm, and construct an initial vehicle collision prediction model according to the dynamic regularization coefficient; Train the initial vehicle collision prediction model according to the historical training set to obtain a vehicle collision prediction model.

[0047] In an implementable manner, the construction module 202 is specifically further configured to: Divide the historical training set into a model training set and a model optimization set based on a time sequence; Train the initial vehicle collision prediction model according to the model training set to obtain a trained vehicle collision prediction model; Perform lightweight compression optimization on the trained vehicle collision prediction model according to the model optimization set to obtain a vehicle collision prediction model.

[0048] In an implementable manner, the determination module 204 is specifically configured to: Determine the collision risk threshold corresponding to the target driving style based on a preset style-threshold database; Calculate the difference between the collision prediction probability and the collision risk threshold, and determine the warning mode based on the difference calculation result.

[0049] In an implementable manner, the determination module 204 is further specifically configured to: Determine the warning interval corresponding to the difference calculation result based on the difference interval rule; When the warning interval is characterized as the first interval, determine the warning mode as a display warning; When the warning interval is characterized as the second interval, determine the warning mode as a speaker warning; When the warning interval is characterized as the third interval, determine the warning mode as a steering wheel warning.

[0050] Those skilled in the art can clearly understand that the technical solutions of the embodiments of the present application can be implemented by means of software and / or hardware. The "units" and "modules" in this specification refer to software and / or hardware that can independently complete or cooperate with other components to complete specific functions. The hardware can be, for example, a Field-Programmable Gate Array (FPGA), an Integrated Circuit (IC), etc.

[0051] Each processing unit and / or module of the embodiments of the present application can be implemented by an analog circuit that implements the functions described in the embodiments of the present application, or can be implemented by software that executes the functions described in the embodiments of the present application.

[0052] Next, please refer to Figure 3 , Figure 3 which shows a schematic structural diagram of an electronic device provided by an embodiment of this specification.

[0053] As Figure 3 shown, the electronic device 300 may include: at least one device processor 301, at least one network interface 303, a user interface 303, a memory 305, and at least one communication bus 302.

[0054] Among them, the communication bus 302 can be used to realize the connection and communication of the above-mentioned various components.

[0055] Among them, the user interface 303 may include buttons, and the optional user interface may further include standard wired interfaces and wireless interfaces.

[0056] Among them, the network interface 304 may include, but is not limited to, a Bluetooth module, an NFC module, a Wi-Fi module, etc.

[0057] Among them, the device processor 301 may include one or more processing cores. The device processor 301 connects various parts within the entire electronic device 300 through various interfaces and lines, and executes various functions of the electronic device 300 and processes data by running or executing instructions, programs, code sets, or instruction sets stored in the memory 305, and by calling the data stored in the memory 305. Optionally, the device processor 301 may be implemented in at least one hardware form of DSP, FPGA, or PLA. The device processor 301 may integrate one or a combination of several of CPU, GPU, and modem, etc. Among them, the CPU mainly processes the operating system, user interface, application programs, etc.; the GPU is responsible for rendering and drawing the content to be displayed on the display screen; the modem is used to process wireless communication. It can be understood that the above-mentioned modem may not be integrated into the device processor 301 and may be implemented separately by a single chip.

[0058] Among them, the memory 305 may include RAM and may also include ROM. Optionally, the memory 305 includes a non-transitory computer-readable medium. The memory 305 can be used to store instructions, programs, codes, code sets, or instruction sets. The memory 305 may include a program storage area and a data storage area. Among them, the program storage area may store instructions for implementing the operating system, instructions for at least one function (such as touch function, sound playback function, image playback function, etc.), instructions for implementing the above-mentioned various method embodiments, etc.; the data storage area may store the data involved in the above-mentioned various method embodiments. Optionally, the memory 305 may also be at least one storage device located far from the aforementioned device processor 301. As Figure 3 shown, the memory 305, as a computer storage medium, may include an operating system, a network communication module, a user interface module, and program instructions.

[0059] Specifically, the device processor 301 may be used to call the vehicle collision warning application program stored in the memory 305 and specifically perform the following operations: Obtain a historical data set corresponding to the vehicle driving based on the vehicle networking cloud platform, where the historical data set is composed of each driving feature data and the collision result data corresponding to each driving feature data; Construct a vehicle collision prediction model corresponding to the historical training set based on the gradient boosting algorithm; Determine the collision prediction probability corresponding to the real-time driving data according to the vehicle collision prediction model; Determine the warning mode corresponding to the collision prediction probability based on the target driving style, and give a warning according to the warning mode.

[0060] As an option of the embodiment of this specification, the obtaining of the historical data set corresponding to the vehicle driving based on the vehicle networking cloud platform includes: Obtain at least two groups of historical record data corresponding to the vehicle driving based on the vehicle networking cloud platform; Determine the driving feature data corresponding to each historical record data according to the feature extraction method; Integrate the driving feature data and the collision result data corresponding to each driving feature data to obtain the historical data set corresponding to the vehicle driving.

[0061] As an option of the embodiment of this specification, after obtaining at least two groups of historical record data corresponding to the vehicle driving based on the vehicle networking cloud platform, it further includes: Perform normalization processing on each historical record data to obtain each standard data; Perform attribute reduction processing on each standard data to obtain each merged data; The determining of the driving feature data corresponding to each historical record data according to the feature extraction method includes: Determine the driving feature data corresponding to each merged data according to the feature extraction method.

[0062] As an option of the embodiment of this specification, the constructing of the vehicle collision prediction model corresponding to the historical training set based on the gradient boosting algorithm includes: Determine the dynamic regularization coefficient in the objective function based on the gradient boosting algorithm, and construct an initial vehicle collision prediction model according to the dynamic regularization coefficient; Train the initial vehicle collision prediction model according to the historical training set to obtain a vehicle collision prediction model.

[0063] As an option of the embodiment of this specification, the training of the initial vehicle collision prediction model according to the historical training set to obtain a vehicle collision prediction model includes: Divide the historical training set into a model training set and a model optimization set based on the time sequence; Train the initial vehicle collision prediction model according to the model training set to obtain a trained vehicle collision prediction model; Perform lightweight compression optimization on the trained vehicle collision prediction model according to the model optimization set to obtain a vehicle collision prediction model.

[0064] As an option in the embodiments of this specification, determining the warning mode corresponding to the collision prediction probability based on the target driving style includes: Determining the collision risk threshold corresponding to the target driving style based on a preset style-threshold database; Calculating the difference between the collision prediction probability and the collision risk threshold, and determining the warning mode based on the result of the difference calculation.

[0065] As an option in the embodiments of this specification, determining the warning mode based on the result of the difference calculation includes: Determining the warning interval corresponding to the result of the difference calculation based on the difference interval rule; When the warning interval is characterized as the first interval, determining the warning mode as a display warning; When the warning interval is characterized as the second interval, determining the warning mode as a speaker warning; When the warning interval is characterized as the third interval, determining the warning mode as a steering wheel warning.

[0066] The embodiments of this specification also provide a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the steps of the above method are implemented. Among them, the computer-readable storage medium may include, but is not limited to, any type of disk, including floppy disks, optical disks, DVDs, CD-ROMs, microdrives, and magneto-optical disks, ROMs, RAMs, EPROMs, EEPROMs, DRAMs, VRAMs, flash memory devices, magnetic cards or optical cards, nanosystems (including molecular memory ICs), or any type of medium or device suitable for storing instructions and / or data.

[0067] It should be noted that for the foregoing method embodiments, for the sake of simple description, they are all expressed as a series of action combinations. However, those skilled in the art should know that this application is not limited by the described action sequence, because according to this application, certain steps can be performed in other sequences or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0068] In the above embodiments, the descriptions of the various embodiments have their own emphases. For the parts not detailed in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0069] In several embodiments provided in the present application, it should be understood that the disclosed device can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed coupling or direct coupling or communication connection between each other can be through some service interfaces. The indirect coupling or communication connection of the device or unit can be in an electrical or other form.

[0070] The units described as separate components may or may not be physically separated. The components displayed as units may or may not be physical units, that is, they can be located in one place, or they can be distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0071] In addition, in each embodiment of the present application, the functional units can be integrated in a processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit. The above-mentioned integrated units can be implemented in the form of hardware or in the form of software functional units.

[0072] If the above-mentioned integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable memory. Based on such an understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a memory and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in each embodiment of the present application. And the aforementioned memory includes: USB flash drive, read-only memory (ROM), random access memory (RAM), mobile hard disk, magnetic disk or optical disk and other media that can store program codes.

[0073] Those of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing relevant hardware through a program. The program can be stored in a computer-readable memory. The memory can include: flash drive, read-only memory (ROM), random access memory (RAM), magnetic disk or optical disk, etc.

[0074] The above description has been made of specific embodiments of this specification. Other embodiments are within the scope of the appended claims. In some cases, the acts or steps recited in the claims may be performed in a different order than in the embodiments and still achieve the desired result. Additionally, the processes depicted in the figures do not necessarily require the particular order shown or sequential order to achieve the desired result. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

Claims

1. A vehicle collision warning method based on the gradient boosting algorithm, characterized in that, The method includes: Obtaining a historical dataset corresponding to vehicle driving based on the vehicle networking cloud platform, where the historical dataset consists of various driving feature data and collision result data corresponding to each driving feature data; Constructing a vehicle collision prediction model corresponding to the historical training set based on the gradient boosting algorithm; Determining a collision prediction probability corresponding to real-time driving data according to the vehicle collision prediction model; Determining a warning mode corresponding to the collision prediction probability based on the target driving style, and giving a warning according to the warning mode.

2. The method according to claim 1, characterized in that, The obtaining of the historical dataset corresponding to vehicle driving based on the vehicle networking cloud platform includes: Obtaining at least two groups of historical record data corresponding to vehicle driving based on the vehicle networking cloud platform; Determining driving feature data corresponding to each historical record data according to the feature extraction method; Integrating the driving feature data and the collision result data corresponding to each driving feature data to obtain a historical dataset corresponding to vehicle driving.

3. The method according to claim 2, wherein After obtaining at least two groups of historical record data corresponding to vehicle driving based on the vehicle networking cloud platform, it further includes: Performing normalization processing on each historical record data to obtain each standard data; Performing attribute reduction processing on each standard data to obtain each merged data; The determining of the driving feature data corresponding to each historical record data according to the feature extraction method includes: Determining driving feature data corresponding to each merged data according to the feature extraction method.

4. The method according to claim 1, characterized in that, The constructing of the vehicle collision prediction model corresponding to the historical training set based on the gradient boosting algorithm includes: Determining a dynamic regularization coefficient in the objective function based on the gradient boosting algorithm, and constructing an initial vehicle collision prediction model according to the dynamic regularization coefficient; Training the initial vehicle collision prediction model according to the historical training set to obtain a vehicle collision prediction model.

5. The method according to claim 4, wherein The training of the initial vehicle collision prediction model according to the historical training set to obtain a vehicle collision prediction model includes: Dividing the historical training set into a model training set and a model optimization set based on the time sequence; Training the initial vehicle collision prediction model according to the model training set to obtain a trained vehicle collision prediction model; Performing lightweight compression optimization on the trained vehicle collision prediction model according to the model optimization set to obtain a vehicle collision prediction model.

6. The method according to claim 1, wherein The determining of the warning mode corresponding to the collision prediction probability based on the target driving style includes: Determining a collision risk threshold corresponding to the target driving style based on a preset style-threshold database; Calculating the difference between the collision prediction probability and the collision risk threshold, and determining a warning mode based on the difference calculation result.

7. The method according to claim 6, wherein The determining of the warning mode based on the difference calculation result includes: Determining a warning interval corresponding to the difference calculation result based on the difference interval rule; When the warning interval is characterized as the first interval, determining the warning mode as a display screen warning; When the warning interval is characterized as the second interval, determining the warning mode as a speaker warning; When the warning interval is characterized as the third interval, determining the warning mode as a steering wheel warning.

8. A vehicle collision warning device based on the gradient boosting algorithm, characterized in that, The device includes: An acquisition module, configured to acquire a historical data set corresponding to vehicle driving based on a vehicle networking cloud platform, where the historical data set is composed of each driving feature data and collision result data corresponding to each driving feature data; A construction module, configured to construct a vehicle collision prediction model corresponding to the historical training set based on a gradient boosting algorithm; A prediction module, configured to determine a collision prediction probability corresponding to real-time driving data according to the vehicle collision prediction model; A determination module, configured to determine a warning mode corresponding to the collision prediction probability based on a target driving style, and perform a warning according to the warning mode.

9. An electronic device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, When the processor executes the computer program, the steps of the method according to any one of claims 1-7 are implemented.

10. A computer-readable storage medium, on which a computer program is stored, where instructions are stored in the computer-readable storage medium, and when the instructions run on a computer or a processor, the computer or the processor is caused to execute the steps of the method according to any one of claims 1-7.

Images