CN114299633A - Automobile driving accident prediction method and device, electronic equipment and storage medium - Google Patents
Automobile driving accident prediction method and device, electronic equipment and storage medium Download PDFInfo
- Publication number
- CN114299633A CN114299633A CN202111623475.2A CN202111623475A CN114299633A CN 114299633 A CN114299633 A CN 114299633A CN 202111623475 A CN202111623475 A CN 202111623475A CN 114299633 A CN114299633 A CN 114299633A
- Authority
- CN
- China
- Prior art keywords
- algorithm
- driving
- automobile
- data
- model
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 81
- 230000004927 fusion Effects 0.000 claims abstract description 83
- 230000008569 process Effects 0.000 claims abstract description 34
- 238000004422 calculation algorithm Methods 0.000 claims description 142
- 238000012549 training Methods 0.000 claims description 78
- 230000006870 function Effects 0.000 claims description 76
- 238000007635 classification algorithm Methods 0.000 claims description 66
- 238000012360 testing method Methods 0.000 claims description 40
- 238000011156 evaluation Methods 0.000 claims description 20
- 238000013138 pruning Methods 0.000 claims description 17
- 238000007781 pre-processing Methods 0.000 claims description 6
- 238000004590 computer program Methods 0.000 claims description 2
- 206010039203 Road traffic accident Diseases 0.000 abstract description 13
- 238000012545 processing Methods 0.000 description 13
- 238000010586 diagram Methods 0.000 description 6
- 238000010276 construction Methods 0.000 description 5
- 238000010606 normalization Methods 0.000 description 5
- 238000006243 chemical reaction Methods 0.000 description 4
- 238000005457 optimization Methods 0.000 description 4
- 230000000007 visual effect Effects 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 230000006872 improvement Effects 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 3
- 230000002159 abnormal effect Effects 0.000 description 2
- 230000001133 acceleration Effects 0.000 description 2
- 238000013145 classification model Methods 0.000 description 2
- 230000001419 dependent effect Effects 0.000 description 2
- 239000013307 optical fiber Substances 0.000 description 2
- 230000000644 propagated effect Effects 0.000 description 2
- 239000004576 sand Substances 0.000 description 2
- 230000035945 sensitivity Effects 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- 230000009471 action Effects 0.000 description 1
- 230000006978 adaptation Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000003491 array Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 230000000295 complement effect Effects 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- DMBHHRLKUKUOEG-UHFFFAOYSA-N diphenylamine Chemical compound C=1C=CC=CC=1NC1=CC=CC=C1 DMBHHRLKUKUOEG-UHFFFAOYSA-N 0.000 description 1
- 238000007499 fusion processing Methods 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000005065 mining Methods 0.000 description 1
- 238000002156 mixing Methods 0.000 description 1
- 238000007500 overflow downdraw method Methods 0.000 description 1
- 238000004806 packaging method and process Methods 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 238000013139 quantization Methods 0.000 description 1
- 230000003252 repetitive effect Effects 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Traffic Control Systems (AREA)
Abstract
The disclosure provides a method and a device for predicting an automobile driving accident, electronic equipment and a storage medium, and relates to the technical field of vehicles. The method comprises the following steps: collecting driving data in the driving process of the automobile in real time; inputting the driving data into a pre-constructed dynamic fusion model to obtain a predicted value of the automobile driving accident; and dynamically predicting the driving risk of the automobile according to the relation between the obtained predicted value of the driving accident of the automobile and a preset driving safety threshold value. The invention provides a method and a device for predicting automobile driving accidents, electronic equipment and a storage medium, which are used for collecting driving data in the automobile driving process, inputting the collected driving data into a pre-constructed dynamic fusion model, automatically predicting the predicted value of the traffic accidents in the automobile driving process, dynamically predicting the automobile driving risks, reducing the probability of the traffic accidents and improving the user experience.
Description
Technical Field
The present disclosure relates to the field of vehicle technologies, and in particular, to a method and an apparatus for predicting an automobile driving accident, an electronic device, and a storage medium.
Background
In the application scenario of vehicle-road cooperation, a prediction algorithm is commonly used for mining the associated information and the distribution rule hidden in the driving data so as to evaluate the future output condition of a specific event. The prediction model fusion is to comprehensively consider the output results of various prediction algorithms, train a plurality of weak classification algorithms by construction, and then integrate and fuse the weak classification algorithms according to a certain weight proportion. The existing model fusion method mainly comprises various methods such as Voting, Blending, Stacking and the like, and different models have respective lengths and have obvious differences, and the model fusion mode can enable the models to exert respective advantages, so that a relatively weak classification algorithm is combined with decision through a certain strategy, and the strong prediction capability is achieved.
In the related technology, in the current weighting decision fusion process, the problem of reasonable distribution of weight coefficients of all classification algorithms is not fully considered, and the prediction capability integration process of all independent classification algorithms under different regression functions is neglected, so that in the application scene of driving accident prediction, the situation that the overall evaluation index improvement effect is poor after model fusion is carried out, and the problem of low model fusion efficiency is caused.
It is to be noted that the information disclosed in the above background section is only for enhancement of understanding of the background of the present disclosure, and thus may include information that does not constitute prior art known to those of ordinary skill in the art.
Disclosure of Invention
The present disclosure provides a method and an apparatus for predicting an automobile driving accident, an electronic device, and a storage medium, which at least overcome to some extent the problems of low model fusion efficiency and poor overall evaluation index improvement effect provided in the related art.
Additional features and advantages of the disclosure will be set forth in the detailed description which follows, or in part will be obvious from the description, or may be learned by practice of the disclosure.
According to an aspect of the present disclosure, there is provided a driving accident prediction method for a vehicle, including:
collecting driving data in the driving process of the automobile in real time;
inputting the driving data into a pre-constructed dynamic fusion model to obtain a predicted value of the automobile driving accident;
and dynamically predicting the driving risk of the automobile according to the relation between the obtained predicted value of the driving accident of the automobile and a preset driving safety threshold value.
In one embodiment of the present disclosure, the dynamic fusion model is constructed by:
collecting test driving data in the driving process of the automobile during the test;
preprocessing the test driving data to obtain training sample data;
and constructing a dynamic fusion model according to the training sample data based on a preset classification algorithm.
In one embodiment of the present disclosure, the preset classification algorithm includes a support vector regression SVR algorithm, a K-nearest-KNN algorithm, and a classification regression tree CART algorithm.
In an embodiment of the present disclosure, the building a dynamic fusion model according to the training sample data based on a preset classification algorithm includes:
respectively inputting training set data in the training sample data into each preset classification algorithm to respectively obtain an algorithm model to be verified of each classification algorithm;
inputting test set data in the training sample data into algorithm models to be verified of all classification algorithms to obtain three evaluation indexes of mean square error, R2 and explained variance fraction of all algorithm models;
and calculating a standard deviation and a correlation coefficient based on the three evaluation indexes to obtain fusion weights among the classification algorithms, and constructing a dynamic fusion model according to the fusion weights among the classification algorithms.
In an embodiment of the present disclosure, the respectively inputting training set data in training sample data into each preset classification algorithm to respectively obtain an algorithm model to be verified of each classification algorithm includes:
inputting the training set data into an SVR algorithm, and constructing an SVR model function and an SVR loss function;
obtaining different width values of the insensitive region, and obtaining an SVR classification function and an SVR function weight corresponding to the width values of the insensitive region to minimize a loss value;
obtaining a driving accident prediction value of the SVR algorithm according to the width value of the non-sensitive area and the weight of the SVR function;
and calculating a SVR algorithm weighting decision prediction value formula according to the SVR function weight and the driving accident prediction value of the SVR algorithm to obtain an SVR algorithm model to be verified.
In an embodiment of the present disclosure, the respectively inputting training set data in training sample data into each preset classification algorithm to respectively obtain an algorithm model to be verified of each classification algorithm includes:
inputting the training set data into a KNN algorithm, and constructing a KNN classification function and a KNN loss function;
acquiring different neighborhood values to obtain a KNN classification function and a KNN function weight corresponding to the neighborhood values, so that the loss value is minimum;
obtaining a driving accident prediction value of the KNN algorithm according to the neighborhood value and the KNN function weight;
and calculating a weighted decision prediction value formula of the KNN algorithm according to the KNN function weight and the driving accident prediction value of the KNN algorithm to obtain the KNN algorithm model to be verified.
In an embodiment of the present disclosure, the respectively inputting training set data in training sample data into each preset classification algorithm to respectively obtain an algorithm model to be verified of each classification algorithm includes:
inputting training set data into a CART algorithm to construct a CART loss function;
obtaining different pruning rates, and obtaining subtrees corresponding to the pruning rates and CART function weights to minimize loss values;
obtaining a driving accident prediction value of the CART algorithm according to the subtrees and the pruning rate;
and calculating a CART algorithm weighting decision prediction value formula according to the CART function weight and the driving accident prediction value of the CART algorithm to obtain a CART algorithm model to be verified.
According to another aspect of the present disclosure, there is provided an automobile driving accident prediction apparatus including:
the acquisition module is used for acquiring driving data in the driving process of the automobile in real time;
the prediction module is used for inputting the driving data into a pre-constructed dynamic fusion model so as to obtain a predicted value of the automobile driving accident;
and the comparison module is used for dynamically predicting the automobile driving risk according to the relation between the obtained automobile driving accident prediction value and a preset driving safety threshold value.
According to another aspect of the present disclosure, there is provided an electronic device including: a processor; and a memory for storing executable instructions of the processor; wherein the processor is configured to perform the above-described automotive driving accident prediction method via execution of the executable instructions.
According to another aspect of the present disclosure, there is provided a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, implements the above-described method for predicting a driving accident of a vehicle.
The method, the device, the electronic equipment and the storage medium for predicting the automobile driving accident, provided by the embodiment of the disclosure, are used for collecting driving data in the automobile driving process, inputting the collected driving data into a pre-constructed dynamic fusion model, and automatically predicting the predicted value of the traffic accident in the automobile driving process, so that the automobile driving risk is dynamically predicted, the probability of the traffic accident is reduced, and the user experience is improved.
Further, according to the automobile driving accident prediction method, the automobile driving accident prediction device, the electronic device and the storage medium provided by the embodiment of the disclosure, based on automobile driving test driving data, a Support Vector Regression (SVR) algorithm, a K nearest KNN classification algorithm and a classification regression tree (CART) algorithm are adopted to respectively train and learn, an integrated prediction result under an optimization strategy when a loss value in each independent classification algorithm is the smallest is obtained by setting different non-sensitive region width values, neighborhood values and pruning rates, a standard deviation and a correlation coefficient are calculated based on evaluation index values of each model, so that a dynamic fusion model capable of predicting automobile driving accidents is obtained, a predicted value of a traffic accident occurring in an automobile driving process is automatically predicted, and therefore, the strong overall prediction capability is achieved, the fusion efficiency of the models is improved, and user experience is good.
It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the disclosure.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and together with the description, serve to explain the principles of the disclosure. It is to be understood that the drawings in the following description are merely exemplary of the disclosure, and that other drawings may be derived from those drawings by one of ordinary skill in the art without the exercise of inventive faculty.
FIG. 1 is a flow chart illustrating a method for predicting a driving accident in an automobile according to an embodiment of the present disclosure;
FIG. 2 illustrates a flow diagram for constructing a dynamic fusion model in an embodiment of the disclosure;
FIG. 3 illustrates a flow diagram for constructing a dynamic fusion model in yet another embodiment of the present disclosure;
FIG. 4 illustrates a flow chart for constructing an SVR algorithm model in an embodiment of the present disclosure;
FIG. 5 shows a flow chart for constructing a KNN classification algorithm model in an embodiment of the disclosure;
FIG. 6 illustrates a flow chart for constructing a CART classification algorithm model in an embodiment of the disclosure;
FIG. 7 illustrates a ROC curve after dynamically fusing models in an embodiment of the disclosure;
FIG. 8 is a schematic diagram of an automobile driving accident prediction device according to an embodiment of the present disclosure;
fig. 9 shows a block diagram of an electronic device in an embodiment of the present disclosure.
Detailed Description
Example embodiments will now be described more fully with reference to the accompanying drawings. Example embodiments may, however, be embodied in many different forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the concept of example embodiments to those skilled in the art. The described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.
Furthermore, the drawings are merely schematic illustrations of the present disclosure and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and thus their repetitive description will be omitted. Some of the block diagrams shown in the figures are functional entities and do not necessarily correspond to physically or logically separate entities. These functional entities may be implemented in the form of software, or in one or more hardware modules or integrated circuits, or in different networks and/or processor devices and/or microcontroller devices.
According to the scheme, by means of big data analysis, based on driving data in the automobile driving process, a Support Vector Regression (SVR) algorithm, a K-Nearest Neighbors (KNN) algorithm And a Classification Regression Tree (CART) algorithm are adopted to respectively train And learn, integrated prediction results under an optimization strategy with the smallest loss value in each independent Classification algorithm are obtained by setting different non-sensitive region width values, neighborhood values And pruning rates, standard differences And correlation coefficients are calculated based on evaluation index values of the models, so that a dynamic fusion model capable of predicting automobile driving accidents is obtained, the prediction values of the traffic accidents in the automobile driving process are automatically predicted, strong overall prediction capability is achieved, the fusion efficiency of the models is improved, And user experience is good.
The scheme provided by the embodiment of the application relates to technologies such as artificial intelligence big data processing, machine learning and automatic driving, is a software program applied to a computer, utilizes driving data in a test automobile driving process to construct a dynamic fusion model, and obtains an automobile driving accident prediction value based on the dynamic fusion model and the driving data in the automobile driving process, thereby dynamically predicting automobile driving risks, reducing the occurrence probability of traffic accidents, improving the fusion efficiency of the model, improving the user experience degree, and specifically explaining the method through the following embodiments:
first, the embodiment of the present disclosure provides a method for predicting a driving accident of an automobile, which may be performed by any system with computing capability.
Fig. 1 shows a flowchart of a method for predicting an automobile driving accident in an embodiment of the present disclosure, and as shown in fig. 1, the method for predicting an automobile driving accident provided in the embodiment of the present disclosure includes the following steps:
s102, collecting driving data in the driving process of the automobile in real time;
in this embodiment, a radar speed meter and a GPS testing system are installed on the vehicle to collect driving data of the driving of the vehicle, where the driving data includes time, speed, driving acceleration, distance to the vehicle ahead, driving lane, geographical position, gear engaged by the driver, braking and deceleration time, weather visibility, and other data.
The acquisition mode to driving data still can adopt other distance measuring means to obtain, like laser rangefinder, big dipper positioning system etc. this application does not specifically restrict, as long as can gather driving data in real time can.
S104, inputting driving data into a pre-constructed dynamic fusion model to obtain a predicted value of the automobile driving accident;
in this embodiment, after the dynamic fusion model is trained, an application program is generated according to the dynamic fusion model in a packaging manner, the application program is deployed and installed in a control system of an automobile, driving data in the driving process of the automobile is collected through a radar velocimeter and a GPS test system, and then the driving data is input into the dynamic fusion model, so that a driving accident prediction value of the automobile can be obtained, the driving risk of the automobile is dynamically predicted, and the occurrence probability of traffic accidents is reduced.
And S106, dynamically predicting the driving risk of the automobile according to the relation between the obtained predicted value of the driving accident of the automobile and a preset driving safety threshold value.
Specifically, when the predicted value of the driving accident of the automobile obtained by the dynamic fusion model is greater than or equal to a preset driving safety threshold value, the risk level of the accident is judged to be low, and no measure is needed; when the driving accident prediction value is smaller than the preset driving safety threshold value, the risk level of accidents is judged to be high, measures are required to be taken, and the conditions of collision, scratch and the like are prevented, such as reducing the speed of the automobile, increasing the distance from the front automobile and the like.
For example, a judgment threshold of a driving safety result is set to be 0.5, and if the obtained dynamic fusion prediction value is greater than or equal to 0.5, the driving safety is determined; if the obtained dynamic fusion predicted value is less than 0.5, the safety risk is determined to exist, so that the automobile driving risk is dynamically predicted, the vigilance awareness of a driver is improved, and driving accidents are avoided.
The preset driving safety threshold value of the embodiment can be determined according to actual conditions, a plurality of driving safety threshold values can be set according to different risk levels, and various different measures are taken according to different risk levels, so that the risk is kept in a reasonable range.
The automobile driving accident prediction method provided by the embodiment of the disclosure collects driving data in the automobile driving process, inputs the collected driving data into a pre-constructed dynamic fusion model, automatically predicts the automobile driving accident prediction value in the automobile driving process, and dynamically predicts the automobile driving risk according to the relationship between the prediction value and the preset driving safety threshold value, thereby reducing the occurrence probability of traffic accidents and improving the user experience.
In order to accurately predict the predicted value of the driving accident, the construction of the dynamic fusion model is a key link, fig. 2 shows a flow chart of constructing the dynamic fusion model in the embodiment of the present disclosure, as shown in fig. 2, in an embodiment of the present disclosure, the dynamic fusion model is constructed in the following manner:
s202, collecting test driving data in the driving process of the automobile during the test;
in this embodiment, a radar speed meter and a GPS test system are installed on a test vehicle to collect test driving data of the test vehicle driving, where the test driving data includes time, vehicle speed, driving acceleration, distance to the front vehicle, driving lane, geographical position, gear engaged by a driver, braking and deceleration time, weather visibility, weather state, driving safety, and other data, where the driving safety is a predicted target row, and its value is a Boolean value, and is taken as 0 or 1.
S204, preprocessing the test driving data to obtain training sample data;
in this embodiment, the collected test driving data is subjected to data preprocessing, including missing value filling, type conversion, normalization processing, and equalization processing.
For example, when filling missing items to process data, for non-fixed-distance data, the distribution of a certain field is statistically analyzed, and the mode with the highest frequency of occurrence is used to complement the missing items, for example, the value of the weather condition may be sunny day, cloudy day, haze weather, sand storm, rainy day, snow day, and the field cannot be calculated.
And for interval-based data, the position of the interval-based data in the ordered arrangement is represented by using numbers, missing items are completed by the average value of the existing values of the field, for example, the visibility is completed by calculating the average value. From this, part of the raw data of table 1 was obtained.
TABLE 1
Because the data set has the text description fields which are not beneficial to later data training, the data need to be subjected to type conversion, the meaning represented by each field is converted into numerical description which is convenient for quantization, by taking the meteorological visibility field as an example, the specific numerical value of visibility has lower value for the process of the data, and the external definition degree has more reference value, so that the visibility is mapped in sequence from high to low, and the corresponding relation is that when the visibility is greater than or equal to 10Km, the visual definition degree is marked as 0; when the visibility is between 1Km and 10Km, the state is light fog, and the clear degree of the visual field is marked as 1; when the visibility is between 0.3Km and 1Km, the fog state is realized, and the clear degree of the visual field is marked as 2; when the visibility is less than 0.3Km, the fog state is achieved, and the clear degree of the visual field is marked as 3. For the weather status field, a clear day is denoted as 0, a rain is denoted as 1, a snow is denoted as 2, a haze is denoted as 3, and a sand storm is denoted as 4. For the driving safety status field, safety is noted as 0 and danger is noted as 1. After the data type conversion and normalization processing, the data application value and efficiency can be improved, and partial data subjected to data type conversion in the table 2 can be obtained.
TABLE 2
In this embodiment, because there is a problem that dimensions of each data set are not consistent, in order to eliminate differences in characteristic dimensions and specifications, a normalization processing method is required to be used to normalize the collected test driving data to obtain normalized test data, and thus training sample data is formed.
The normalization process is performed according to the following formula, which can be used but is not limited to the following formula:
wherein i is the data number, j is the variable number, xi,jDenotes the j variable, X, in the non-normalized i group of datajRepresenting the set formed by variable data values corresponding to all j, and min represents the experimental driving data after abnormal points are removedAnd the max represents the maximum value of the relevant variable in the test driving data after the abnormal point is removed.
As a common method for scaling the interval, the normalization process uses the boundary value information to scale the value interval of the feature to a specific range, and normalizes the feature in the data set by using the MinMaxScaler class in the preprocessing library so that the numerical range of each feature is limited to 0 to 1, thereby obtaining the normalized partial data in table 3.
TABLE 3
And S206, constructing a dynamic fusion model according to the training sample data based on a preset classification algorithm.
Specifically, the preset classification algorithm of the present embodiment includes a support vector regression SVR algorithm, a K-nearest neighbor KNN algorithm, and a classification regression tree CART algorithm.
During training, in order to avoid the problems of over-fitting and under-fitting caused by the particularity of data and the limitation of an algorithm in the model training process, training sample data is divided into training set data and test set data according to a preset proportion, and generally, the training set data and the test set data are divided into 7: 3, the resolution is carried out. The training set data is used for training each algorithm model to be verified, and the test set data is used for determining the model weight of each algorithm model, so that a dynamic fusion model is constructed according to each algorithm model and the model weight, predicted values obtained by the dynamic fusion model are more in line with actual conditions, and the accuracy of automobile driving accident prediction is improved.
When the model is trained, the optimal parameter combination is obtained by using a grid searching parameter adjusting mode, each hyper-parameter combination and the cross verification times are received by using a grid searching method GridSearchCV (), the hyper-parameter combination with the optimal performance is selected, the characteristics of a user are selected as X, the driving safety condition label result of a test vehicle is collected as Y, and the model function of each algorithm is constructed.
According to the automobile driving accident prediction method provided by the embodiment of the disclosure, the test driving data of the test vehicle is collected and preprocessed, and then a dynamic fusion model is constructed based on the SVR algorithm, the KNN algorithm and the CART algorithm, so that the automobile driving accident prediction value is predicted according to the real-time driving data, the overall prediction capability is improved, and the model fusion efficiency is improved.
In an embodiment of the present disclosure, step S206 is based on a preset algorithm, and a dynamic fusion model is constructed according to training sample data, which specifically includes:
s302, respectively inputting training set data in training sample data into each preset classification algorithm to respectively obtain an algorithm model to be verified of each classification algorithm;
s304, inputting the test set data in the training sample data into the algorithm model to be verified of each classification algorithm to obtain three evaluation indexes of Mean Square Error (MSE), R2 and the explained variance fraction of each algorithm model;
s306, calculating a standard deviation and a correlation coefficient based on the three evaluation indexes to obtain fusion weights among the classification algorithms, and constructing a dynamic fusion model according to the fusion weights among the classification algorithms.
Specifically, the training sample data is as follows: and 3, dividing the training set data and the test set data in proportion, presetting classification algorithms including an SVR algorithm, a KNN algorithm and a CART algorithm, and inputting the training set data into the three algorithms respectively to obtain an algorithm model to be verified of the three algorithms.
And inputting the remaining 30% of the test set data into the obtained algorithm model to be verified to obtain three evaluation indexes of mean square error, R2 and interpretation variance fraction of each algorithm model.
Wherein the mean square error is
yiFor the true result value of the sample, yi' predict result value for sample data.
The R2 metric is the degree to which the variance of the dependent variable of value is interpreted by the independent variable, reflecting the degree to which the model fits the data, and R2 can be expressed as:
wherein,
is the average true value of the sample data.
The Explained variance score (extended variance score) refers to the degree of influence of an independent variable on the variance change of a dependent variable, and is obtained by the following formula:
and calculating a standard deviation and a correlation coefficient according to three evaluation indexes of mean square error, R2 and the explained variance fraction of each algorithm model to obtain fusion weight among all classification algorithms, thereby constructing a dynamic fusion model.
In an embodiment of the present disclosure, as shown in fig. 4, step S302 respectively inputs training set data in training sample data into each preset classification algorithm, and respectively obtains an algorithm model to be verified of each classification algorithm, where the method includes:
s402, inputting the training set data into an SVR algorithm, and constructing an SVR model function and an SVR loss function;
s404, obtaining different non-sensitive area width values, and obtaining SVR classification functions and SVR function weights corresponding to the non-sensitive area width values to minimize loss values;
s406, obtaining a driving accident prediction value of the SVR algorithm according to the width value of the non-sensitive area and the SVR function weight;
and S408, calculating a SVR algorithm weighting decision prediction value formula according to the SVR function weight and the driving accident prediction value of the SVR algorithm to obtain an SVR algorithm model to be verified.
Specifically, an SVR function S (x) and a loss function l (x) are constructed based on an input feature x, a tag result y and a mapping function phi (x), wherein,
S(x)＝w×φ(x)+b
l (x) ═ wx phi (x) + b-y formula two
Setting different non-sensitive area width values epsiloni(i belongs to n) so as to obtain a plurality of corresponding SVR functions, realize the minimization of loss value as the target adjustment and distribute the initial weight, and ensure the weight w of each SVR functionεiThe following condition is satisfied:
according to n different non-sensitive area width values epsiloniAnd SVR function weight wεiThe corresponding predicted value S of the driving accident can be obtaineds1(x)，Ss2(x)…Ssn(x) (i belongs to n), so as to obtain a predicted value W' of the weighting decision of the SVR algorithmSVRThe formula:
the above formula is the obtained SVR algorithm model to be verified.
In an embodiment of the present disclosure, as shown in fig. 5, step S302 respectively inputs training set data in training sample data into each preset classification algorithm, and respectively obtains an algorithm model to be verified of each classification algorithm, where the method includes:
s502, inputting the training set data into a KNN algorithm, and constructing a KNN classification function and a KNN loss function;
s504, obtaining different neighborhood values, and obtaining a KNN classification function and a KNN function weight corresponding to the neighborhood values to minimize a loss value;
s506, obtaining a driving accident prediction value of the KNN algorithm according to the neighborhood value and the KNN function weight;
and S508, calculating a weighted decision prediction value formula of the KNN algorithm according to the KNN function weight and the driving accident prediction value of the KNN algorithm to obtain the KNN algorithm model to be verified.
After the KNN algorithm is constructed, a KNN classification function S (x) and a loss function l (x) are constructed based on the input features x and the label result y, wherein the KNN classification function can be an Euclidean distance model function,
S(x)＝w(x)+b
l (x) ═ w (x) + b-y equation five
Setting different neighborhood values ki(i belongs to n) to obtain a plurality of corresponding KNN functions, so as to realize the minimization of loss value and allocate initial weight for target adjustment, and ensure the weight W of each functionkiThe following condition is satisfied:
according to n different neighborhood values kiSum function weight WkiThe corresponding predicted value S of the driving accident can be obtainedk1(x)，Sk2(x)…Skn(x) (i belongs to n), and calculating to obtain a predicted value W' of the weighting decision of the KNN algorithmKNNThe formula:
the above formula is the obtained KNN algorithm model to be verified.
In an embodiment of the present disclosure, as shown in fig. 6, step S302 respectively inputs training set data in training sample data into each preset classification algorithm, and respectively obtains an algorithm model to be verified of each classification algorithm, where the method includes:
s602, inputting training set data into a CART algorithm to construct a CART loss function;
s604, obtaining different pruning rates, obtaining subtrees corresponding to the pruning rates and CART function weights, and enabling loss values to be minimum;
s606, obtaining a driving accident prediction value of the CART algorithm according to the subtrees and the pruning rate;
s608, calculating a CART algorithm weighting decision prediction value formula according to the CART function weight and the driving accident prediction value of the CART algorithm to obtain a CART algorithm model to be verified.
In particular, in the construction ofAfter the CART algorithm, a loss function C alpha (T) ═ Gini (x) + alpha n is constructed based on input features x and the number n of subnodes, wherein alpha is a pruning rate and is used for balancing the fitting degree and the subtree complexity of training data, and a plurality of subtrees T are obtained by adjusting alpha values1，T2…Tn(i belongs to n), distributing initial weight for realizing minimization of loss value as target adjustment, and ensuring weight w of each subtreeTiThe following conditions are satisfied:
from n different subtrees TiAnd subtree weight wTiThe driving accident prediction value S of the CART algorithm can be obtainedT1(x)，ST2(x)…STn(x) (i belongs to n), and calculating to obtain a predicted value W' of the weighting decision of the CART algorithmCARTThe formula:
the above formula is the obtained CART algorithm model to be verified.
In the embodiment, after the algorithm model to be verified is obtained, the test set data is input into the SVR algorithm model, the KNN algorithm model and the CART algorithm model to be verified respectively, the mean square error, R2 and the interpretation variance score of each algorithm model are obtained through calculation, and the three classification algorithms are evaluated and analyzed through three dimensions.
Calculating standard deviations and correlation coefficients of the three evaluation indexes, fitting fusion weights among the algorithm models, and recording the fusion weights of the SVR algorithm models as W'SVRAnd the fusion weight of the KNN algorithm model is recorded as W'KNNThe fusion weight of the CART algorithm model is recorded as W'CARTThen, the formula of the finally obtained dynamic fusion model is as follows:
in order to verify the performance of the dynamic fusion model improvement, the model performance evaluation work is carried out, and the dynamic fusion model is evaluated according to four evaluation indexes of the model performance, namely Accuracy, Precision, Recall and F1-score.
The label result of the input test data is compared with the prediction result obtained by each classification algorithm in the test to obtain the evaluation index data of the classification algorithm and the dynamic fusion model, as shown in table 4.
TABLE 4
| Model (model) | Accuracy of | Accuracy of | Recall rate | F1-score |
| SVR | 0.779 | 0.793 | 0.811 | 0.801 |
| KNN | 0.821 | 0.837 | 0.803 | 0.817 |
| CART | 0.801 | 0.831 | 0.817 | 0.825 |
| Dynamic fusion | 0.832 | 0.852 | 0.801 | 0.833 |
As can be seen from table 4, in each of the independent classification models, the KNN model is first located at 0.821 in the accuracy dimension, the KNN model is first located at 0.837 in the accuracy dimension, and the CART model is first located at 0.817 in the regression dimension. The dynamic fusion model is slightly higher than the KNN model by 0.832 in accuracy, and is centered at the top by 0.852 in accuracy, but the regression rate is lower. At this time, it is necessary to focus on the analysis of the F1-score index, because the F1 value is taken as a harmonic mean of the accuracy and the regression rate, and the model performance can be more comprehensively evaluated, so that from the point of view of the F1-score index, the dynamic fusion model shows more excellent overall prediction performance with 0.833 than three independent classification models.
A Receiver Operator Characteristic Curve (ROC Curve) can reflect the relation between continuous variables of sensitivity and specificity, and the Curve is drawn by taking the sensitivity as an ordinate to represent a true positive rate and taking a difference value between 1 and the specificity as an abscissa to represent a false positive rate. As shown in FIG. 7, the ROC curve is close to the upper left corner, and the AUC (area Under Current) value represented by the area enclosed by the whole ROC curve reaches 0.84, so that the dynamic fusion model has excellent prediction effect and generalization performance.
The automobile driving accident prediction method provided by the embodiment of the disclosure is based on experimental driving data of experimental automobile driving, training and learning are respectively performed by adopting a Support Vector Regression (SVR) algorithm, a K nearest KNN (K nearest neighbor) classification algorithm and a classification regression tree (CART) algorithm, an integrated prediction result under an optimization strategy with the smallest loss value in each independent classification algorithm is obtained by setting different non-sensitive region width values, neighborhood values and pruning rates, a standard deviation and a correlation coefficient are calculated based on evaluation index values of each model, so that a dynamic fusion model capable of predicting automobile driving accidents is obtained, a prediction value of traffic accidents occurring in the automobile driving process is automatically predicted, and therefore, the strong overall prediction capability is achieved, the fusion efficiency of the models is improved, and the user experience is good.
Based on the same inventive concept, the embodiment of the present disclosure further provides a device for predicting a driving accident of an automobile, as described in the following embodiments. Because the principle of the embodiment of the apparatus for solving the problem is similar to that of the embodiment of the method, the embodiment of the apparatus can be implemented by referring to the implementation of the embodiment of the method, and repeated details are not described again.
Fig. 8 is a schematic diagram illustrating an automobile driving accident prediction apparatus according to an embodiment of the present disclosure, and as shown in fig. 8, the apparatus includes:
the acquisition module 801 is used for acquiring driving data in the driving process of the automobile in real time;
the prediction module 802 is configured to input the driving data into a pre-constructed dynamic fusion model to obtain a predicted value of the driving accident;
and the comparison module 803 is configured to dynamically predict the driving risk of the vehicle according to the relationship between the obtained predicted value of the driving accident of the vehicle and a preset driving safety threshold.
In one embodiment of the present disclosure, the apparatus further comprises a data processing module and a fusion model building module, not shown in the figures,
the acquisition module 801 is further used for acquiring test driving data in the automobile driving process during the test;
the data processing module is used for preprocessing the test driving data to obtain training sample data;
and the fusion model construction module is used for constructing a dynamic fusion model according to the training sample data based on a preset classification algorithm.
Specifically, the preset classification algorithm of the present embodiment includes a support vector regression SVR algorithm, a K-nearest neighbor KNN algorithm, and a classification regression tree CART algorithm.
In one embodiment of the present disclosure, the fusion model construction module includes a pre-training module, a predicted value output sub-module, and a fusion weight calculation sub-module, which are not shown in the drawings,
the pre-training module is used for respectively inputting training set data in the training sample data into each preset classification algorithm to respectively obtain an algorithm model to be verified of each classification algorithm;
the model verification submodule is used for inputting the test set data in the training sample data into the algorithm model to be verified of each classification algorithm to obtain three evaluation indexes of mean square error, R2 and interpretation variance fraction of each algorithm model;
and the fusion weight calculation submodule is used for calculating the standard deviation and the correlation coefficient based on the three evaluation indexes to obtain the fusion weight among the classification algorithms, and constructing a dynamic fusion model according to the fusion weight among the classification algorithms.
In one embodiment of the present disclosure, the pre-training module includes an SVR training module,
the SVR training module is used for inputting the training set data into an SVR algorithm to construct an SVR model function and an SVR loss function;
obtaining different width values of the insensitive region, and obtaining an SVR classification function and an SVR function weight corresponding to the width values of the insensitive region to minimize a loss value;
obtaining a driving accident prediction value of the SVR algorithm according to the width value of the non-sensitive area and the weight of the SVR function;
and calculating a SVR algorithm weighting decision prediction value formula according to the SVR function weight and the driving accident prediction value of the SVR algorithm to obtain an SVR algorithm model to be verified.
In one embodiment of the present disclosure, the pre-training module includes a KNN training module,
the KNN training module is used for inputting the training set data into a KNN algorithm to construct a KNN classification function and a KNN loss function;
acquiring different neighborhood values to obtain a KNN classification function and a KNN function weight corresponding to the neighborhood values, so that the loss value is minimum;
obtaining a driving accident prediction value of the KNN algorithm according to the neighborhood value and the KNN function weight;
and calculating a weighted decision prediction value formula of the KNN algorithm according to the KNN function weight and the driving accident prediction value of the KNN algorithm to obtain the KNN algorithm model to be verified.
In one embodiment of the present disclosure, the pre-training module further comprises a CART training module,
the CART training module is used for inputting training set data into a CART algorithm to construct a CART loss function;
obtaining different pruning rates, and obtaining subtrees corresponding to the pruning rates and CART function weights to minimize loss values;
obtaining a driving accident prediction value of the CART algorithm according to the subtrees and the pruning rate;
and calculating a CART algorithm weighting decision prediction value formula according to the CART function weight and the driving accident prediction value of the CART algorithm to obtain a CART algorithm model to be verified.
As will be appreciated by one skilled in the art, aspects of the present invention may be embodied as a system, method or program product. Thus, various aspects of the invention may be embodied in the form of: an entirely hardware embodiment, an entirely software embodiment (including firmware, microcode, etc.) or an embodiment combining hardware and software aspects that may all generally be referred to herein as a "circuit," module "or" system.
The embodiment of the disclosure provides a method and a device for predicting automobile driving accidents, which collects driving data in the driving process of an automobile, inputs the collected driving data into a pre-constructed dynamic fusion model, automatically predicts the predicted value of the traffic accidents in the driving process of the automobile, dynamically predicts the driving risks of the automobile, reduces the probability of the traffic accidents, simultaneously, adopts SVR algorithm, KNN algorithm and CART algorithm to respectively train and learn based on the experimental driving data of the automobile driving, obtains an integrated prediction result under an optimization strategy with the minimum loss value in each independent classification algorithm by setting different non-sensitive area width values, neighborhood values and pruning rates, calculates standard deviation and correlation coefficient based on the evaluation index value of each model, thereby obtaining a dynamic fusion model capable of predicting the automobile driving accidents, and automatically predicts the predicted value of the traffic accidents in the driving process of the automobile, therefore, the method has the advantages of achieving stronger overall prediction capability, improving the fusion efficiency of the models and having good user experience.
An electronic device 900 according to this embodiment of the invention is described below with reference to fig. 9. The electronic device 900 shown in fig. 9 is only an example and should not bring any limitations to the function and scope of use of the embodiments of the present invention.
As shown in fig. 9, the electronic device 900 is embodied in the form of a general purpose computing device. Components of electronic device 900 may include, but are not limited to: the at least one processing unit 910, the at least one memory unit 920, and a bus 930 that couples various system components including the memory unit 920 and the processing unit 910.
Wherein the storage unit stores program code that is executable by the processing unit 910 to cause the processing unit 910 to perform steps according to various exemplary embodiments of the present invention described in the above section "exemplary methods" of the present specification. For example, the processing unit 910 may perform the real-time collection of driving data during the driving process of the automobile as shown in fig. 1; inputting the driving data into a pre-constructed dynamic fusion model to obtain a predicted value of the automobile driving accident; and dynamically predicting the driving risk of the automobile according to the relation between the obtained predicted value of the driving accident of the automobile and a preset driving safety threshold value.
The storage unit 920 may include a readable medium in the form of a volatile storage unit, such as a random access memory unit (RAM)9201 and/or a cache memory unit 9202, and may further include a read only memory unit (ROM) 9203.
The electronic device 900 may also communicate with one or more external devices 940 (e.g., keyboard, pointing device, bluetooth device, etc.), with one or more devices that enable a user to interact with the electronic device 900, and/or with any devices (e.g., router, modem, etc.) that enable the electronic device 900 to communicate with one or more other computing devices. Such communication may occur via input/output (I/O) interface 950. Also, the electronic device 900 may communicate with one or more networks (e.g., a Local Area Network (LAN), a Wide Area Network (WAN) and/or a public network, such as the Internet) via the network adapter 960. As shown, the network adapter 960 communicates with the other modules of the electronic device 900 via the bus 930. It should be appreciated that although not shown, other hardware and/or software modules may be used in conjunction with the electronic device 900, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems, among others.
Through the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described herein may be implemented by software, or by software in combination with necessary hardware. Therefore, the technical solution according to the embodiments of the present disclosure may be embodied in the form of a software product, which may be stored in a non-volatile storage medium (which may be a CD-ROM, a usb disk, a removable hard disk, etc.) or on a network, and includes several instructions to enable a computing device (which may be a personal computer, a server, a terminal device, or a network device, etc.) to execute the method according to the embodiments of the present disclosure.
In an exemplary embodiment of the present disclosure, there is also provided a computer-readable storage medium having stored thereon a program product capable of implementing the above-described method of the present specification. In some possible embodiments, aspects of the invention may also be implemented in the form of a program product comprising program code means for causing a terminal device to carry out the steps according to various exemplary embodiments of the invention described in the above section "exemplary methods" of the present description, when said program product is run on the terminal device.
A program product for implementing the above method according to an embodiment of the present invention is described, which may employ a portable compact disc read only memory (CD-ROM) and include program code, and may be run on a terminal device, such as a personal computer. However, the program product of the present invention is not limited in this regard and, in the present document, a readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device.
The program product may employ any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the foregoing. More specific examples (a non-exhaustive list) of the readable storage medium include: an electrical connection having one or more wires, a portable disk, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
A computer readable signal medium may include a propagated data signal with readable program code embodied therein, for example, in baseband or as part of a carrier wave. Such a propagated data signal may take many forms, including, but not limited to, electro-magnetic, optical, or any suitable combination thereof. A readable signal medium may also be any readable medium that is not a readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device.
Program code embodied on a readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.
Program code for carrying out operations for aspects of the present invention may be written in any combination of one or more programming languages, including an object oriented programming language such as Java, C + + or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computing device, partly on the user's device, as a stand-alone software package, partly on the user's computing device and partly on a remote computing device, or entirely on the remote computing device or server. In the case of a remote computing device, the remote computing device may be connected to the user computing device through any kind of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or may be connected to an external computing device (e.g., through the internet using an internet service provider).
It should be noted that although in the above detailed description several modules or units of the device for action execution are mentioned, such a division is not mandatory. Indeed, the features and functionality of two or more modules or units described above may be embodied in one module or unit, according to embodiments of the present disclosure. Conversely, the features and functions of one module or unit described above may be further divided into embodiments by a plurality of modules or units.
Moreover, although the steps of the methods of the present disclosure are depicted in the drawings in a particular order, this does not require or imply that the steps must be performed in this particular order, or that all of the depicted steps must be performed, to achieve desirable results. Additionally or alternatively, certain steps may be omitted, multiple steps combined into one step execution, and/or one step broken down into multiple step executions, etc.
Through the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described herein may be implemented by software, or by software in combination with necessary hardware. Therefore, the technical solution according to the embodiments of the present disclosure may be embodied in the form of a software product, which may be stored in a non-volatile storage medium (which may be a CD-ROM, a usb disk, a removable hard disk, etc.) or on a network, and includes several instructions to enable a computing device (which may be a personal computer, a server, a mobile terminal, or a network device, etc.) to execute the method according to the embodiments of the present disclosure.
Other embodiments of the disclosure will be apparent to those skilled in the art from consideration of the specification and practice of the disclosure disclosed herein. This application is intended to cover any variations, uses, or adaptations of the disclosure following, in general, the principles of the disclosure and including such departures from the present disclosure as come within known or customary practice within the art to which the disclosure pertains. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the disclosure being indicated by the following claims.
Claims (10)
1. A method for predicting a driving accident of an automobile, comprising:
collecting driving data in the driving process of the automobile in real time;
inputting the driving data into a pre-constructed dynamic fusion model to obtain a predicted value of the automobile driving accident;
and dynamically predicting the driving risk of the automobile according to the relation between the obtained predicted value of the driving accident of the automobile and a preset driving safety threshold value.
2. The method of claim 1, wherein the dynamic fusion model is constructed by:
collecting test driving data in the driving process of the automobile during the test;
preprocessing the test driving data to obtain training sample data;
and constructing a dynamic fusion model according to the training sample data based on a preset classification algorithm.
3. The method of claim 2, wherein the predetermined classification algorithm comprises Support Vector Regression (SVR) algorithm, K-nearest neighbor (KNN) algorithm and classification regression tree (CART) algorithm.
4. The method according to claim 2, wherein the building a dynamic fusion model according to the training sample data based on a preset classification algorithm comprises:
respectively inputting training set data in the training sample data into each preset classification algorithm to respectively obtain an algorithm model to be verified of each classification algorithm;
inputting test set data in the training sample data into algorithm models to be verified of all classification algorithms to obtain three evaluation indexes of mean square error, R2 and explained variance fraction of all algorithm models;
and calculating a standard deviation and a correlation coefficient based on the three evaluation indexes to obtain fusion weights among the classification algorithms, and constructing a dynamic fusion model according to the fusion weights among the classification algorithms.
5. The method according to claim 4, wherein the step of inputting training set data in the training sample data into each preset classification algorithm to obtain the algorithm model to be verified of each classification algorithm comprises:
inputting the training set data into an SVR algorithm, and constructing an SVR model function and an SVR loss function;
obtaining different width values of the insensitive region, and obtaining an SVR classification function and an SVR function weight corresponding to the width values of the insensitive region to minimize a loss value;
obtaining a driving accident prediction value of the SVR algorithm according to the width value of the non-sensitive area and the weight of the SVR function;
and calculating a SVR algorithm weighting decision prediction value formula according to the SVR function weight and the driving accident prediction value of the SVR algorithm to obtain an SVR algorithm model to be verified.
6. The method according to claim 4, wherein the step of inputting training set data in the training sample data into each preset classification algorithm to obtain the algorithm model to be verified of each classification algorithm comprises:
inputting the training set data into a KNN algorithm, and constructing a KNN classification function and a KNN loss function;
acquiring different neighborhood values to obtain a KNN classification function and a KNN function weight corresponding to the neighborhood values, so that the loss value is minimum;
obtaining a driving accident prediction value of the KNN algorithm according to the neighborhood value and the KNN function weight;
and calculating a weighted decision prediction value formula of the KNN algorithm according to the KNN function weight and the driving accident prediction value of the KNN algorithm to obtain the KNN algorithm model to be verified.
7. The method according to claim 4, wherein the step of inputting training set data in the training sample data into each preset classification algorithm to obtain the algorithm model to be verified of each classification algorithm comprises:
inputting training set data into a CART algorithm to construct a CART loss function;
obtaining different pruning rates, and obtaining subtrees corresponding to the pruning rates and CART function weights to minimize loss values;
obtaining a driving accident prediction value of the CART algorithm according to the subtrees and the pruning rate;
and calculating a CART algorithm weighting decision prediction value formula according to the CART function weight and the driving accident prediction value of the CART algorithm to obtain a CART algorithm model to be verified.
8. An automobile driving accident prediction apparatus, comprising:
the acquisition module is used for acquiring driving data in the driving process of the automobile in real time;
the prediction module is used for inputting the driving data into a pre-constructed dynamic fusion model so as to obtain a predicted value of the automobile driving accident;
and the comparison module is used for dynamically predicting the automobile driving risk according to the relation between the obtained automobile driving accident prediction value and a preset driving safety threshold value.
9. An electronic device, comprising: a processor; and a memory for storing executable instructions of the processor; wherein the processor is configured to perform the method of predicting an automobile driving accident according to any one of claims 1-7 via execution of the executable instructions.
10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out a method of predicting a driving accident of a vehicle according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111623475.2A CN114299633B (en) | 2021-12-28 | 2021-12-28 | Automobile driving accident prediction method and device, electronic equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111623475.2A CN114299633B (en) | 2021-12-28 | 2021-12-28 | Automobile driving accident prediction method and device, electronic equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114299633A true CN114299633A (en) | 2022-04-08 |
| CN114299633B CN114299633B (en) | 2024-07-23 |
Family
ID=80971761
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111623475.2A Active CN114299633B (en) | 2021-12-28 | 2021-12-28 | Automobile driving accident prediction method and device, electronic equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114299633B (en) |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090063115A1 (en) * | 2007-08-31 | 2009-03-05 | Zhao Lu | Linear programming support vector regression with wavelet kernel |
| CN102005135A (en) * | 2010-12-09 | 2011-04-06 | 上海海事大学 | Genetic algorithm-based support vector regression shipping traffic flow prediction method |
| CN104408907A (en) * | 2014-10-31 | 2015-03-11 | 重庆大学 | Highway traffic incident duration time prediction method with on-line optimization capability |
| CN110288096A (en) * | 2019-06-28 | 2019-09-27 | 江苏满运软件科技有限公司 | Prediction model training and prediction technique, device, electronic equipment, storage medium |
| CN110489790A (en) * | 2019-07-10 | 2019-11-22 | 合肥工业大学 | Based on the IGBT junction temperature prediction technique for improving ABC-SVR |
| CN110807930A (en) * | 2019-11-07 | 2020-02-18 | 中国联合网络通信集团有限公司 | Dangerous vehicle early warning method and device |
| CN110827088A (en) * | 2019-11-07 | 2020-02-21 | 深圳鼎然信息科技有限公司 | Vehicle cost prediction method and device based on big data and storage medium |
| CN111080551A (en) * | 2019-12-13 | 2020-04-28 | 太原科技大学 | Multi-label image completion method based on depth convolution characteristics and semantic neighbor |
| CN112200293A (en) * | 2020-11-02 | 2021-01-08 | 吉林大学 | CART-AMV improved random forest algorithm |
| CN112435077A (en) * | 2020-12-11 | 2021-03-02 | 四川长虹电器股份有限公司 | Sales forecasting method |
-
2021
- 2021-12-28 CN CN202111623475.2A patent/CN114299633B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090063115A1 (en) * | 2007-08-31 | 2009-03-05 | Zhao Lu | Linear programming support vector regression with wavelet kernel |
| CN102005135A (en) * | 2010-12-09 | 2011-04-06 | 上海海事大学 | Genetic algorithm-based support vector regression shipping traffic flow prediction method |
| CN104408907A (en) * | 2014-10-31 | 2015-03-11 | 重庆大学 | Highway traffic incident duration time prediction method with on-line optimization capability |
| CN110288096A (en) * | 2019-06-28 | 2019-09-27 | 江苏满运软件科技有限公司 | Prediction model training and prediction technique, device, electronic equipment, storage medium |
| CN110489790A (en) * | 2019-07-10 | 2019-11-22 | 合肥工业大学 | Based on the IGBT junction temperature prediction technique for improving ABC-SVR |
| CN110807930A (en) * | 2019-11-07 | 2020-02-18 | 中国联合网络通信集团有限公司 | Dangerous vehicle early warning method and device |
| CN110827088A (en) * | 2019-11-07 | 2020-02-21 | 深圳鼎然信息科技有限公司 | Vehicle cost prediction method and device based on big data and storage medium |
| CN111080551A (en) * | 2019-12-13 | 2020-04-28 | 太原科技大学 | Multi-label image completion method based on depth convolution characteristics and semantic neighbor |
| CN112200293A (en) * | 2020-11-02 | 2021-01-08 | 吉林大学 | CART-AMV improved random forest algorithm |
| CN112435077A (en) * | 2020-12-11 | 2021-03-02 | 四川长虹电器股份有限公司 | Sales forecasting method |
Non-Patent Citations (2)
| Title |
|---|
| 张伟: "基于机器学习的航天器故障预测算法研究", 基于机器学习的航天器故障预测算法研究, pages 10 - 28 * |
| 许岩岩 等: "高速路交通流短时预测方法", 交通运输工程学报, vol. 13, no. 02, pages 114 - 119 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114299633B (en) | 2024-07-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| KR102539942B1 (en) | Method and apparatus for training trajectory planning model, electronic device, storage medium and program | |
| WO2022005576A1 (en) | Systems and methods for optimizing trajectory planner based on human driving behaviors | |
| JP2021157833A (en) | Road condition prediction method, method for establishing road condition prediction model and corresponding device thereof, electronic apparatus, recording medium, and computer program product | |
| CN113283665A (en) | Urban traffic accident risk prediction method based on road network | |
| Wang et al. | ARIMA Model and Few‐Shot Learning for Vehicle Speed Time Series Analysis and Prediction | |
| CN116205024A (en) | Self-adaptive automatic driving dynamic scene general generation method for high-low dimension evaluation scene | |
| CN118470974A (en) | Driving safety early warning system and method based on Internet of vehicles | |
| CN116453343A (en) | Intelligent traffic signal control optimization algorithm, software and system based on flow prediction in intelligent networking environment | |
| CN117932763A (en) | Expressway traffic model construction method based on digital twin | |
| CN117273964B (en) | Intelligent vehicle insurance policy generation system and method for self-adaptive driving data | |
| CN113792598B (en) | Vehicle-mounted camera-based vehicle collision prediction system and method | |
| CN113096405B (en) | Construction method of prediction model, and vehicle accident prediction method and device | |
| CN117392834B (en) | Intelligent route planning method and system | |
| CN114299633B (en) | Automobile driving accident prediction method and device, electronic equipment and storage medium | |
| CN109934496A (en) | Interregional current influence determines method, apparatus, equipment and medium | |
| CN116300928A (en) | Data processing method for vehicle and training method for data processing model | |
| US20230280457A1 (en) | Radar detector with velocity profiling | |
| CN115796590A (en) | Method and device for predicting road fog accident risk, electronic equipment and medium | |
| Galanis et al. | Weather-based road condition estimation in the era of Internet-of-Vehicles (IoV) | |
| Xuan et al. | Performance Comparison of Machine-learning Models for Analyzing Weather and Traffic Accident Correlations | |
| CN114399788B (en) | Object detection method and system | |
| CN116125996B (en) | Safety monitoring method and system for unmanned vehicle | |
| CN117272840B (en) | Expressway bad weather early warning method and system | |
| CN115662167B (en) | Automatic driving map construction method, automatic driving method and related devices | |
| US20230243952A1 (en) | Unified radar perception architecture |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |