CN109785595A - A kind of vehicle abnormality track real-time identification method based on machine learning - Google Patents
A kind of vehicle abnormality track real-time identification method based on machine learning Download PDFInfo
- Publication number
- CN109785595A CN109785595A CN201910141745.2A CN201910141745A CN109785595A CN 109785595 A CN109785595 A CN 109785595A CN 201910141745 A CN201910141745 A CN 201910141745A CN 109785595 A CN109785595 A CN 109785595A
- Authority
- CN
- China
- Prior art keywords
- model
- data
- training
- vehicle
- abnormality
- 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.)
- Pending
Links
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a kind of vehicle abnormality track real-time identification method based on machine learning, is related to track of vehicle anomalous identification field；It obtains comprising: carry out cleaning to the data of acquisition completely, without the training data for repeating, being no different constant value；Model training is carried out using based on unsupervised isolated forest method and training data, obtains abnormality detection model；Abnormality detection model is put into stream calculation engine and is predicted in real time, and prediction result is sent to car owner；It is automatically updated according to the feedback information of car owner and correction model, and updated model is put into and is predicted in real time in stream calculation engine and prediction result is sent to car owner；The present invention is by periodically acquiring information of vehicles, use unsupervised isolated forest algorithm, real-time forecast analysis is carried out to track of vehicle in stream calculation engine, provide the probability value of vehicle abnormality behavior, the feedback data provided according to vehicle user, periodically adjustment model realizes dynamic more new model, improves the recognition accuracy of model.
Description
Technical field
The present invention relates to track of vehicle anomalous identification field, especially a kind of vehicle abnormality track based on machine learning is real
When recognition methods.
Background technique
With Chinese society's expanding economy, the ownership of the vehicle (automobile, electric vehicle, motorcycle) in Chinese market is got over
Come it is bigger, vehicle lose problem also become a more intractable social security problem of management.The theft one of car owner at present
As trace historical track and monitoring using post-event alarm, afterwards method, however many theft gangs have it is huger and mature
Network of disposing of stolen goods, speed of disposing of stolen goods is especially fast, once vehicle is lost, the probability recovered of alarm rear vehicle is relatively low, while being chased after
The human cost and social resources that stolen goods are spent are relatively high.
The existing track of vehicle anomalous identification technical functionality point for taking machine learning model all focuses on machine learning algorithm
Optimization, for example the data cleansings of track of vehicle data, dwell point are extracted, interest region is excavated, visualization interest region, movement
Mode excavation etc..But its have the disadvantage in that one, trained model using customize training, training after use when without
Method is updated according to feedback data, can not be modified to model, model gradually loses accuracy；Two, vehicle massive information
Under, the handling capacity and performance of system are unable to satisfy.Therefore, it is necessary to a kind of vehicle abnormality track recognizing methods can overcome above ask
Topic.
Summary of the invention
It is an object of the invention to: the present invention provides a kind of vehicle abnormality track based on the machine learning sides of identification in real time
Method, solving the problems, such as that existing vehicle abnormality track identification model can not update causes recognition accuracy low.
The technical solution adopted by the invention is as follows:
A kind of vehicle abnormality track real-time identification method based on machine learning, includes the following steps:
Cleaning is carried out to the data of acquisition to obtain completely, without the training data for repeating, being no different constant value；
Model training is carried out using based on unsupervised isolated forest method and training data, obtains abnormality detection model；
Abnormality detection model is put into stream calculation engine and is predicted in real time, and prediction result is sent to car owner；
It is automatically updated according to the feedback information of car owner and correction model, and updated model is put into stream calculation engine
It is predicted in real time and prediction result is sent to car owner.
Preferably, the data of the acquisition include the longitude data of vehicle, the latitude data of vehicle and Unix time；It is described
The preservation system of data uses Hadoop distributed file system, that is, HDFS；The cleaning means of the data uses Hive.
Preferably, the model training includes the following steps:
Step a: rejecting outliers model training is carried out using training data, obtains abnormality detection model；
Step b: the assessment of model accuracy rate is carried out to abnormality detection model, judges whether its model accuracy rate is up to standard, if reaching
Mark then obtains abnormality detection model PMML file by processing；If below standard, skip to step a and continue to train.
Preferably, described put completion training pattern into carries out predicting in real time in stream calculation engine including the following steps:
Step aa: according to the time period or warning strategies and alarm level is arranged in continuous trigger number, obtains warning strategies group；
Step bb: predicting output probability value after vehicle data using abnormality detection model in real time, by probability value and alarm plan
Slightly group matching obtains warning information；
Step cc: being sent to car owner for warning information, sending method include wechat touching reach, short message and multimedia message.
Preferably, the feedback information according to car owner is automatically updated includes the following steps: with correction model
Step aaa: historical data and latest data based on vehicle carry out regular exercise to rejecting outliers model automatically
Obtain new model；
Step bbb: whether the accuracy rate of automatic detection new model is up to standard, if up to standard, skips to step ccc more new model,
If it does not meet the standards, then not more new model；
Step ccc: the model up to standard to detection is replaced deployment automatically.
Preferably, regular exercise is carried out to rejecting outliers model automatically in the step aaa to include the following steps:
The training dataset of modeling, the training dataset include all historical training datasets on HDFS and most
The data newly obtained；
Timing training script is set；
Training dataset is put into model according to timing training script timing and is trained, completion timing training；Wherein
Historical training dataset of this training dataset as next time.
In conclusion by adopting the above-described technical solution, the beneficial effects of the present invention are:
1. the present invention is in such a way that high amount of traffic calculates and machine learning combines, by the warp for periodically acquiring vehicle
The data informations such as latitude, time, using Outlier Detection Algorithms such as unsupervised isolated forests, to vehicle rail in stream calculation engine
Mark carries out real-time forecast analysis, provides the probability value of vehicle abnormality behavior, the feedback data provided according to vehicle user, periodically
Model is adjusted, dynamic more new model is realized, solves existing vehicle abnormality track identification model and can not update to lead to recognition accuracy
Low problem periodically automatically updates machine learning prediction model, improves the recognition accuracy of model；
2. the present invention is based on the probability value of machine learning model real-time detection track of vehicle exception, according to the alarm plan of setting
A possibility that slightly judging theft, can look-ahead theft；
3. the present invention is based on the rejecting outliers methods such as unsupervised isolated forest (Isolation Forest) to carry out mould
Type training, isolated forest algorithm directly find out abnormal data, do not have to construct model to normal data, compared to traditional exception
The Stability and veracity of data detection method, model is higher, and detection efficiency is higher in mass data；
4. the present invention is strong using this big data framework extensibility of streaming engine, handling capacity is big, and processing data are high in real time
Effect, uninterruptedly.
Detailed description of the invention
In order to illustrate the technical solution of the embodiments of the present invention more clearly, below will be to needed in the embodiment attached
Figure is briefly described, it should be understood that the following drawings illustrates only certain embodiments of the present invention, therefore is not construed as pair
The restriction of range for those of ordinary skill in the art without creative efforts, can also be according to this
A little attached drawings obtain other relevant attached drawings.
Fig. 1 is method block diagram of the invention；
Fig. 2 is flow chart of the method for the present invention；
Fig. 3 is specific implementation effect diagram of the invention；
Fig. 4 is model modification flow chart of the invention；
Specific embodiment
In order to make the objectives, technical solutions, and advantages of the present invention clearer, with reference to the accompanying drawings and embodiments, right
The present invention is further elaborated.It should be appreciated that described herein, specific examples are only used to explain the present invention, not
For limiting the present invention, i.e., described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is logical
The component for the embodiment of the present invention being often described and illustrated herein in the accompanying drawings can be arranged and be designed with a variety of different configurations.
Therefore, the detailed description of the embodiment of the present invention provided in the accompanying drawings is not intended to limit below claimed
The scope of the present invention, but be merely representative of selected embodiment of the invention.Based on the embodiment of the present invention, those skilled in the art
Member's every other embodiment obtained without making creative work, shall fall within the protection scope of the present invention.
It should be noted that the relational terms of term " first " and " second " or the like be used merely to an entity or
Operation is distinguished with another entity or operation, and without necessarily requiring or implying between these entities or operation, there are any
This actual relationship or sequence.Moreover, the terms "include", "comprise" or its any other variant be intended to it is non-exclusive
Property include so that include a series of elements process, method, article or equipment not only include those elements, but also
Further include other elements that are not explicitly listed, or further include for this process, method, article or equipment it is intrinsic
Element.In the absence of more restrictions, the element limited by sentence "including a ...", it is not excluded that including described
There is also other identical elements in the process, method, article or equipment of element.
Technical problem: solving the problems, such as that existing vehicle abnormality track identification model can not update causes recognition accuracy low；
Technological means:
A kind of vehicle abnormality track real-time identification method based on machine learning, includes the following steps:
Cleaning is carried out to the data of acquisition to obtain completely, without the training data for repeating, being no different constant value；
Model training is carried out using based on unsupervised isolated forest method and training data, obtains abnormality detection model；
Abnormality detection model is put into stream calculation engine and is predicted in real time, and prediction result is sent to car owner；
It is automatically updated according to the feedback information of car owner and correction model, and updated model is put into stream calculation engine
It is predicted in real time and prediction result is sent to car owner.
The data of acquisition include the longitude data of vehicle, the latitude data of vehicle and Unix time；The preservation of the data
System uses Hadoop distributed file system, that is, HDFS；The cleaning means of the data uses Hive.
Model training includes the following steps:
Step a: rejecting outliers model training is carried out using training data, obtains abnormality detection model；
Step b: the assessment of model accuracy rate is carried out to abnormality detection model, judges whether its model accuracy rate is up to standard, if reaching
Mark then obtains abnormality detection model PMML file by processing；If below standard, skip to step a and continue to train.
Completion training pattern is put into and carries out predicting in real time in stream calculation engine including the following steps:
Step aa: according to the time period or warning strategies and alarm level is arranged in continuous trigger number, obtains warning strategies group；
Step bb: predicting output probability value after vehicle data using abnormality detection model in real time, by probability value and alarm plan
Slightly group matching obtains warning information；
Step cc: being sent to car owner for warning information, sending method include wechat touching reach, short message and multimedia message.
Preferably, the feedback information according to car owner is automatically updated includes the following steps: with correction model
Step aaa: historical data and latest data based on vehicle carry out regular exercise to rejecting outliers model automatically
Obtain new model；
Step bbb: whether the accuracy rate of automatic detection new model is up to standard, if up to standard, skips to step ccc more new model,
If it does not meet the standards, then not more new model；
Step ccc: the model up to standard to detection is replaced deployment automatically.
Regular exercise is carried out to rejecting outliers model automatically in step aaa to include the following steps:
The training dataset of modeling, the training dataset include all historical training datasets on HDFS and most
The data newly obtained；
Timing training script is set；
Training dataset is put into model according to timing training script timing and is trained, completion timing training；Wherein
Historical training dataset of this training dataset as next time.
Technical effect: the present invention passes through periodical acquisition in such a way that high amount of traffic calculates and machine learning combines
The data informations such as longitude and latitude, the time of vehicle, using Outlier Detection Algorithms such as unsupervised isolated forests, in stream calculation engine
Real-time forecast analysis is carried out to track of vehicle, provides the probability value of vehicle abnormality behavior, the feedback coefficient provided according to vehicle user
According to periodically adjusting model, realize dynamic more new model, solve existing vehicle abnormality track identification model and can not update to cause to know
The low problem of other accuracy, periodically automatically updates machine learning prediction model, improves the recognition accuracy of model；The present invention
Based on the probability value of machine learning model real-time detection track of vehicle exception, theft is judged according to the warning strategies of setting
Possibility, can look-ahead theft；Based on rejecting outliers such as unsupervised isolated forests (Isolation Forest)
Method carries out model training, and isolated forest algorithm directly finds out abnormal data, does not have to construct model to normal data, compared to
The Stability and veracity of traditional abnormal deviation data examination method, model is higher, and detection efficiency is higher in mass data；It uses
This big data framework extensibility of streaming engine is strong, and handling capacity is big, processing data real-time high-efficiency, uninterrupted.
Feature and performance of the invention are described in further detail with reference to embodiments.
Embodiment 1
Technical term is explained:
PMML, full name Predictive Model Markup Language (Predictive Model Markup Language), is retouched using XML
State with storing data mining model, PMML is a kind of de facto standard language, for rendering data mining model.
Hadoop is a distributed system infrastructure developed by apache foundation.
HDFS, Hadoop realize a distributed file system (Hadoop Distributed File System),
Abbreviation HDFS.HDFS has the characteristics of high fault tolerance, and is designed to be deployed on cheap (low-cost) hardware；And it
High-throughput (high throughput) data for carrying out access application are provided, those is suitble to have super large data set
The application program of (large data set).
The data file of structuring can be mapped as a number by Hive, a Tool for Data Warehouse based on Hadoop
According to library table, and simple sql query function is provided, sql sentence can be converted to MapReduce task and run.
Spark, Apache Spark are the computing engines for the Universal-purpose quick for aiming at large-scale data processing and designing, including
Four kinds of functions such as SQL, ML, graphx, streaming.
Flink, novel streaming computing frame.
Kafka, the open source stream process platform developed by Apache Software Foundation, is a kind of distribution of high-throughput
Formula distribution subscription message system, official website: http://kafka.apache.org/.
A kind of vehicle abnormality track real-time identification method based on machine learning, includes the following steps:
Cleaning is carried out to the data of acquisition to obtain completely, without the training data for repeating, being no different constant value；
Model training is carried out using based on unsupervised isolated forest method and training data, obtains abnormality detection model；
Abnormality detection model is put into stream calculation engine and is predicted in real time, and prediction result is sent to car owner；
It is automatically updated according to the feedback information of car owner and correction model, and updated model is put into stream calculation engine
It is predicted in real time and prediction result is sent to car owner.
It is specific as follows:
Step 1: data acquisition and data cleansing:
In order to use machine learning algorithm to carry out model training to vehicle abnormality track, traditional Relational DataBase is used
It is not applicable that (Mysql, Oracle etc.) saves magnanimity information of vehicles, because relevant database is due to locking mechanisms, in data
Amount just needs a point library to divide table, poor expandability after reaching million ranks.Therefore the application is based on the processing of Hadoop big data frame
Mass data information saves data using HDFS, carries out data cleansing to data using Hive.
| Data name | Type | Range | Explanation |
| Longitude | double | 0~180 | The longitude data of vehicle |
| Latitude | double | 0~90 | The latitude information of vehicle |
| Time | big int | 1~2^64 | Unix time, i.e. number of seconds since on January 1st, 1970 |
Data cleansing mode:
Longitude information, latitude information, temporal information have any one missing then to remove this record；
Remove the record i.e. exceptional value not in data area；
Remove the record of data type inaccuracy；
Remove repeated data using distinct to record；
Step 2: model training:
Model training is carried out based on unsupervised isolated forest (Isolation Forest) method；Isolated forest algorithm is straight
It connects and finds out abnormal data, do not have to construct model to normal data, compared to traditional abnormal deviation data examination method, the standard of model
True property and stability are higher, and detection efficiency is higher in mass data.
Probability value more may be exceptional value closer to 1, more may be normal value closer to 0.When the probability of most of data
When value is 0.5, then it represents that data are no different constant value.
Model training should meet the following conditions:
The data volume of training data is sufficiently large, time cycle long enough；
Training data has carried out data cleansing, without missing data, without abnormal data, and data duplicate removal；
The accuracy rate and recall rate of model training should be higher than desired threshold value (such as 90%)；
Model training tool uses big data Machine learning tools spark-ML or Flink-ML；
It saves after model training success into PMML file, is put into stream calculation engine and carries out predicting abnormality.PMML filename
It is set using " business name-version number-renewal time ", the purpose for saving as PMML file is to save into general model lattice
Formula, it is convenient to be loaded, predicted in stream calculation engine (flink, storm, spark streaming).
Step 3: predicting abnormality:
Trained model is put into stream calculation engine and is predicted in real time, output abnormality probability value (0~1).It is abnormal
Probability value is higher, illustrates that a possibility that vehicle is currently in abnormal track is higher.
The warning strategies of diversified forms are set:
Outlier threshold: it is set as 0.85 herein；Frequency of abnormity: it is set as 5 herein.
Warning strategies one: not distinguishing the period, as long as continuous five times are all higher than predicting abnormality threshold value, is alerted；
Warning strategies two: the period is distinguished, such as in midnight (morning 0 ---) at 5 points, as long as being all higher than for continuous five times abnormal
Prediction threshold value is just alerted；
Customized warning strategies: meet other combination conditions and alerted.
The probability value matching warning strategies of output are obtained into warning information, warning information (including vehicle violation position) is logical
It crosses the modes such as short message or wechat triggering and notifies vehicle user, and require whether user feedback is reported by mistake.If user replys wrong report,
It is then recorded as reporting by mistake, if user is back to normal, be recorded as normal；If user does not reply, it is considered as normal.
Step 4: model modification: being automatically updated according to the feedback information of car owner and correction model, and by updated model
It puts into and is predicted in real time in stream calculation engine and prediction result is sent to car owner.
When practical application, by judge vehicle in different time points (on and off duty, weekend, festivals or holidays etc.) motion profile come
Summarize the characteristics of motion of vehicle；By judging the frequency and the time point of vehicle shift track, in conjunction with the vehicle movement rule of summary
Carry out the probability of real-time judge vehicle loss；Car owner will receive warning information, can provide simply to alarm in information as user
It explains, such as: it is far that vehicle deviates regular course suddenly；Vehicle when abnormal between point (such as: the late into the night) appear in unconventional area
Domain；As shown in figure 3, blue solid lines are vehicle normal trace, red dotted line is vehicle abnormality track, can quickly identify track of vehicle
It is whether abnormal；In identification process the recognition accuracy of model can be improved according to the feedback more new model of warning information；The present invention adopts
The mode combined with machine learning is calculated with high amount of traffic, passes through the data such as longitude and latitude, the time for periodically acquiring vehicle and believes
Breath carries out prediction point in real time to track of vehicle in stream calculation engine using Outlier Detection Algorithms such as unsupervised isolated forests
Analysis, provides the probability value of vehicle abnormality behavior, the feedback data provided according to vehicle user, periodically adjusts model, realizes dynamic
State more new model, solving the problems, such as that existing vehicle abnormality track identification model can not update causes recognition accuracy low, periodically
Ground automatically updates machine learning prediction model, improves the recognition accuracy of model.
Embodiment 2
On the basis of the embodiment of the present invention one, refined model updates, as shown in figure 4, specific as follows:
Step 4: model modification:
It is automatically updated according to the feedback information of car owner and is included the following steps: with correction model
Step aaa: historical data and latest data based on vehicle carry out regular exercise to rejecting outliers model automatically
Obtain new model；
Step bbb: whether the accuracy rate of automatic detection new model is up to standard, if up to standard, skips to step ccc more new model,
If it does not meet the standards, then not more new model；
Step ccc: the model up to standard to detection is replaced deployment automatically.
Regular exercise is carried out to rejecting outliers model automatically in step aaa to include the following steps:
The training dataset of modeling, the training dataset include all historical training datasets on HDFS and most
The data newly obtained；
Timing training script is set；
Training dataset is put into model according to timing training script timing and is trained, completion timing training；Wherein
Historical training dataset of this training dataset as next time.
Model needs are periodically updated and correct, and could persistently guarantee the accuracy and validity of model, update cycle
Can be customized, the update cycle can be day, week, the moon, year.Update cycle is longer, and the system resource used is fewer, but model
Accuracy can be lower and lower；The period of model modification is shorter, and the system resource used is more, but the accuracy of model is protected
Card.
Model is set as training automatically, automatic detection, automatic deployment.Based on latest data, existing historical data and user
Feedback data periodically carries out model training, whether uses the automatic detection model of accuracy rate P (precision) index after model training
Meet it is up to standard, accuracy rate refer to model identification whole accuracy rate, that is, being predicted as has much ratios in abnormal vehicle be pre-
It surveys accurate.
Model Selection Strategy:
The model accuracy rate P of single training >=90%, model is up to standard；
Twice in succession trained model accuracy rate [80%, 90%), model is up to standard；
Model accuracy rate P < 80% of single training, model be not up to standard.
Specific threshold value can be adjusted according to the security level of business, and safety coefficient is demanding just by two threshold value tune
It is whole it is high a bit, safety coefficient requires low by the lower of two adjusting thresholds.
It is not updated if model is not up to standard, still uses old model, new model is used if model is up to standard.
When automatically updating, two prediction links are set in streaming computing engine, both links are mutually backups；When one it is pre-
When in use, another prediction link B is for updating by surveyor's chain road A；After model modification comes into force, link B is for predicting, link A
For updating, loop back and forth like this.
The feedback data provided by vehicle user periodically adjusts model, realizes dynamic more new model, avoids existing pre-
The disadvantage that error model causes model discrimination lower and lower is continued to use under measured data error situation, is periodically automatically updated
Machine learning prediction model, updates according to feedback data, improves the validity and recognition accuracy of model.
The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the invention, all in essence of the invention
Made any modifications, equivalent replacements, and improvements etc., should all be included in the protection scope of the present invention within mind and principle.
Claims (6)
1. a kind of vehicle abnormality track real-time identification method based on machine learning, characterized by the following steps:
Cleaning is carried out to the data of acquisition to obtain completely, without the training data for repeating, being no different constant value；
Model training is carried out using based on unsupervised isolated forest method and training data, obtains abnormality detection model；
Abnormality detection model is put into stream calculation engine and is predicted in real time, and prediction result is sent to car owner；
It is automatically updated according to the feedback information of car owner and correction model, and updated model is put into stream calculation engine and is carried out
It predicts in real time and prediction result is sent to car owner.
2. a kind of vehicle abnormality track real-time identification method based on machine learning according to claim 1, feature exist
In: the data of the acquisition include the longitude data of vehicle, the latitude data of vehicle and Unix time；The preservation system of the data
System uses Hadoop distributed file system, that is, HDFS；The cleaning means of the data uses Hive.
3. a kind of vehicle abnormality track real-time identification method based on machine learning according to claim 1, feature exist
In: the model training includes the following steps:
Step a: rejecting outliers model training is carried out using training data, obtains abnormality detection model；
Step b: carrying out the assessment of model accuracy rate to abnormality detection model, judge whether its model accuracy rate is up to standard, if up to standard,
Abnormality detection model PMML file is obtained by processing；If below standard, skip to step a and continue to train.
4. according to claim 1 or a kind of vehicle abnormality track real-time identification method based on machine learning described in 3, special
Sign is: described put completion training pattern into carries out predicting in real time in stream calculation engine including the following steps:
Step aa: according to the time period or warning strategies and alarm level is arranged in continuous trigger number, obtains warning strategies group；
Step bb: output probability value after vehicle data is predicted in real time using abnormality detection model, by probability value and warning strategies group
Matching obtains warning information；
Step cc: being sent to car owner for warning information, sending method include wechat touching reach, short message and multimedia message.
5. a kind of vehicle abnormality track real-time identification method based on machine learning according to claim 2, feature exist
In: the feedback information according to car owner is automatically updated to be included the following steps: with correction model
Step aaa: historical data and latest data based on vehicle carry out regular exercise to rejecting outliers model automatically and obtain
New model；
Step bbb: whether the accuracy rate of automatic detection new model is up to standard, if up to standard, step ccc more new model is skipped to, if not
It is up to standard, then not more new model；
Step ccc: the model up to standard to detection is replaced deployment automatically.
6. a kind of vehicle abnormality track real-time identification method based on machine learning according to claim 5, feature exist
In: regular exercise is carried out to rejecting outliers model automatically in the step aaa and is included the following steps:
The training dataset of modeling, the training dataset include all historical training datasets on HDFS and newest obtain
The data taken；
Timing training script is set；
Training dataset is put into model according to timing training script timing and is trained, completion timing training；Wherein this
Training dataset as next time historical training dataset.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910141745.2A CN109785595A (en) | 2019-02-26 | 2019-02-26 | A kind of vehicle abnormality track real-time identification method based on machine learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910141745.2A CN109785595A (en) | 2019-02-26 | 2019-02-26 | A kind of vehicle abnormality track real-time identification method based on machine learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109785595A true CN109785595A (en) | 2019-05-21 |
Family
ID=66487076
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910141745.2A Pending CN109785595A (en) | 2019-02-26 | 2019-02-26 | A kind of vehicle abnormality track real-time identification method based on machine learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109785595A (en) |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110203128A (en) * | 2019-06-18 | 2019-09-06 | 东风小康汽车有限公司重庆分公司 | Construction method, turn signal autocontrol method and the system of turn signal submodel |
| CN110599748A (en) * | 2019-09-20 | 2019-12-20 | 英华达（上海）科技有限公司 | Reminding method and reminding device applying same |
| CN111010393A (en) * | 2019-12-16 | 2020-04-14 | 陕西数图行信息科技有限公司 | Anomaly detection and elimination method for big data cleaning |
| CN111026925A (en) * | 2019-11-29 | 2020-04-17 | 中电福富信息科技有限公司 | Flink-based anomaly detection method and device for parallelization of isolated forest algorithm |
| CN111325463A (en) * | 2020-02-18 | 2020-06-23 | 深圳前海微众银行股份有限公司 | Data quality detection method, device, equipment and computer readable storage medium |
| CN111353890A (en) * | 2020-03-30 | 2020-06-30 | 中国工商银行股份有限公司 | Application log-based application anomaly detection method and device |
| CN111599165A (en) * | 2020-01-21 | 2020-08-28 | 南京中新赛克科技有限责任公司 | Multi-source big data-based electric vehicle robbery real-time warning method and system |
| CN111784966A (en) * | 2020-06-15 | 2020-10-16 | 武汉烽火众智数字技术有限责任公司 | Personnel management and control method and system based on machine learning |
| CN111860561A (en) * | 2019-12-31 | 2020-10-30 | 北京嘀嘀无限科技发展有限公司 | Method, device, equipment and storage medium for identifying abnormal stopping behavior of network appointment vehicle |
| CN111988342A (en) * | 2020-09-18 | 2020-11-24 | 大连理工大学 | Online automobile CAN network anomaly detection system |
| CN112959960A (en) * | 2021-02-26 | 2021-06-15 | 北京车和家信息技术有限公司 | Abnormal vehicle use detection method and device, medium, equipment, vehicle and server |
| CN113627521A (en) * | 2021-08-09 | 2021-11-09 | 西华大学 | Intelligent logistics unmanned aerial vehicle abnormal behavior identification method based on isolated forest method |
| CN113781797A (en) * | 2021-11-11 | 2021-12-10 | 成都古河云科技有限公司 | Automatic optical recognition system for vehicle outline |
| CN113792782A (en) * | 2021-09-13 | 2021-12-14 | 一汽出行科技有限公司 | Track monitoring method and device for operating vehicle, storage medium and computer equipment |
| CN115984992A (en) * | 2022-12-22 | 2023-04-18 | 云控智行(上海)汽车科技有限公司 | Method, device and equipment for detecting vehicle operation data |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106528865A (en) * | 2016-12-02 | 2017-03-22 | 航天科工智慧产业发展有限公司 | Quick and accurate cleaning method of traffic big data |
| CN107240264A (en) * | 2017-06-26 | 2017-10-10 | 安徽四创电子股份有限公司 | A kind of non-effective driving trace recognition methods of vehicle and urban road facility planing method |
| US10026506B1 (en) * | 2015-02-06 | 2018-07-17 | Brain Trust Innovations I, Llc | System, RFID chip, server and method for capturing vehicle data |
| CN108921440A (en) * | 2018-07-11 | 2018-11-30 | 平安科技（深圳）有限公司 | Pollutant method for monitoring abnormality, system, computer equipment and storage medium |
| CN109191354A (en) * | 2018-08-21 | 2019-01-11 | 安徽讯飞智能科技有限公司 | A kind of whole people society pipe task distribution method based on natural language processing |
| CN109214107A (en) * | 2018-09-26 | 2019-01-15 | 大连海事大学 | A kind of ship's navigation behavior on-line prediction method |
-
2019
- 2019-02-26 CN CN201910141745.2A patent/CN109785595A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10026506B1 (en) * | 2015-02-06 | 2018-07-17 | Brain Trust Innovations I, Llc | System, RFID chip, server and method for capturing vehicle data |
| CN106528865A (en) * | 2016-12-02 | 2017-03-22 | 航天科工智慧产业发展有限公司 | Quick and accurate cleaning method of traffic big data |
| CN107240264A (en) * | 2017-06-26 | 2017-10-10 | 安徽四创电子股份有限公司 | A kind of non-effective driving trace recognition methods of vehicle and urban road facility planing method |
| CN108921440A (en) * | 2018-07-11 | 2018-11-30 | 平安科技（深圳）有限公司 | Pollutant method for monitoring abnormality, system, computer equipment and storage medium |
| CN109191354A (en) * | 2018-08-21 | 2019-01-11 | 安徽讯飞智能科技有限公司 | A kind of whole people society pipe task distribution method based on natural language processing |
| CN109214107A (en) * | 2018-09-26 | 2019-01-15 | 大连海事大学 | A kind of ship's navigation behavior on-line prediction method |
Cited By (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110203128A (en) * | 2019-06-18 | 2019-09-06 | 东风小康汽车有限公司重庆分公司 | Construction method, turn signal autocontrol method and the system of turn signal submodel |
| CN110599748A (en) * | 2019-09-20 | 2019-12-20 | 英华达（上海）科技有限公司 | Reminding method and reminding device applying same |
| CN110599748B (en) * | 2019-09-20 | 2021-05-11 | 英华达（上海）科技有限公司 | Reminding method and reminding device applying same |
| CN111026925A (en) * | 2019-11-29 | 2020-04-17 | 中电福富信息科技有限公司 | Flink-based anomaly detection method and device for parallelization of isolated forest algorithm |
| CN111010393A (en) * | 2019-12-16 | 2020-04-14 | 陕西数图行信息科技有限公司 | Anomaly detection and elimination method for big data cleaning |
| CN111010393B (en) * | 2019-12-16 | 2021-11-05 | 陕西数图行信息科技有限公司 | Anomaly detection and elimination method for big data cleaning |
| CN111860561A (en) * | 2019-12-31 | 2020-10-30 | 北京嘀嘀无限科技发展有限公司 | Method, device, equipment and storage medium for identifying abnormal stopping behavior of network appointment vehicle |
| CN111860561B (en) * | 2019-12-31 | 2024-02-27 | 北京嘀嘀无限科技发展有限公司 | Abnormal stay behavior identification method, device and equipment of network appointment vehicle and storage medium |
| CN111599165A (en) * | 2020-01-21 | 2020-08-28 | 南京中新赛克科技有限责任公司 | Multi-source big data-based electric vehicle robbery real-time warning method and system |
| CN111325463A (en) * | 2020-02-18 | 2020-06-23 | 深圳前海微众银行股份有限公司 | Data quality detection method, device, equipment and computer readable storage medium |
| CN111353890A (en) * | 2020-03-30 | 2020-06-30 | 中国工商银行股份有限公司 | Application log-based application anomaly detection method and device |
| CN111784966A (en) * | 2020-06-15 | 2020-10-16 | 武汉烽火众智数字技术有限责任公司 | Personnel management and control method and system based on machine learning |
| CN111988342A (en) * | 2020-09-18 | 2020-11-24 | 大连理工大学 | Online automobile CAN network anomaly detection system |
| CN112959960A (en) * | 2021-02-26 | 2021-06-15 | 北京车和家信息技术有限公司 | Abnormal vehicle use detection method and device, medium, equipment, vehicle and server |
| CN113627521A (en) * | 2021-08-09 | 2021-11-09 | 西华大学 | Intelligent logistics unmanned aerial vehicle abnormal behavior identification method based on isolated forest method |
| CN113792782A (en) * | 2021-09-13 | 2021-12-14 | 一汽出行科技有限公司 | Track monitoring method and device for operating vehicle, storage medium and computer equipment |
| CN113781797A (en) * | 2021-11-11 | 2021-12-10 | 成都古河云科技有限公司 | Automatic optical recognition system for vehicle outline |
| CN113781797B (en) * | 2021-11-11 | 2022-02-18 | 成都古河云科技有限公司 | Automatic optical recognition system for vehicle outline |
| CN115984992A (en) * | 2022-12-22 | 2023-04-18 | 云控智行(上海)汽车科技有限公司 | Method, device and equipment for detecting vehicle operation data |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109785595A (en) | A kind of vehicle abnormality track real-time identification method based on machine learning | |
| CN111475804B (en) | Alarm prediction method and system | |
| CN105095048A (en) | Processing method for alarm correlation of monitoring system based on business rules | |
| CN113179173B (en) | Operation and maintenance monitoring system for expressway system | |
| CN105917625A (en) | Classification of detected network anomalies using additional data | |
| CN114465874B (en) | Fault prediction method, device, electronic equipment and storage medium | |
| CN103281341A (en) | Network event processing method and device | |
| CN110333983A (en) | Business monitoring and searching service monitoring method and device | |
| CN116863723B (en) | Use method of digital twin base | |
| US11775530B2 (en) | Method to improve global query performance in an edge network | |
| CN115514619B (en) | Alarm convergence method and system | |
| Yan et al. | SeTraStream: semantic-aware trajectory construction over streaming movement data | |
| Ding et al. | Massive heterogeneous sensor data management in the Internet of Things | |
| CN116051330A (en) | Intelligent campus security management system and method based on big data | |
| CN111581056B (en) | Software engineering database maintenance and early warning system based on artificial intelligence | |
| Tang et al. | Multidimensional analysis of atypical events in cyber-physical data | |
| Woodard et al. | A survey of data cleansing techniques for cyber-physical critical infrastructure systems | |
| CN109992621A (en) | Foods supervision information resources increment ETL system and method | |
| CN114268567A (en) | Abnormal terminal identification method, abnormal terminal analysis device, abnormal terminal analysis equipment and abnormal terminal storage medium | |
| Dogan et al. | Evaluation of network trust using provenance based on distributed local intelligence | |
| CN117671961A (en) | Urban road traffic flow state prediction method and system based on block chain | |
| Brown et al. | From rocks to pebbles: Smoothing spatiotemporal data streams in an overlay of sensors | |
| CN117540228A (en) | Fault processing method, device and storage medium | |
| CN115033464A (en) | Distributed disaster recovery drilling data early warning method and system based on artificial intelligence | |
| Hendawi et al. | The microsoft reactive framework meets the internet of moving things |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190521 |