CN115878998A - Vehicle lane change identification method - Google Patents
Vehicle lane change identification method Download PDFInfo
- Publication number
- CN115878998A CN115878998A CN202211559485.9A CN202211559485A CN115878998A CN 115878998 A CN115878998 A CN 115878998A CN 202211559485 A CN202211559485 A CN 202211559485A CN 115878998 A CN115878998 A CN 115878998A
- Authority
- CN
- China
- Prior art keywords
- data
- vehicle
- lane
- lane change
- nearest neighbor
- 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
- 238000000034 method Methods 0.000 title claims abstract description 36
- 238000007635 classification algorithm Methods 0.000 claims abstract description 30
- 238000010606 normalization Methods 0.000 claims abstract description 10
- 230000001133 acceleration Effects 0.000 claims abstract description 9
- 230000006399 behavior Effects 0.000 claims description 10
- 238000010801 machine learning Methods 0.000 description 5
- 230000006870 function Effects 0.000 description 4
- 238000013075 data extraction Methods 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 206010000117 Abnormal behaviour Diseases 0.000 description 1
- 206010039203 Road traffic accident Diseases 0.000 description 1
- 238000007418 data mining Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Landscapes
- Traffic Control Systems (AREA)
Abstract
The invention discloses a vehicle lane change identification method, which comprises the following steps: respectively extracting features of lane change data and lane invariant data of a vehicle to obtain corresponding feature data sets, wherein the feature data sets comprise variances and maximum deviation values; merging the feature data sets after normalization processing, training a nearest neighbor classification algorithm based on the merged data, and obtaining the trained nearest neighbor classification algorithm; and identifying the lane change behavior of the vehicle based on the trained nearest neighbor classification algorithm. The invention reduces the cost of the current common vehicle lane change identification method, can judge whether the vehicle changes the lane by only acquiring the acceleration data in the x direction and the y direction, and can assist a global satellite navigation system such as a Beidou satellite navigation system to more accurately detect whether the vehicle changes the lane.
Description
Technical Field
The invention belongs to the technical field of machine learning, and particularly relates to a vehicle lane change identification method.
Background
The lane change behavior is one of the basic vehicle driving behaviors, which greatly influences the normal operation of road traffic, at the present stage, the traffic accidents caused by frequent lane change and illegal lane change of drivers are increasingly increased, and the abnormal behavior of vehicles represented by unreasonable lane change is gradually the main reason of traffic jam in an intersection area of a highway and failure of main line traffic flow. Therefore, the identification of the lane change of the vehicle can help people to better know the specific situation of road traffic and improve the road traffic.
The current common vehicle lane change identification method is machine vision, and the current commercially used lane departure early warning systems are all based on machine vision and can be divided into an overlooking system and a forward-looking system according to the installation mode of a sensor. The overlooking system can only be used on a structured road, and a road mark must exist and can be effectively recognized; while forward looking systems can be used on roads without road markings, they are highly susceptible to interference from other vehicles or pedestrians. Meanwhile, machine vision is greatly influenced by environment and weather, under the working condition of insufficient light, the judgment error is large and even the machine vision is invalid, and if the factors such as road curvature and lane line blurring are considered and processed, the image processing workload is large and the real-time performance is insufficient. Accordingly, it is desirable to provide a method that is less costly and that accurately identifies a lane change in a vehicle.
Disclosure of Invention
The invention aims to provide a vehicle lane change identification method to solve the problems in the prior art.
In order to achieve the above object, the present invention provides a lane change recognition method for a vehicle, comprising the steps of:
respectively extracting features of lane change data and lane invariant data of a vehicle to obtain corresponding feature data sets, wherein the feature data sets comprise variances and maximum deviation values;
merging the feature data sets after normalization processing, training a nearest neighbor classification algorithm based on merged data, and obtaining the trained nearest neighbor classification algorithm;
and identifying the lane change behavior of the vehicle based on the trained nearest neighbor classification algorithm.
Optionally, the process of acquiring lane change data and lane invariant data of the vehicle includes: extracting a plurality of rows of data before and after the lane id changes from the HighD data set, and extracting x-direction acceleration data, y-direction acceleration data, lane id and vehicle id in the plurality of rows of data for storage; if the numbers of the lane id and the vehicle id are not changed, the extracted data are lane-invariant data; and if the numbers of the lane id and the vehicle id are changed, the extracted data are lane change data.
Optionally, the process of performing the normalization process includes: and respectively extracting the x variance, the y variance, the x maximum deviation value and the y maximum deviation value of the lane change data and the lane unchangeable data from the characteristic data set, and respectively carrying out normalization processing.
Optionally, the training of the nearest neighbor classification algorithm further comprises: and dividing the merged data into a training set and a test set according to a preset proportion.
Optionally, the process of training the nearest neighbor classification algorithm comprises: training a nearest neighbor classification algorithm based on the training set, obtaining the accuracy of the test set, obtaining the optimal k value of the nearest neighbor classification algorithm based on the accuracy, and further obtaining the trained nearest neighbor classification algorithm.
Optionally, the process of predicting the lane change behavior of the vehicle includes: acquiring a plurality of rows of data of application data, performing normalization processing, inputting the data into a trained nearest neighbor classification algorithm, and identifying lane change behaviors of each row of data based on a prediction function in the trained nearest neighbor classification algorithm; the application data is data to be predicted.
Optionally, the process of normalizing the application data includes: and performing feature extraction on a plurality of rows of data of the application data to obtain an x variance, a y variance, an x maximum deviation value and a y maximum deviation value, and performing normalization processing on the x variance, the y variance, the x maximum deviation value and the y maximum deviation value.
The invention has the technical effects that:
the invention reduces the cost of the current common vehicle lane change identification method, can judge whether the vehicle changes the lane by only acquiring the acceleration data in the x direction and the y direction, and the cost of the sensor for acquiring the vehicle acceleration is far less than the cost of the overlooking system and the foresight system required by the machine vision of the current common vehicle lane change identification method. Meanwhile, the invention can also assist a global satellite navigation system such as a Beidou satellite navigation system to more accurately detect whether the vehicle changes lanes.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this application, illustrate embodiments of the application and, together with the description, serve to explain the application and are not intended to limit the application. In the drawings:
fig. 1 is a flowchart of a lane change recognition method for a vehicle according to an embodiment of the present invention.
Detailed Description
It should be noted that the embodiments and features of the embodiments in the present application may be combined with each other without conflict. The present application will be described in detail below with reference to the embodiments with reference to the attached drawings.
It should be noted that the steps illustrated in the flowcharts of the figures may be performed in a computer system such as a set of computer-executable instructions and that, although a logical order is illustrated in the flowcharts, in some cases, the steps illustrated or described may be performed in an order different than presented herein.
Example one
As shown in fig. 1, in the present embodiment, a lane change recognition method for a vehicle is provided, in which a HighD data set is subjected to data extraction, feature values of the extracted data are normalized, a KNN (K-nearest neighbor) nearest neighbor classification algorithm is used for machine learning, and a prediction function of the KNN (K-nearest neighbor) nearest neighbor classification algorithm is used for predicting data with accelerations in x direction and y direction, so as to determine whether the vehicle changes lanes.
Extracting a data set
And extracting lane change data and lane change data in the HighD data set. Then, four columns of xAccelection, yAccelection, laneId, id in each of 5 lines before and after laneId digit change are selected for extraction. Finally, 10 lines of each lane change process are stored as a csv file and placed into each folder according to the sequence. For later machine learning.
And defining lane change in the HighD data set as the lane change process, wherein the lane change process is that the id number of the vehicle does not change, but the laneID number of the lane changes. And the id number of the vehicle and the laneId number of the lane are not changed, namely the lane is not changed. Extracting four columns of data of the variable channel and the invariable channel xAccele, yAccele, laneID and the like respectively, storing the four columns of data as csv files, and putting the csv files into all folders according to the sequence.
Normalizing the extracted data characteristic value
And extracting two characteristic values, namely variance and a maximum value deviation value, from the file of each extracted data set. Thus, four characteristic value data, i.e., x variance, y variance, x maximum value offset value, and y maximum value offset value, can be obtained. The four characteristic value data are normalized. After the normalized value is obtained, a 'bian' label is added beside the normalized value to represent lane change, so that the machine learning is facilitated later. These 5 data are stored in the csv file in sequence.
And similarly, extracting four characteristic value data of x variance, y variance, x maximum value deviation value and y maximum value deviation value from the data which does not change in the HighD data set, adding a label to the characteristic value data as 'bubian' and storing the characteristic value data in the csv file in sequence.
Machine learning by KNN nearest neighbor classification algorithm
And respectively reading the lane-changed data and the lane-unchanged data, and then combining the data. Then, according to the following steps of 3: the ratio of 1 splits this merged data into a training set and a test set. A KNN (K-nearest neighbor) nearest neighbor classification algorithm in data mining classification (classification) is employed for training. And obtaining the optimal k value required by the KNN proximity classification algorithm by calculating the accuracy of the test set. In this embodiment, a training set is used for training, and an optimal k value required by the KNN proximity classification algorithm is obtained as 6 by calculating the accuracy obtained by the test set.
Applied to raw data
The 18\ u tracks. Csv file in the HighD dataset was selected as the object for the application of the present invention. Of course, 18 v tracks. Csv was not extracted in the previous data extraction.
After the acceleration data in the x-direction and the y-direction in 18 v tracks.csv was collected, they were normalized in groups of 10 in order. And then forming a new file named 18.Csv by the normalized data according to the sequence.
Reading 18.Csv line by line, and judging whether each line of data represents lane change behavior through a prediction function of a trained KNN (K-nearest neighbor) nearest neighbor classification algorithm.
The final result was 79 lane change behaviors in 18 \ u tracks.csv, as well as 79 for the prediction function. The predicted lane change occurrence time also corresponds to the raw data.
The above description is only for the preferred embodiment of the present application, but the scope of the present application is not limited thereto, and any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope of the present application should be covered within the scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.
Claims (7)
1. A vehicle lane change identification method is characterized by comprising the following steps:
respectively extracting features of lane change data and lane invariant data of a vehicle to obtain corresponding feature data sets, wherein the feature data sets comprise variances and maximum deviation values;
merging the feature data sets after normalization processing, training a nearest neighbor classification algorithm based on the merged data, and obtaining the trained nearest neighbor classification algorithm;
and identifying the lane change behavior of the vehicle based on the trained nearest neighbor classification algorithm.
2. The vehicle lane change recognition method according to claim 1,
the process of acquiring lane change data and lane invariant data of a vehicle comprises the following steps: extracting a plurality of rows of data before and after the lane id changes from the HighD data set, and extracting x-direction acceleration data, y-direction acceleration data, lane id and vehicle id in the plurality of rows of data for storage; if the numbers of the lane id and the vehicle id are not changed, the extracted data are unchanged lane data; and if the numbers of the lane id and the vehicle id are changed, the extracted data are lane change data.
3. The vehicle lane-change recognition method according to claim 1,
the process of performing normalization processing includes: and respectively extracting the x variance, the y variance, the x maximum deviation value and the y maximum deviation value of the lane change data and the lane unchangeable data from the characteristic data set, and respectively carrying out normalization processing.
4. The vehicle lane change recognition method according to claim 1,
before training the nearest neighbor classification algorithm, the method further comprises the following steps: and dividing the merged data into a training set and a test set according to a preset proportion.
5. The vehicle lane change recognition method according to claim 4,
the process of training the nearest neighbor classification algorithm includes: training a nearest neighbor classification algorithm based on the training set, obtaining the accuracy of the test set, obtaining the optimal k value of the nearest neighbor classification algorithm based on the accuracy, and further obtaining the trained nearest neighbor classification algorithm.
6. The vehicle lane-change recognition method according to claim 1,
the process of predicting the lane change behavior of the vehicle comprises the following steps: acquiring a plurality of rows of data of application data, performing normalization processing, inputting the data into a trained nearest neighbor classification algorithm, and identifying lane change behaviors of each row of data based on a prediction function in the trained nearest neighbor classification algorithm; the application data is data to be predicted.
7. The vehicle lane change recognition method according to claim 6,
the process of normalizing the application data comprises the following steps: the method comprises the steps of extracting characteristics of a plurality of rows of data of application data, obtaining an x variance, a y variance, an x maximum deviation value and a y maximum deviation value of the application data, and normalizing the x variance, the y variance, the x maximum deviation value and the y maximum deviation value.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211559485.9A CN115878998A (en) | 2022-12-06 | 2022-12-06 | Vehicle lane change identification method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211559485.9A CN115878998A (en) | 2022-12-06 | 2022-12-06 | Vehicle lane change identification method |
Publications (1)
Publication Number | Publication Date |
---|---|
CN115878998A true CN115878998A (en) | 2023-03-31 |
Family
ID=85766230
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202211559485.9A Pending CN115878998A (en) | 2022-12-06 | 2022-12-06 | Vehicle lane change identification method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115878998A (en) |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105320966A (en) * | 2015-10-30 | 2016-02-10 | 东软集团股份有限公司 | Vehicle driving state recognition method and apparatus |
CN110329271A (en) * | 2019-06-18 | 2019-10-15 | 北京航空航天大学杭州创新研究院 | A kind of multisensor vehicle driving detection system and method based on machine learning |
CN111209838A (en) * | 2019-12-31 | 2020-05-29 | 清华大学 | Driving intention-based dynamic identification method for lane change behavior of surrounding vehicles |
CN112560782A (en) * | 2020-12-26 | 2021-03-26 | 浙江天行健智能科技有限公司 | Vehicle lane changing behavior identification method based on random forest algorithm |
CN114132340A (en) * | 2021-08-16 | 2022-03-04 | 中南大学 | Lane change trajectory prediction method and device and computer storage medium |
CN115018016A (en) * | 2022-08-03 | 2022-09-06 | 苏州大学 | Method and system for identifying lane changing intention of manually-driven vehicle |
-
2022
- 2022-12-06 CN CN202211559485.9A patent/CN115878998A/en active Pending
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105320966A (en) * | 2015-10-30 | 2016-02-10 | 东软集团股份有限公司 | Vehicle driving state recognition method and apparatus |
CN110329271A (en) * | 2019-06-18 | 2019-10-15 | 北京航空航天大学杭州创新研究院 | A kind of multisensor vehicle driving detection system and method based on machine learning |
CN111209838A (en) * | 2019-12-31 | 2020-05-29 | 清华大学 | Driving intention-based dynamic identification method for lane change behavior of surrounding vehicles |
CN112560782A (en) * | 2020-12-26 | 2021-03-26 | 浙江天行健智能科技有限公司 | Vehicle lane changing behavior identification method based on random forest algorithm |
CN114132340A (en) * | 2021-08-16 | 2022-03-04 | 中南大学 | Lane change trajectory prediction method and device and computer storage medium |
CN115018016A (en) * | 2022-08-03 | 2022-09-06 | 苏州大学 | Method and system for identifying lane changing intention of manually-driven vehicle |
Non-Patent Citations (2)
Title |
---|
HAILUN ZHANG 等: "An Ensemble Learning–Online Semi-Supervised Approach for Vehicle Behavior Recognition", 《IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS》, vol. 23, no. 8, 31 August 2022 (2022-08-31), pages 10610 - 10626 * |
郑志晓 等: "基于车辆运行状态的高速公路驾驶员换道行为辨识", 《2015 中国汽车安全技术学术会议》, 31 December 2015 (2015-12-31), pages 55 - 64 * |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US9436879B2 (en) | Method for recognizing traffic signs | |
US9360332B2 (en) | Method for determining a course of a traffic lane for a vehicle | |
CN111016921A (en) | System and method for determining familiarity of a vehicle data set | |
CN110879950A (en) | Multi-stage target classification and traffic sign detection method and device, equipment and medium | |
KR20200024934A (en) | Lane information management method, driving control method and lane information management device | |
CN110175620B (en) | Detection based on fusion of multiple sensors | |
CN110335467B (en) | Method for realizing highway vehicle behavior detection by using computer vision | |
CN113127466B (en) | Vehicle track data preprocessing method and computer storage medium | |
CN111038522A (en) | System and method for assessing training data set familiarity of driver assistance systems | |
KR102197449B1 (en) | Enforcement system for enforcement a certain section in the section enforcement point | |
CN110675637A (en) | Vehicle illegal video processing method and device, computer equipment and storage medium | |
JP2008165326A (en) | Feature recognition device, own vehicle location recognizing device, navigation device and feature recognition method | |
CN111971724A (en) | Action selection device, action selection program, and action selection method | |
CN111284501A (en) | Apparatus and method for managing driving model based on object recognition, and vehicle driving control apparatus using the same | |
CN110386088B (en) | System and method for performing vehicle variance analysis | |
CN111352414A (en) | Decoy removal apparatus and method for vehicle and vehicle including the same | |
CN115878998A (en) | Vehicle lane change identification method | |
JP5327241B2 (en) | Object identification device | |
CN112559968A (en) | Driving style representation learning method based on multi-situation data | |
EP3786845A1 (en) | Difficulty-adaptive training for machine learning modules | |
WO2018030103A1 (en) | Displayed content recognition device and vehicle control device | |
US20210180982A1 (en) | Map generation device and map generation method | |
CN110741379A (en) | Method for determining the type of road on which a vehicle is travelling | |
CN114822044A (en) | Driving safety early warning method and device based on tunnel | |
CN114333414A (en) | Parking yield detection device, parking yield detection system, and recording medium |
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 |