WO2011060725A1

WO2011060725A1 - Abnormal vehicle speed data identifying method and device thereof

Info

Publication number: WO2011060725A1
Application number: PCT/CN2010/078875
Authority: WO
Inventors: 昝艳; 付新刚; 贾学力; 李建军
Original assignee: 北京世纪高通科技有限公司
Priority date: 2009-11-19
Filing date: 2010-11-18
Publication date: 2011-05-26
Also published as: CN101727749A; CN101727749B

Abstract

An abnormal vehicle speed data identifying method and device thereof are provided. The method includes: computing the distance of each vehicle speed data object in a target vehicle speed data set according to a preset parameter k; a radius parameter of the target set is selected from the distance; the vehicle object corresponding to the distance larger than the radius parameter is marked as an abnormal vehicle speed data.

Description

Method and device for identifying abnormal vehicle speed data This application claims to be submitted to the Chinese Patent Office on January 19, 2010, and the application number is 2 009 1 02 3794 0. K. The invention is entitled "Identification method and device for abnormal vehicle speed data". Priority of the Chinese Patent Application, the entire contents of which is incorporated herein by reference. Technical field

The present invention relates to the field of intelligent transportation, and in particular to a method and apparatus for identifying abnormal vehicle speed data.

Background technique

The intelligent transportation system is currently the most effective way to solve the transportation problems, especially traffic congestion, traffic congestion, traffic accidents and traffic pollution. Among them, dynamic traffic information service is one of the core research directions of intelligent transportation system. It can dynamically reflect the traffic conditions in the area in real time, guide the optimal driving route, improve the efficiency of road and vehicle use, and is important to alleviate traffic congestion. Measures. In the study of dynamic traffic information, it is a hot issue to analyze the road historical speed value and find the parameters that can reflect the cyclical trend of the road, and then extract the dynamic traffic characteristics of the road. However, due to various interferences in the process of collecting and processing the speed data, the vehicle speed data will be polluted by different degrees of noise. Using contaminated data modeling and statistical analysis may lead to erroneous results, so the vehicle speed data is required. Denoising, that is, identifying and deleting abnormal vehicle speed values. The main methods used in the prior art for identifying anomalous data are statistical methods, distance-based methods, and density-based clustering-based methods.

However, in the process of using statistical methods and distance-based methods for identifying abnormal vehicle speed data, the inventors have found that at least the following problems exist in the prior art: When using statistical methods to identify abnormal vehicle speed data, it is generally required to know knowledge about data set parameters. , such as distribution models, distribution parameters, etc. But in most cases, the distribution may be unknown. When using the distance-based method to identify abnormal vehicle speed data, the distance between data objects needs to be calculated according to a distance function, which is a data object with a higher distance than all other objects. But it requires the user to provide the minimum acceptable distance directly. And this is hard to determine. The density clustering based method can identify abnormal data from the data of unknown distribution form, but the density clustering based method has not been used in the prior art to identify abnormal vehicle speed data.

Summary of the invention

Embodiments of the present invention provide a method and apparatus for identifying abnormal vehicle speed data, which employs a density clustering based method to identify abnormal vehicle speed data from a target vehicle speed data set.

In order to achieve the above object, the embodiment of the present invention adopts the following technical solutions:

A method for identifying abnormal vehicle speed data, comprising:

Calculating the distance of each vehicle speed data object in the target vehicle speed data set according to the preset parameters;

A radius parameter of the target vehicle speed data set is selected from all of the above-mentioned distances; and a vehicle speed data object corresponding to the distance of the radius parameter is identified as abnormal vehicle speed data.

An apparatus for identifying abnormal vehicle speed data, comprising: a calculating unit, configured to calculate a distance of each vehicle speed data object in the target vehicle speed data set according to a preset parameter;

a selection unit, configured to select a radius parameter of the target vehicle speed data set from all of the above-distances;

And a first identification unit, configured to identify a vehicle speed data object corresponding to the distance greater than the radius parameter as abnormal vehicle speed data.

In the embodiment of the present invention described by the foregoing solution, the preset parameter is a density threshold in a density clustering based method, and the fc of each vehicle speed data object in the target vehicle speed data set may be calculated according to the preset parameter. -distance. The distance of the vehicle speed data object is the maximum distance of the vehicle speed data object to its nearest vehicle speed data object. After calculating the distance of all the vehicle speed data objects in the target data set, the radius parameters of the target vehicle speed data set are selected from all of the above-distances. The radius parameter of the target vehicle speed data set is the radius of the density clustering in the density clustering based method. By sequentially comparing the distance of each vehicle speed data object in the target data set with the radius parameter of the target vehicle speed data set Small, the vehicle speed data object corresponding to the distance corresponding to the radius parameter is identified as abnormal vehicle speed data. Thereby, a method based on density clustering is used to identify abnormal vehicle speed data.

DRAWINGS

1 is a flow chart showing a method for identifying abnormal vehicle speed data in Embodiment 1;

2 is a structural block diagram of an apparatus for identifying abnormal vehicle speed data of Embodiment 1;

3 is a flow chart showing a method for identifying abnormal vehicle speed data in Embodiment 2;

Fig. 4 is a block diagram showing the configuration of an apparatus for identifying abnormal vehicle speed data of the second embodiment.

detailed description

The technical solutions in the embodiments of the present invention are clearly and completely described in conjunction with the drawings in the embodiments of the present invention. It is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative efforts are within the scope of the present invention.

Example 1

An embodiment of the present invention provides a method for identifying abnormal vehicle speed data. As shown in FIG. 1, the method includes the following steps:

1 01. Calculate the distance of each vehicle speed data object in the target vehicle speed data set according to the preset parameters.

Since the density clustering based method identifies the anomaly data from the cluster of arbitrary shapes by setting the density threshold and the domain radius, when the density clustering based method is used to identify the abnormal vehicle speed data, the density threshold and the domain radius are also set. value. The preset parameter is a density threshold in a method based on density clustering, and the parameter can be set by the experience of a technician.

The distance of the vehicle speed data object may be defined as: a vehicle speed data object P and a vehicle speed data object. The distance between e D is di _S t P, o, and satisfies:

1 at least one vehicle speed data object object ^e , such that dist(p, q) < dist(p, o);

2 and at most there is a vehicle speed data object object, such that dist(p, q) < dist(p, o) 102. Select, from all the above-distances, a radius parameter of the target vehicle speed data set. Each vehicle speed data object in the target data set has a corresponding-distance, and one of all the above-distances is selected as a radius parameter of the target vehicle speed data set. The radius parameter is the value of the domain radius in the density clustering based method.

103. After determining the values of the density threshold and the domain radius, according to the density clustering based method, the vehicle speed data object corresponding to the distance-distance parameter is identified as the abnormal vehicle speed data. Thereby, the method based on density clustering is applied to identify abnormal vehicle speed data.

In order to implement the above method, an embodiment of the present invention further provides an identification device for abnormal vehicle speed data. As shown in FIG. 2, the device includes: a calculation unit 21, a selection unit 22, and a first identification unit 23.

The calculating unit 21 is configured to calculate an fc-distance of each vehicle speed data object in the target vehicle speed data set according to the preset parameter. After the calculating unit 21 calculates the distance of each vehicle speed data object, the selecting unit 22 is configured to select the radius parameter of the target vehicle speed data set from all of the above-mentioned distances. The radius parameter is the value of the radius of the field in the method based on density clustering. After the density threshold and the domain radius are determined in the density clustering based method, the first identification unit 23 is configured to identify the vehicle speed data object corresponding to the distance of the radius parameter as the abnormal vehicle speed data. The apparatus determines a density threshold and a domain radius in the density clustering based method by the calculating unit 21 and the selecting unit 22, and the first identifying unit 23 compares each vehicle speed data in the target data set according to the density clustering based method. The object-to-distance and the size of the radius parameter of the target vehicle speed data set identify abnormal vehicle speed data from the target data set. Thereby, a method based on density clustering is used to identify abnormal vehicle speed data.

Example 2:

In the embodiment of the present invention, a method for identifying abnormal vehicle speed data is described in detail by taking a historical traffic flow on a designated road link as an example. As shown in FIG. 3, the method includes the following steps:

301. Load the target data set. The target data set may be a vehicle speed value at a specific time point of a week characteristic day on a specified road link. For example, when the target data set is taken from the vehicle speed data object of one year, there may be a loss of the vehicle speed value, and the target data set has a maximum of 52 vehicles. Speed value. The week feature day refers to a set of certain flood seasons in which the traffic flow data changes in the week have certain similar characteristics. For example, the date of the week can be divided into two types of weekdays, weekdays and Sundays. It can also be divided into seven characteristic days from Monday to Sunday according to the weekly similarity of traffic flow.

302. Calculate a distance of each vehicle speed data object in the target vehicle speed data set according to the preset parameter. The preset parameter is a density threshold in a method based on density clustering, and the parameter can be set by the experience of a technician.

303. Sort all of the above-distances in ascending order. For all of the above vehicle speed data objects, the fc - distance is sorted in ascending order. Then, according to the ordinate as fc - distance, the abscissa is sorted by the number of points, such as: 1, 2, 3, etc., respectively used to represent the first vehicle speed data, the second vehicle speed data, the third vehicle speed data, etc. Draw the sorted values.

304. Select the distance-correspondence corresponding to the maximum distance change as the radius parameter of the target vehicle speed data set. A derivative method can be used to determine the magnitude of the -distance increment change. The fc - distance is used as a dependent variable, and the number of the points is used as an argument. The magnitude of the change in the increment of the distance is calculated relative to the increment of the value of the point, and the corresponding distance-distance is taken as the radius parameter of the target vehicle speed data set. The radius parameter is the value of the radius of the field in the method based on density clustering.

305. After determining the values of the density threshold and the domain radius, according to the density clustering-based method, the vehicle speed data object corresponding to the distance of the radius parameter is identified as the abnormal vehicle speed data. After the abnormal vehicle speed data is identified, the abnormal vehicle speed data can be deleted.

306. Identify a vehicle speed data object corresponding to the distance of the radius parameter as boundary vehicle speed data.

307. Identify a vehicle speed data object corresponding to a k-distance smaller than the radius parameter as core vehicle speed data. The method based on density clustering not only realizes the identification of abnormal vehicle speed data from the target dataset, but also identifies the boundary vehicle speed data and the core vehicle speed data, which provides a basis for studying the vehicle speed change in the historical traffic flow.

The embodiment of the present invention further provides an apparatus for identifying abnormal vehicle speed data. As shown in FIG. 4, the method includes: a loading unit 41, a calculating unit 42, a selecting unit 43, a first identifying unit 44, and a second label. The unit 45 and the third identification unit 46 are identified.

The loading unit 41 is configured to load the target data set into the cache. The target data set may be a vehicle speed value at a specific time point of a week characteristic day on a specified road link. After the loading unit 41 loads the target data set, the calculating unit 42 is configured to calculate a distance of each vehicle speed data object in the target vehicle speed data set according to the preset parameter.

Then, the selecting unit 43 is configured to select a radius parameter of the target vehicle speed data set from all of the above-mentioned distances. The selection unit 43 includes a sorting module 43A and a selection module 43B. The sorting module 43A is used to sort all of the above-distances in ascending order. The selection module 43B is configured to select a distance corresponding to the maximum distance change maximum as the radius parameter of the target vehicle speed data set. The radius parameter is the value of the radius of the field in the method based on density clustering. The density threshold and the domain radius in the density cluster based method are determined by the calculation unit 42 and the selection unit 43.

Then, according to the density clustering based method, the first identification unit 44 is configured to be larger than the radius by comparing the distance of each vehicle speed data object in the target data set with the radius parameter of the target vehicle speed data set. The vehicle speed data object corresponding to the k-distance of the parameter is identified as abnormal vehicle speed data. The second identification unit 45 is configured to identify a vehicle speed data object corresponding to the distance of the radius parameter as the boundary vehicle speed data. The third identification unit 46 is configured to identify a vehicle speed data object corresponding to a distance smaller than the radius parameter as core vehicle speed data. Therefore, based on the density clustering method, the abnormal vehicle speed data is identified from the target data set.

The embodiments of the present invention are mainly applied to the field of intelligent transportation, and the method based on density clustering is used to identify abnormal vehicle speed data.

Through the description of the above embodiments, those skilled in the art can clearly understand that the present invention can be implemented by means of software plus necessary general hardware, and of course, by hardware, but in many cases, the former is a better implementation. . Based on such understanding, the technical solution of the present invention, which is essential or contributes to the prior art, may be embodied in the form of a software product stored in a readable storage medium, such as a floppy disk of a computer. , hard disk or CD, etc., including a number of instructions to make a computer device (can be an individual) A computer, server, or network device, etc.) performs the methods described in various embodiments of the present invention. The above is only a specific embodiment of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art can easily think of changes or substitutions within the technical scope of the present invention. It should be covered by the scope of the present invention. Therefore, the scope of the invention should be determined by the scope of the claims.

Claims

Claim

A method for identifying abnormal vehicle speed data, comprising:

Selecting a radius parameter of the target vehicle speed data set from all of the above-distances;

The vehicle speed data object corresponding to the distance corresponding to the radius parameter is identified as abnormal vehicle speed data.

2. The method for identifying abnormal vehicle speed data according to claim 1, wherein the radius parameter of the target vehicle speed data set selected from the distances of all vehicle speed data objects comprises:

Sort all of the above - distances in ascending order;

Selecting - the distance corresponding to the maximum distance change maximum is used as the radius parameter of the target vehicle speed data set.

The method for identifying abnormal vehicle speed data according to claim 1, further comprising:

Identifying a vehicle speed data object corresponding to the distance parameter of the radius parameter as boundary vehicle speed data;

A vehicle speed data object corresponding to the distance-less distance parameter is identified as core vehicle speed data.

4. An apparatus for identifying abnormal vehicle speed data, comprising:

a calculating unit, configured to calculate a distance of each vehicle speed data object in the target vehicle speed data set according to the preset parameter;

a selection unit, configured to select a radius parameter of the target vehicle speed data set from all the above-distances;

5. The apparatus for identifying abnormal vehicle speed data according to claim 4, wherein The selection unit includes:

a sorting module for sorting all of the above-distances in ascending order;

And a selection module, configured to select a distance corresponding to the maximum distance increment change as a radius parameter of the target vehicle speed data set.

The device for identifying abnormal vehicle speed data according to claim 4, further comprising:

a second identification unit, configured to identify a vehicle speed data object corresponding to the distance corresponding to the radius parameter as boundary vehicle speed data;

And a third identification unit, configured to identify a vehicle speed data object corresponding to the distance smaller than the radius parameter as core vehicle speed data.