CN111275084A - K-nearest neighbor algorithm-based method and device for determining collision danger category of encountering ship - Google Patents
K-nearest neighbor algorithm-based method and device for determining collision danger category of encountering ship Download PDFInfo
- Publication number
- CN111275084A CN111275084A CN202010043558.3A CN202010043558A CN111275084A CN 111275084 A CN111275084 A CN 111275084A CN 202010043558 A CN202010043558 A CN 202010043558A CN 111275084 A CN111275084 A CN 111275084A
- Authority
- CN
- China
- Prior art keywords
- ship
- collision
- encountering
- nearest neighbor
- sample
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2413—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on distances to training or reference patterns
Abstract
The application belongs to the field of intelligent ship navigation control, and particularly relates to a method and a device for determining the type of a collision danger of an encountering ship based on a K nearest neighbor algorithm, aiming at solving the problems that the calculation process is complex, the data statistics is large when indexes are too many, and the weight is difficult to determine in the existing method. The method comprises the steps of obtaining characteristic values of a plurality of encountering ship collision risk evaluation indexes based on position information and motion information of a ship and a target ship which are obtained in advance, taking the characteristic values of the plurality of encountering ship collision risk evaluation indexes as data samples to be classified, classifying the data samples to be classified through a collision risk classification model which is established in advance, and obtaining corresponding encountering ship collision risk classes; the collision danger classification model is established based on a K nearest neighbor algorithm. The method has the advantages of simple calculation, no need of parameter estimation, insensitivity to abnormal data and the like; the classification result obtained is more in line with the actual situation of navigation than just obtaining the collision risk.
Description
Technical Field
The application belongs to the field of intelligent ship navigation control, and particularly relates to a method and a device for determining the collision danger category of a meeting ship based on a K nearest neighbor algorithm.
Background
With the continuous development of shipping business, the number of ships is increased, the tonnage is increased, the speed is increased, and ship collision accidents happen occasionally. The ship collision accident not only endangers the life safety of people, but also causes the pollution of marine environment. Therefore, many scholars are dedicated to the research of ship collision avoidance decision and the research of collision avoidance automation, the intelligent degree of ships is improved, and many navigation aids are introduced into navigation practice, so that more choices are provided for reducing ship collision accidents. The key technology of ship collision avoidance is the judgment of collision danger. The existing method for detecting the ship collision risk degree obtains dynamic data of the ship threatening navigation by a ship automatic identification system, calculates values of influencing elements and membership values of all elements, distributes weights to the five elements by utilizing a hierarchical parting method, calculates confidence values of all elements on corresponding evaluation levels and confidence vectors of the ship threatening on all evaluation levels, and obtains the ship collision risk degree of the ship threatening through definite processing of the threat ship threat degree. Generally, pairwise comparison of the analytic hierarchy process is performed by using 1 to 9 to illustrate relative importance, if more and more indexes exist, judgment of importance degree between every two indexes is possibly difficult, and even consistency of single-level sequencing and total sequencing is affected, so that consistency inspection cannot pass.
In summary, the existing method is complex in calculation process, when the indexes are too many, the data statistics are large, and the weight is difficult to determine. Therefore, in order to make the decision of the intelligent ship more consistent with the actual navigation, it is necessary to provide a method for determining the collision risk category of the ship based on the K-nearest neighbor algorithm.
Disclosure of Invention
Technical problem to be solved
In order to solve the problems, the application provides a method and a device for determining the collision danger category of the encountering ship based on a K-nearest neighbor algorithm.
(II) technical scheme
In order to achieve the purpose, the invention adopts the main technical scheme that:
a method for determining the collision risk category of an encountered ship based on a K nearest neighbor algorithm comprises the following steps:
step S10, acquiring characteristic values of a plurality of risk evaluation indexes of collision of the meeting ship based on the position information and the motion information of the ship and the target ship which are acquired in advance;
step S20, using the characteristic values of the multiple encountering ship collision risk evaluation indexes as data samples to be classified, and classifying the data samples to be classified through a pre-established collision risk class classification model to obtain corresponding encountering ship collision risk classes;
the collision danger classification model is established based on a K nearest neighbor algorithm, the training sample characteristics of a training sample set adopted by the model are meeting ship collision danger evaluation indexes, and sample labels are corresponding meeting ship collision danger classes of the training samples.
As an improvement of the method of the present invention, the types of the ship collision risk include: safe and collision-free, safe but collision-free, dangerous but passable-to-collision-free and extremely dangerous to avoid.
As an improvement of the method of the present invention, the method for obtaining the training samples in the training sample set comprises:
step S201, collecting position information and motion information of different meeting situations of the ship in meeting;
step S202, acquiring characteristic values of ship motion indexes in different meeting situations based on the acquired position information and motion information, and taking the characteristic values as characteristics of a sample to be marked; the meeting ship motion index comprises the meeting ship collision risk evaluation index;
step S203, according to the meeting condition of the ship, carrying out sample marking on the sample to be marked;
and S204, carrying out expert verification on the marked sample to obtain a training sample.
As an improvement of the method of the present invention, the step S20 of classifying the data samples to be classified by a pre-established collision risk classification model to obtain corresponding collision risk classes of the encountering ship specifically includes:
step S211, calculating the sample distance between the data sample to be classified and each training sample in the training sample set according to a preset distance measurement mode;
step S212, sequencing the sample distances calculated in the step S211 according to an increasing relation to obtain K training samples with the minimum sample distance, wherein K is a positive integer;
s213, acquiring the category of each training sample in the K training samples;
and S214, determining the meeting ship collision danger category corresponding to the data sample to be classified according to the category of each training sample through a preset classification decision rule.
As an improvement of the method of the present invention, the preset distance measure is euclidean distance.
As an improvement of the method of the present invention, the predetermined classification decision rule is a majority decision method.
As an improvement of the method of the invention, the value range of K is as follows: k is greater than or equal to 1 and less than the square root of the number of samples in the training sample set.
As an improvement of the method, the evaluation index of the collision danger of the ship comprises the following steps: the minimum encounter distance and the minimum encounter time, and one or more of a ship spacing, a ship speed ratio, an azimuth angle and a relative heading.
As a modification of the method of the present invention, the "position information and motion information of the own ship and the target ship" in step S10 includes: and (4) real-time position coordinates, navigation speed and course of the ship.
The invention provides a device for determining the collision risk category of a rendezvous ship based on a K nearest neighbor algorithm, which comprises: a risk evaluation index characteristic value calculation module and a risk category classification module;
the risk evaluation index characteristic value calculation module is configured to obtain a plurality of characteristic values of risk evaluation indexes in collision with ships based on position information and motion information of the ships and the target ships which are obtained in advance;
the danger classification module is configured to classify the data samples to be classified through a pre-established collision danger classification model by taking the characteristic values of the multiple encountering ship collision danger evaluation indexes as sample characteristics of the data samples to be classified to obtain corresponding encountering ship collision danger classes;
the collision danger classification model is established based on a K nearest neighbor algorithm, the training sample characteristics of a training sample set adopted by the model are meeting ship collision danger evaluation indexes, and sample labels are corresponding meeting ship collision danger classes of the training samples.
(III) advantageous effects
The beneficial effect of this application is: according to the method for determining the encountering ship collision danger category based on the K nearest neighbor algorithm, the encountering ship collision danger category is used as the classification label to perform danger category classification on meeting situations of different categories, and compared with the existing method, the method has the advantages of being simple in calculation, free of parameter estimation, small in high index calculation amount, insensitive to abnormal values and the like.
The multiple factors of the nearest meeting distance, the minimum meeting time, the ship spacing, the ship speed ratio, the azimuth angle and the relative course are used as risk degree analysis indexes, so that the inaccuracy of judgment caused by a single index can be avoided.
The data in the training sample set and the corresponding labels are verified and confirmed by experts, and the input samples are classified based on the training sample set, so that the classification is more accurate, and a more reliable basis can be provided for the safe navigation decision of the intelligent ship.
When the risk evaluation of the ship meeting is carried out to be used as a basis for determining key avoidance of the ship, the risk category of the ship meeting collision is divided into four categories, namely safe and collision-free possibility, safe but collision-free possibility, dangerous but collision-free passing and extreme dangerous and collision-free, and compared with the case of only obtaining the collision risk degree, the method is more in line with the decision-making habit of experts in navigation and the actual situation of navigation.
The method can obtain the classification result of 'extreme danger and no collision avoidance', and guide the intelligent ship to take collision avoidance decisions for protecting personnel and reducing loss.
Drawings
The invention is described with the aid of the following figures:
fig. 1 is a schematic flow chart of a method for determining a collision risk category of a ship encountered based on a K-nearest neighbor algorithm in an embodiment of the present invention.
Detailed Description
For the purpose of better explaining the present invention and to facilitate understanding, the present invention will be described in detail by way of specific embodiments with reference to the accompanying drawings.
The method classifies the evaluation index set of the collision danger of the ships in the meeting through a K nearest neighbor algorithm so as to determine the collision danger category of the ships in the meeting, and as shown in figure 1, the method comprises the following steps:
step S10, acquiring characteristic values of a plurality of risk evaluation indexes of collision of the meeting ship based on the position information and the motion information of the ship and the target ship which are acquired in advance;
step S20, using the characteristic values of the multiple encountering ship collision risk evaluation indexes as data samples to be classified, and classifying the data samples to be classified through a pre-established collision risk classification model to obtain corresponding encountering ship collision risk classes;
the collision danger classification model is established based on a K nearest neighbor algorithm, training sample characteristics of a training sample set adopted by the model are meeting ship collision danger evaluation indexes, and sample labels are corresponding meeting ship collision danger classes of the training samples.
The steps of the method of the present invention are described in detail to facilitate an understanding of the method of the present invention.
Step S1, acquiring the position, speed and course of the ship and the target ship through onboard sensors such as a Global Positioning System (GPS), an Automatic Identification System (AIS) of the ship, a radar and the like, and calculating to obtain a nearest meeting Distance (DCPA) and a minimum meeting Distance (DCPA)Forming data sample points data to be classified according to index characteristic values of Time (TCPA), ship spacing (dist), ship speed ratio lambda, azimuth angle B and relative course Rctest(DCPAtest，TCPAtest，Dtest，λtest，Btest，Rctest)。
Step S2, calculating euclidean distances between the data sample points to be classified in step S1 and the data sample points in the training sample set according to formula (1):
where i represents the data number in the training sample set, DCPAtestRepresenting the closest encounter distance, TCPA, of the data samples to be classifiedtestRepresenting the minimum encounter time, D, of the data samples to be classifiedtestRepresenting the ship spacing, λ, of the data samples to be classifiedtestRepresenting the ship speed ratio of the data samples to be classified, BtestRepresenting the azimuth, Rc, of the data sample to be classifiedtestRepresenting the relative heading of the data sample to be classified.
The method for obtaining the training sample set comprises the following steps:
step S21, collecting position information and motion information of different meeting situations of the ship in meeting;
step S22, acquiring characteristic values of the ship motion indexes in different meeting situations based on the acquired position information and motion information, and taking the characteristic values as the characteristics of the sample to be marked; the meet ship motion indexes comprise meet ship collision risk evaluation indexes;
step S23, according to the meeting condition of the ship, carrying out sample marking on the sample to be marked; the categories of data samples include: the safety is safe without collision possibility, the safety is safe but has collision possibility and danger but can avoid collision and pass and the extreme danger can not avoid collision;
and step S24, carrying out expert verification on the marked sample to obtain a training sample.
In this embodiment, the meeting ship motion indexes are a closest meeting Distance (DCPA), a minimum meeting Time (TCPA), a ship distance (dist), a ship speed ratio λ, an azimuth angle B, and a relative heading Rc.
And step S3, sorting the Euclidean distances calculated in the step S2 according to an increasing relationship.
And S4, selecting the data sample points in the K training sample sets with the minimum distance according to the sequence in the step S3. The value range of K is as follows: k is greater than or equal to 1 and less than the square root of the number of samples in the training sample set.
And S5, determining the occurrence frequency of the class of the K data sample points according to the K data sample points in the training sample set in the step S4.
And step S6, according to the frequency in the step S5, the class with the highest frequency of occurrence is used as the ship collision danger class of the target ship and the ship.
The invention embodiment is a device for determining the collision danger category of an encountering ship based on a K nearest neighbor algorithm, which comprises: a risk evaluation index characteristic value calculation module and a risk category classification module;
the risk evaluation index characteristic value calculation module is configured to obtain a plurality of characteristic values of risk evaluation indexes in collision with ships based on position information and motion information of the ships and the target ships which are obtained in advance;
the danger classification module is configured to classify the data samples to be classified through a pre-established collision danger classification model by taking the characteristic values of the multiple encountering ship collision danger evaluation indexes as sample characteristics of the data samples to be classified to obtain corresponding encountering ship collision danger classes;
the collision danger classification model is established based on a K nearest neighbor algorithm, training sample characteristics of a training sample set adopted by the model are meeting ship collision danger evaluation indexes, and sample labels are corresponding meeting ship collision danger classes of the training samples.
It should be noted that, the device for determining the risk category of collision of a ship based on the K-nearest neighbor algorithm provided in the foregoing embodiment is only illustrated by the division of the functional modules, and in practical applications, the functions may be allocated to different functional modules according to needs, that is, the modules or steps in the embodiment of the present invention are further decomposed or combined, for example, the modules in the foregoing embodiment may be combined into one module, or may be further split into multiple sub-modules, so as to complete all or part of the functions described above. The names of the modules and steps involved in the embodiments of the present invention are only for distinguishing the modules or steps, and are not to be construed as unduly limiting the present invention.
It can be clearly understood by those skilled in the art that, for convenience and brevity of description, the specific working process and related descriptions of the above-described apparatus may refer to the corresponding process in the foregoing method embodiments, and are not described herein again.
It should be understood that the above description of specific embodiments of the present invention is only for the purpose of illustrating the technical lines and features of the present invention, and is intended to enable those skilled in the art to understand the contents of the present invention and to implement the present invention, but the present invention is not limited to the above specific embodiments. It is intended that all such changes and modifications as fall within the scope of the appended claims be embraced therein.
Claims (10)
1. A method for determining the collision risk category of an encountered ship based on a K nearest neighbor algorithm is characterized by comprising the following steps:
step S10, acquiring characteristic values of a plurality of risk evaluation indexes of collision of the meeting ship based on the position information and the motion information of the ship and the target ship which are acquired in advance;
step S20, using the characteristic values of the multiple encountering ship collision risk evaluation indexes as data samples to be classified, and classifying the data samples to be classified through a pre-established collision risk class classification model to obtain corresponding encountering ship collision risk classes;
the collision danger classification model is established based on a K nearest neighbor algorithm, the training sample characteristics of a training sample set adopted by the model are meeting ship collision danger evaluation indexes, and sample labels are corresponding meeting ship collision danger classes of the training samples.
2. The K-nearest neighbor algorithm-based method for determining the collision risk category of the encountering ship, according to claim 1, wherein the collision risk category of the encountering ship comprises: safe and collision-free, safe but collision-free, dangerous but passable-to-collision-free and extremely dangerous to avoid.
3. The K-nearest neighbor algorithm-based method for determining the collision risk class of encountering ships according to claim 2, wherein the method for obtaining the training samples in the training sample set comprises:
step S201, collecting position information and motion information of different meeting situations of the ship in meeting;
step S202, acquiring characteristic values of ship motion indexes in different meeting situations based on the acquired position information and motion information, and taking the characteristic values as characteristics of a sample to be marked; the meeting ship motion index comprises the meeting ship collision risk evaluation index;
step S203, according to the meeting condition of the ship, carrying out sample marking on the sample to be marked;
and S204, carrying out expert verification on the marked sample to obtain a training sample.
4. The method for determining the encountering ship collision risk category based on the K-nearest neighbor algorithm of claim 3, wherein the step S20 of classifying the data samples to be classified through a pre-established collision risk category classification model to obtain the corresponding encountering ship collision risk category specifically comprises:
step S211, calculating the sample distance between the data sample to be classified and each training sample in the training sample set according to a preset distance measurement mode;
step S212, sequencing the sample distances calculated in the step S211 according to an increasing relation to obtain K training samples with the minimum sample distance, wherein K is a positive integer;
s213, acquiring the category of each training sample in the K training samples;
and S214, determining the meeting ship collision danger category corresponding to the data sample to be classified according to the category of each training sample through a preset classification decision rule.
5. The K-nearest neighbor algorithm-based method for determining collision risk category of encountering ships according to claim 4, wherein the preset distance metric is Euclidean distance.
6. The K-nearest neighbor algorithm-based method for determining the collision risk category of an encountering ship according to claim 4, wherein the preset classification decision rule is a majority decision method.
7. The method for determining the encountering ship collision danger category based on the K nearest neighbor algorithm according to claim 4, wherein the value range of K is as follows: k is greater than or equal to 1 and less than the square root of the number of samples in the training sample set.
8. The K-nearest neighbor algorithm-based method for determining collision risk of encountering ships according to any one of claims 1-7, wherein the evaluation index of collision risk of encountering ships comprises: the minimum encounter distance and the minimum encounter time, and one or more of a ship spacing, a ship speed ratio, an azimuth angle and a relative heading.
9. The K-nearest neighbor algorithm-based method for determining collision risk of encountering ships according to any one of claims 1-7, wherein the step S10 of determining the position information and the motion information of the ship and the target ship comprises: and (4) real-time position coordinates, navigation speed and course of the ship.
10. A device for determining the collision risk category of an encountered ship based on a K-nearest neighbor algorithm is characterized by comprising: a risk evaluation index characteristic value calculation module and a risk category classification module;
the risk evaluation index characteristic value calculation module is configured to obtain a plurality of characteristic values of risk evaluation indexes in collision with ships based on position information and motion information of the ships and the target ships which are obtained in advance;
the danger classification module is configured to classify the data samples to be classified through a pre-established collision danger classification model by taking the characteristic values of the multiple encountering ship collision danger evaluation indexes as sample characteristics of the data samples to be classified to obtain corresponding encountering ship collision danger classes;
the collision danger classification model is established based on a K nearest neighbor algorithm, the training sample characteristics of a training sample set adopted by the model are meeting ship collision danger evaluation indexes, and sample labels are corresponding meeting ship collision danger classes of the training samples.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010043558.3A CN111275084A (en) | 2020-01-15 | 2020-01-15 | K-nearest neighbor algorithm-based method and device for determining collision danger category of encountering ship |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010043558.3A CN111275084A (en) | 2020-01-15 | 2020-01-15 | K-nearest neighbor algorithm-based method and device for determining collision danger category of encountering ship |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111275084A true CN111275084A (en) | 2020-06-12 |
Family
ID=70999039
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010043558.3A Pending CN111275084A (en) | 2020-01-15 | 2020-01-15 | K-nearest neighbor algorithm-based method and device for determining collision danger category of encountering ship |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111275084A (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104050329A (en) * | 2014-06-25 | 2014-09-17 | 哈尔滨工程大学 | Method for detecting degree of risk of ship collision |
| CN110009936A (en) * | 2019-03-15 | 2019-07-12 | 北京海兰信数据科技股份有限公司 | A kind of ship auxiliary collision prevention method for crowded waters |
| CN110543907A (en) * | 2019-08-29 | 2019-12-06 | 交控科技股份有限公司 | fault classification method based on microcomputer monitoring power curve |
-
2020
- 2020-01-15 CN CN202010043558.3A patent/CN111275084A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104050329A (en) * | 2014-06-25 | 2014-09-17 | 哈尔滨工程大学 | Method for detecting degree of risk of ship collision |
| CN110009936A (en) * | 2019-03-15 | 2019-07-12 | 北京海兰信数据科技股份有限公司 | A kind of ship auxiliary collision prevention method for crowded waters |
| CN110543907A (en) * | 2019-08-29 | 2019-12-06 | 交控科技股份有限公司 | fault classification method based on microcomputer monitoring power curve |
Non-Patent Citations (1)
| Title |
|---|
| XIN GAO ET AL.: "A rapidly exploring random tree optimization algorithm for space robotic manipulators guided by obstacle avoidance independent potential field", INTERNATIONAL JOURNAL OF ADVANCED ROBOTIC SYSTEMS, pages 1 - 17 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Zhou et al. | Ship classification based on ship behavior clustering from AIS data | |
| Alizadeh et al. | Vessel trajectory prediction using historical automatic identification system data | |
| CN110888126B (en) | Unmanned ship information perception system data comprehensive processing method based on multi-source sensor | |
| CN110188093A (en) | A kind of data digging system being directed to AIS information source based on big data platform | |
| CN109543715A (en) | A kind of ship air route is extracted and the method for track deviation detection | |
| CN108960421B (en) | Improved online forecasting method for speed of unmanned surface vehicle based on BP neural network | |
| CN111142538B (en) | Unmanned ship operation strategy real-time decision-making method for water environment monitoring | |
| CN111444769B (en) | Laser radar human leg detection method based on multi-scale self-adaptive random forest | |
| CN110196962A (en) | A kind of air speed abnormality recognition method based on Density Estimator | |
| CN114333423B (en) | Narrow water channel navigation collision risk assessment method | |
| CN115050214B (en) | AIS data-based ship collision risk prediction method | |
| CN113128826A (en) | Navigation risk prediction method, navigation risk prediction device and storage medium | |
| Lu et al. | Shape-based vessel trajectory similarity computing and clustering: A brief review | |
| CN110826891B (en) | Relative collision risk degree obtaining method based on ship cluster situation | |
| CN111275084A (en) | K-nearest neighbor algorithm-based method and device for determining collision danger category of encountering ship | |
| CN112733935A (en) | Fishing boat fishing mode prediction method based on Stacking algorithm | |
| Zhang et al. | Research on ship classification based on trajectory association | |
| CN114331042A (en) | Ship collision risk analysis method | |
| CN115456258A (en) | Method for predicting transportation capacity of competitor ship and computer readable medium | |
| CN111882135B (en) | Internet of things equipment intrusion detection method and related device | |
| CN114047508A (en) | Sea radar detection range statistical evaluation method and system | |
| Westerdijk et al. | Classifying vessel types based on AIS data | |
| Wang et al. | Analysis of navigation characteristics of inland watercraft based on DBSCAN clustering algorithm | |
| Li et al. | A Novel Hybrid Approach for Detecting Abnormal Vessel Behavior in Maritime Traffic | |
| Newaliya et al. | A review of maritime spatio-temporal data analytics |
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 |