CN105897488B - A kind of method for visualizing of radio-signal data - Google Patents
A kind of method for visualizing of radio-signal data Download PDFInfo
- Publication number
- CN105897488B CN105897488B CN201610419842.XA CN201610419842A CN105897488B CN 105897488 B CN105897488 B CN 105897488B CN 201610419842 A CN201610419842 A CN 201610419842A CN 105897488 B CN105897488 B CN 105897488B
- Authority
- CN
- China
- Prior art keywords
- signal
- radio
- timeslice
- cluster
- frequency
- 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.)
- Active
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/22—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks comprising specially adapted graphical user interfaces [GUI]
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04B—TRANSMISSION
- H04B17/00—Monitoring; Testing
- H04B17/30—Monitoring; Testing of propagation channels
- H04B17/309—Measuring or estimating channel quality parameters
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04B—TRANSMISSION
- H04B17/00—Monitoring; Testing
- H04B17/30—Monitoring; Testing of propagation channels
- H04B17/309—Measuring or estimating channel quality parameters
- H04B17/318—Received signal strength
Abstract
The present invention provides a kind of method for visualizing of radio-signal data, step 1: obtaining the radio-signal data extracted from frequency spectrum data and original level sampled data；Step 2: drawing frequency-bandwidth scatter plot；Step 3: radio-signal data being clustered using clustering algorithm；Step 4: dividing timeslice；Step 5: to mean center frequency, average bandwidth, average signal-to-noise ratio and the average signal strength of each timeslice of each cluster calculation；Step: 6: drawing signal flow diagram.Utilize the various features of signal flow diagram efficient coding radio signal, a variety of important features of signal data more discrete on time-frequency are smoothly shown, preferably show the multiple features time-varying mode of radio signal, accelerates analysis personnel to the macroscopic view perception efficiency of radio signal time-varying mode.
Description
Technical field
The present invention relates to information visualization field, in particular to a kind of method for visualizing of radio-signal data.
Background technique
Radio-frequency spectrum is a kind of limited natural resources, is the underlying carrier of modern wireless information transfer.In order to reduce
Interfering with each other between radio signal, in order to guarantee important communication link it is unimpeded, in order to preferably be provided using limited frequency spectrum
Source, especially in the key areas such as airport and border and in Security ensuring of important activities task, enhanced radio spectrum monitoring and
Radio signal management, to safeguard national security and promote national economy steady development have own strategic significance.
Being monitored to radio-frequency spectrum in certain frequency range is the movable foundation stone of radio control, various monitoring device acquisitions
Spectrum monitoring data be frequency spectrum occupancy rate calculate, interference signal monitoring and frequency spectrum resource division provide information source.Traditional frequency
Modal data analysis first shows the method for visualizing such as frequency spectrum amplitude frequency diagram, time-frequency figure and sunset glow figure, then by analyzing personnel
According to the time-frequency distributions of various spectrogram perceptual signals and find signal of interest.But this way is to analysis personnel specialty quality
It is required that very high, ordinary user does not pass through professional training and accumulates practical operation experience and is difficult timely and accurately to identify nothing from spectrogram
Line electric signal, and when monitoring time is elongated or frequency range in there is more radio signal when, manual analysis efficiency can fast prompt drop
It is low, it is associated with so as to cause active detecting unknown signal is difficult to the complexity analyzed between signal, radio control excessively relies on report
Equal passive managements mode.
Converting radio-signal data for radio frequency line modal data by signal processing and mathematical analysis is radio tube
The new direction of reason, each data record of radio-signal data represent the radio signal of a physical presence, each data
Item reflects the multidimensional characteristics such as the centre frequency, bandwidth, signal-to-noise ratio of corresponding radio signal.This structuring degree is higher, believes
Number more readily apparent radio-signal data of feature is the useful supplement of radio frequency line modal data, be conducive to merge public frequency library and
The basic datas resources such as station library are authorized, solidification expertise is also beneficial to, are promoted and various intelligent data analysis means
Using.But radio-signal data the problem of there is also some urgent need to resolve in practical applications, first radio-signal data
It extracts at any time, a large amount of signal characteristic records may correspond to the non-instantaneous radio signal of same reality, due to monitoring environment
Unstability and be time-multiplexed same frequency there may be interference signal and multi-user, be difficult correctly using automatic method
Ground is by actual wireless electric signal cluster signal characteristic record；Then, the visualizations side such as traditional scatter plot, line chart and parallel coordinates
Method is all difficult to show the time-varying mode of radio signal multiple features, and it is visual that brainstrust lacks valid wireless electrical signal data
Radio signal is observed, compares and analyzed to change method.
Summary of the invention
The purpose of the present invention is to provide a kind of effective method for visualizing to observe, compare and analyze radio signal,
The present invention provides it is a kind of new can accurately show radio signal multiple features time-varying mode and convenient for comparative analysis can
Depending on change method.
The technical solution of the present invention is as follows:
A kind of method for visualizing of radio-signal data, including the following steps:
Step 1: obtain radio-signal data:
Radio-signal data is included in a period [tstart,tend] in all n nothings for detecting in detection frequency range
Line electric signal { S1,S2,…Si,…,Sn, the feature that each radio signal includes has centre frequency (freq), bandwidth
(baud), signal-to-noise ratio (snr), signal strength (dbm) and timestamp, timestamp indicate the time point that the signal is detected, mark
Remember signal SiCentre frequency, bandwidth, signal-to-noise ratio, signal strength and the timestamp of (1≤i≤n) be respectively
Step 2: frequency-bandwidth scatter plot is drawn according to step 1 radio-signal data obtained:
X-axis indicates that frequency, Y-axis indicate that bandwidth draws scatter plot in scatter plot, and frequency range is collected for radio signal
Frequency separation, bandwidth range is 0 maximum value that bandwidth in all radio signals is obtained to step 1；According to each letter
Number centre frequency and bandwidth value find its position in frequency-bandwidth scatter plot, draw all signaling points；Record each nothing
Line electric signal SiCoordinate S in frequency-bandwidth scatter ploti(xi,yi)(1≤i≤n)；It is encoded, is filled out according to signal strength color
Fill signaling point；Signal strength color coding refers to the ascending RGB color degree item for being encoded to gradual change of signal strength indication；
Step 3: step 1 radio-signal data obtained being clustered using clustering algorithm, obtains k cluster
{C1,C2,C3,…,Ck, M noise spot；
Radio-signal data extraction carries out at timed intervals, and the non-instantaneous radio signal detected in frequency range can quilt
Repeatedly record, due to the unstability of signal and acquisition equipment, the feature of same radio signal will appear a certain range of wave
It is dynamic, so needing to cluster using radio-signal data of the clustering algorithm to acquisition.
Step 4: divide timeslice:
By entire radio signal acquisition time section [tstart,tend] it is divided into length p identical, the i.e. p time
Piece, j-th of period of label is tj(1≤j≤p)
To each cluster Ci(1≤i≤k), when all signals in the cluster are divided into corresponding according to its timestamp
Between in piece, each cluster Ci(1≤i≤k) obtains a timeslice dividing sequence Csti={ sti1,sti2,…,stij,…,
stip(1≤i≤k), wherein stij(1≤j≤p) (1≤i≤k) indicates cluster CiTimeslice t is divided into (1≤i≤k)j(1
≤ j≤p) in signal set, when the signal number in the set be 0 when,Otherwise by stijIt is denoted as set
{Sq,…,Sq+r, signal number is r+1 in the set, 0≤r≤n-1, and the timestamp of all signals in the set is in the time
Piece tj(1≤j≤p) is interior, wherein SqAnd Sq+rRespectively indicate all signaling points in the set temporally stab sequence after timestamp most
Small and maximum signal, q and q+r are respectively signal SqAnd Sq+rIn signal set { S1,S2,…Si,…,SnIn serial number；
Step 5: each cluster C according to obtained in step 4iThe timeslice dividing sequence Cst of (1≤i≤k)i={ sti1,
sti2,…,stip(1≤i≤k), calculate each cluster CiTimeslice t is divided into (1≤i≤k)jIt is all in (1≤j≤p)
Mean center frequency (mFreq), average bandwidth (mBaud), average signal-to-noise ratio (mSnr) and the average signal strength of signal
(mDbm), ifThen without calculating；
Step 6: signal flow diagram is drawn according to the calculated result of step 3 cluster result obtained and step 5, it is specific to wrap
Include following steps:
Step 6.1): drawing signal flow diagram coordinate system, and wherein X-axis indicates that frequency, Y-axis indicate the time；Frequency range is nothing
The collected frequency separation of line electrical signal data, time range are the period that radio-signal data is collected；Each cluster
It is drawn into a bars stream；
Step 6.2): to cluster Ci(1≤i≤k), according to stijThe value mFreq of the mean center frequency of interior signalij, believing
Corresponding frequency and timeslice t in number flow graph coordinate systemjInitial position graphical pointv A, according to being divided into timeslice tj+1Interior signal
Mean center frequency mFreqi(j+1), corresponding frequency and timeslice t in signal flow diagram coordinate systemj+1Initial position, that is, timeslice
tjFinal position graphical pointv B, tie point A and point B, draw a line segment AB, with determination the cluster corresponding to signal stream exist
Timeslice tjCenter；
If timeslice tjWithout subsequent time piece t continuous therewithj+1OrThen timeslice tjCorresponding signal stream
It does not draw；
Step 6.3): according to the clustering to timeslice tjThe average bandwidth of interior signal determines the width of signal stream, and
The left margin and right margin of signal stream are drawn, left margin and right margin are symmetrical about line segment AB；
Step 6.4): according to the clustering to timeslice tjThe value and signal strength face of the average signal strength of interior signal
Color coding, filling left margin to the region between line segment AB；
Step 6.5): according to the clustering to timeslice tjThe average signal-to-noise ratio and signal-to-noise ratio color of interior signal are compiled
Code, filling line segment AB to right border area；Signal-to-noise ratio color coding, which refers to, is encoded to gradual change gradually for snr value is ascending
Graying angle value；
Step 6.6): it repeats the above steps 6.2) to step 6.5), successively draws Ci(1≤i≤k) all timeslices,
And all clusters.
It is that R circle indicates a signal with a radius in the scatter plot of the step 2, according to the signal strength of the signal,
The circle of color filling corresponding to its value is found in signal strength color coding.
The signal strength color encodes concrete methods of realizing are as follows: color mode uses RGB mode, in color RGB
Parameter is 0 to 255 with (incremental) the realization gradual change of certain increments, the value range of three parameters of RGB color mode.Initially
Color is red (255,0,0), it is assumed that step-length i, the second parameter of priming color i incremented by successively, fade to yellow (255,
255,0), then first parameter is successively successively decreased i, fades to green (0,255,0), then third parameter i incremented by successively, gradually
Become cyan (0,255,255), then second parameter successively successively decreases i, fades to blue (0,0,255), then third ginseng
The i that successively successively decreases is counted, black (0,0,0) is finally faded to.The distance length that black is passed through is gradient to from priming color red
Distance is (5*255)/i, converts (0, (5*255)/i) range for the signal strength indication of (0, -140) dbm range, i.e.,
Each signal strength indication s (dbm) in (0, -140) range is in corresponding one unique value of (0, (5*255)/i) range
Mindex, obtaining mindex, passing through following manner obtains color color (r, g, b) corresponding to the signal strength indication later:
If mindex >=0 and mindex≤255/i, enables
R=255；
G=map (mindex, 0,255/i, 0,255)；
B=0；
If mindex > 255/i and mindex≤255*2/i, enable
R=map (mindex, 255/i, 255*2/i, 0,255)；
G=255；
B=0；
If mindex > 255*2/i and mindex≤255*3/i, enable
R=0；
G=255
B=map (mindex, 255*2/i, 255*3/i, 0,255)；
If mindex > 3*255/i and mindex≤255*4/i, enable
R=0；
G=map (mindex, 255*3/i, 255*4/i, 0,255)；
B=255；
If mindex > 4*255/i and mindex≤255*5/i, enable
R=0；
G=0；
B=map (mindex, 255*4/i, 255*5/i, 0,255)；
Wherein map (x, x1, x2, x3, x4) function is to return to x of the value in (x1, x2) range to be transformed into (x3, x4) model
The respective value enclosed；
In this way, make 0 to -140dbm signal strength indication be encoded to by red fade to yellow fade to again it is green
Color fades to cyan again and fades to blue again, finally fades to the coloration item of black.
The signal-to-noise ratio color encodes concrete methods of realizing are as follows: uses greyscale color mode, i.e., with 0 to 255 difference ash
Angle value indicates color, and 0 indicates black, 255 indicate white；0 to 255 ranges are converted by the snr value that range is 0 to 100
Color value, i.e., a color value of corresponding (0, the 255) range of each snr value in (0,100) range, which is
It is encoded for the greyscale color of the signal-to-noise ratio；0 to 100 snr value is encoded to gradual change gray scale from white to black；
The step 3 clusters radio-signal data using DBScan clustering algorithm, the specific steps are as follows:
Step 3.1: two parameters needed for setting DBScan clustering algorithm: the quantity at least put in radius eps and neighborhood
minpts；The original state for marking all the points is not visited (unvisited)；
Step 3.2: calculate step 2 frequency-bandwidth scatter plot Euclidean distance:Make
For two signal S in clustering algorithmi,SjThe distance between (1≤i, j≤n, i ≠ j)；
The specific method of DBScan algorithm: the point of optional one not visited (unvisited) starts, and finds out and its distance
All points nearby within eps (including eps).If quantity >=the minpts nearby put, current point and its neighbouring dot
At a cluster, and starting point is marked as having accessed (visited).Then recurrence handles institute in the cluster in the same way
There is the point for being not labeled as having accessed (visited), to be extended to cluster.It, should if the quantity < minpts nearby put
Point is temporarily labeled to be used as noise spot.If cluster is fully extended, i.e., all the points in cluster are marked as having accessed, and then use
Same algorithm goes to handle not visited point；
Step 3.3: obtain DBScan clustering algorithm as a result, total k cluster, M noise spot.
DBScan clustering algorithm is a more representational density-based algorithms.With division and hierarchical clustering
Method is different, and cluster is defined as the maximum set of the connected point of density by it, can be having the region division highdensity to be enough
Cluster, and the cluster of arbitrary shape can be found in the spatial database of noise.
Euclidean distance in DBScan clustering algorithm, in frequency of use-bandwidth scatter plot between signaling pointMeasure two signal Si,SjThe distance between (1≤i, j≤n, i ≠ j)；This be because
Centre frequency and bandwidth are the core features for identifying a radio signal, and centre frequency (unit Mhz) and bandwidth (unit
Db the value difference between) is larger, and the directly meeting of use judges the distance between two signaling points by the shadow of bandwidth when so that clustering
Sound is too big, and the influence of centre frequency can be ignored substantially, so as to obtain result error too big for cluster.
In the step 3, radio-signal data is clustered using clustering algorithm, and the face of customized each cluster
Color；When defining the color of each cluster, the distribution situation of combining wireless electrical signal data signal strength, selection signal intensity coding
It is not used and the color high with the signaling point Fill Color discrimination in the cluster, to better discriminate between out cluster and should
Signaling point in cluster.
In the step 4, the time span of the basis of design radio signal acquisition of p and sample rate selection；Such as: letter
Number acquisition interval time is 1 second, then the length that timeslice can be set is 1 second, then p is equal to [tstart,tend] number of seconds.
In the step 4, in the step 5, C is clusterediTimeslice t is divided into (1≤i≤k)jInstitute in (1≤j≤p)
There are mean center frequency (mFreq), average bandwidth (mBaud), average signal-to-noise ratio (mSnr) and the average signal strength of signal
(mDbm) calculation is as follows:
In the step 6, when drawing signal stream, if there is there is no signal appearance in continuous a period of time, because often
May all there be signal in not all timeslice in a cluster, then discontinuous situation may occur in this stream, be
Two rendering parameters: timeslice signaling point degree of rarefication e and continuous sparse timeslice are arranged in the time discontinuity of display signal
Number z, timeslice signaling point degree of rarefication indicate that the minimum signal number in a timeslice, continuous sparse time the piece number indicate signal
Number is less than the continuous time the piece number of timeslice signal degree of rarefication, if there is the letter in the sequential time slices more than or equal to z
Number point is less than e, then not drawing this segment signal stream.
Beneficial effect
The present invention provides a kind of method for visualizing of radio-signal data, step 1: from frequency spectrum data and original level
Radio-signal data is extracted in sampled data；Step 2: radio-signal data being clustered using clustering algorithm；Step
3: to mean center frequency, average bandwidth, average signal-to-noise ratio and the average signal strength of each timeslice of each cluster calculation；Step
Rapid 4: drawing signal flow diagram.Using the various features of signal flow diagram efficient coding radio signal, by letter more discrete on time-frequency
A variety of important features of number are smoothly shown, preferably the effective multiple features time-varying mould for showing radio signal
Formula convenient for analysis personnel observation, compares and analyzes radio signal, improves analysis personnel to radio signal quantity and microcosmic spy
The identification capability of sign accelerates analysis personnel to the macroscopic view perception efficiency of radio signal time-varying mode.
Detailed description of the invention
Fig. 1 is the flow chart of the method for the invention；
Fig. 2 is frequency-bandwidth two dimension scatter plot；
Fig. 3 is using the frequency bandwidth two dimension scatterplot for distributing color after DBScan clustering algorithm cluster and to each cluster
Figure；
Fig. 4 signal flow diagram.
Specific embodiment
The present invention is described in further detail in the following with reference to the drawings and specific embodiments.
A kind of method for visualizing of radio-signal data, including the following steps:
Step 1: obtaining radio-signal data, radio-signal data is 2016/4/7 15:16:08 to 2016/4/7
15:17:42, all signaling points that frequency range is 952-961Mhz in 94 seconds periods altogether, radio-signal data
Each data record corresponds to a radio signal, and the feature that each radio signal includes has centre frequency (freq), band
Wide (baud), signal-to-noise ratio (snr), signal strength (dbm) and timestamp, indicate time point quilt of the signal represented by timestamp
It detects, altogether includes 6437 radio signals, marking signal SiThe centre frequency of (1≤i≤6437), bandwidth, signal-to-noise ratio, letter
Number intensity and timestamp are respectively
Step 2: frequency-bandwidth scatter plot being drawn according to step 1 radio-signal data obtained, wherein X-axis indicates
Frequency, frequency range 952-961Mhz, Y-axis indicate bandwidth, and bandwidth range 0 arrives 4050db.In frequency-bandwidth scatter plot, often
A signal is represented as the circle that a radius is R, and round color indicates that the signal strength of the signal, signal strength color are encoded to
It fades to yellow by red and fades to green again to fade to cyan again and fade to blue again, the coloration item for finally fading to black is compiled
Code 0 arrives the signal strength indication of -140dbm, tracer signal point SiCoordinate S in frequency-bandwidth scatter ploti(xi,yi)(1≤i≤
n).Frequency-bandwidth scatter plot is as shown in Figure 2；
Step 3: step 1 radio-signal data obtained being clustered using DBScan clustering algorithm.Radio
Signal data extraction carries out at timed intervals, and the non-instantaneous radio signal monitored in frequency range can be recorded repeatedly, due to
The unstability of signal and acquisition equipment, the feature of same radio signal will appear a certain range of fluctuation, so needing to make
It is clustered with radio-signal data of the clustering algorithm to acquisition.DBScan clustering algorithm is a more representational base
In the clustering algorithm of density.Different from division and hierarchy clustering method, cluster is defined as the maximum set of the connected point of density by it,
It can be cluster having region division highdensity enough, and can find the poly- of arbitrary shape in the spatial database of noise
Class.Two signal S in clustering algorithmi,SjThe distance between (1≤i, j≤n, i ≠ j) uses step 2 frequency-bandwidth scatter plot
Euclidean distance:Parameter eps=30, minpts=40 needed for DBScan clustering algorithm is set.
DBScan clustering algorithm obtains 7 cluster { C1,C2,C3,…,C7, 640 noise spots, the color of customized each cluster, letter
The value of number intensity relatively concentrates between -30 to -80dbm, that is, the color put mainly in the orange section for being gradient to green, so
It selects that other unused other colors can be used when cluster color.Frequency-bandwidth after being clustered using DBScan
Scatter plot is as shown in figure 3, each cluster is surrounded using a polygon；
Step 4: timeslice is divided, by entire radio signal acquisition time section 2016/4/715:16:08 to 2016/
4/715:17:42 is divided into 94 sections, and every segment length is 1 second, and j-th of period of label is tj(1≤j≤94).According to step 3 institute
7 obtained clusters, to each cluster CiAll signaling points in the cluster are divided by (1≤i≤7) according to its timestamp
In corresponding timeslice, each cluster Ci(1≤i≤7) obtain a time series Csti={ sti1,sti2,…,sti94}(1≤
I≤7), wherein stij(1≤j≤94) (1≤i≤7) indicate the set { S of a signaling pointq,…,Sq+r, gather interior signal
Number is r (0≤r≤6437), and the timestamp of all signaling points in the set is in tjIn (1≤j≤94) timeslice；
Step 5: each cluster C according to obtained in step 4iSequence C st after the division of (1≤i≤7) timeslicei=
{sti1,sti2,…,sti94(1≤i≤7) calculate each timeslice t to the timeslice dividing sequence of each clusterj(1≤j≤
94) the mean center frequency (mFreq) of signal, average bandwidth (mBaud), average signal-to-noise ratio (mSnr) and average signal strength in
(mDbm), calculation is as follows:
Step 6: the timeslice for each cluster being calculated according to step 5 divide after each timeslice mean center
Frequency (mFreq), average bandwidth (mBaud), average signal-to-noise ratio (mSnr) and average signal strength (mDbm) draw signal flow diagram,
X-axis indicates frequency in signal flow diagram, and Y-axis indicates the time, and the time is incremented by from top to bottom, and frequency range is aerogram number
According to collected frequency separation, time range is the period that radio-signal data is collected；Each cluster is drawn into one
Bars stream.Take first cluster first, draw its first timeslice, according to the value of the centre frequency of first timeslice with
And the center frequency value of second timeslice corresponding frequency and time corresponding to the timeslice in signal flow diagram coordinate system
Straight line is drawn in region, determines signal stream corresponding to the cluster in the center of the timeslice；Then according to the cluster
Average signal bandwidth in the timeslice draws the left margin and right margin of signal stream, and the distance of left margin to right margin is it
The value of signal bandwidth can be multiplied by a coefficient less than 1 since signal bandwidth may be larger, and left margin is straight to centre frequency
The distance of line is the half of its average bandwidth, and the distance of right margin to centre frequency straight line is also two points of its average bandwidth
One of, left margin and right margin are parallel to the straight line of centre frequency straight line；Then according to average signal strength in the timeslice
Value filling left margin arrive centre frequency line region, color encode it is identical as the color coding of signal dot in step 2；Then root
Centre frequency line is filled to right border area according to average signal-to-noise ratio in the timeslice, and color coding uses from white to black gradually
Become the snr value of gray-coded 0 to 100；It repeats the above steps and successively draws remaining timeslice, the last one timeslice is not
It draws.Similarly draw all clusters.When drawing signal stream, if certain stream all goes out without signal in continuous a period of time
Existing, then discontinuous situation may occur in this stream, in order to show the time discontinuity of signal, we are arranged two here
A rendering parameter: signaling point degree of rarefication per second and continuous sparse number of seconds, it is all more sparse if there is continuous z seconds signaling point, that
Signal stream will not draw this section.The signal flow diagram drawn is as shown in Figure 4.
Claims (9)
1. a kind of method for visualizing of radio-signal data, which is characterized in that including the following steps:
Step 1: obtain radio-signal data:
Radio-signal data is included in a period [tstart,tend] in all n radio for detecting in detection frequency range
Signal { S1,S2,…Si,…,Sn, the feature that each radio signal includes has centre frequency, bandwidth, signal-to-noise ratio, signal strength
And timestamp, timestamp indicate the time point that the signal is detected, marking signal SiThe centre frequency of (1≤i≤n), bandwidth,
Signal-to-noise ratio, signal strength and timestamp are respectively
Step 2: frequency-bandwidth scatter plot is drawn according to step 1 radio-signal data obtained:
X-axis indicates that frequency, Y-axis indicate that bandwidth draws scatter plot in scatter plot, and frequency range is the frequency that radio signal is collected
Rate section, bandwidth range obtain the maximum value of the bandwidth in all radio signals for 0 to step 1；According to each signal
Centre frequency and bandwidth value find its position in frequency-bandwidth scatter plot, draw all signaling points；Record each radio
Signal SiCoordinate S in frequency-bandwidth scatter ploti(xi,yi)(1≤i≤n)；It is encoded according to signal strength color, filling letter
Number point；Signal strength color coding refers to the ascending RGB color degree item for being encoded to gradual change of signal strength indication；
Step 3: step 1 radio-signal data obtained being clustered using DBScan clustering algorithm, obtains k cluster
{C1,C2,C3,…,Ck, M noise spot；Specific step is as follows:
Step 3.1: two parameters needed for setting DBScan clustering algorithm: the quantity at least put in radius eps and neighborhood
minpts；It is not visited for marking the original state of all the points；
Step 3.2: calculate step 2 frequency-bandwidth scatter plot Euclidean distance:As
Two signal S in DBScan clustering algorithmi,SjThe distance between (1≤i, j≤n, i ≠ j)；
The specific method of DBScan algorithm: optional one not visited point starts, and finds out all within eps with its distance
Neighbouring point；If quantity >=the minpts nearby put, current point and its one cluster of point formation nearby, and starting point is labeled
To have accessed；Then recurrence handles all points for being not labeled as having accessed in the cluster in the same way, to carry out to cluster
Extension；If quantity < the minpts nearby put, which is temporarily labeled and is used as noise spot；If cluster is fully extended, i.e.,
All the points in cluster are marked as having accessed, and are then gone to handle not visited point with same algorithm again；
Step 3.3: obtain DBScan clustering algorithm as a result, symbiosis at k cluster, i.e., k cluster, M noise spot；
Step 4: divide timeslice:
By entire radio signal acquisition time section [tstart,tend] be divided into length p sections identical, i.e. p timeslice, mark
Remember that j-th of period is tj(1≤j≤p)；
To each cluster CiAll signals in the cluster are divided into corresponding timeslice according to its timestamp by (1≤i≤k)
In, each cluster Ci(1≤i≤k) obtains a timeslice dividing sequence Csti={ sti1,sti2,…,stij,…,stip}(1
≤ i≤k), wherein stij(1≤j≤p) (1≤i≤k) indicates cluster CiTimeslice t is divided into (1≤i≤k)j(1≤j≤p)
The set of interior signal, when the signal number in the set is 0,Otherwise by stijIt is denoted as set { Sq,…,
Sq+r, signal number is r+1 in the set, 0≤r≤n-1, and the timestamp of all signals in the set is in timeslice tj(1≤
J≤p) in, wherein SqAnd Sq+rIt respectively indicates all signaling points in the set and temporally stabs timestamp minimum and maximum after sequence
Signal, q and q+r are respectively signal SqAnd Sq+rIn signal set { S1,S2,…Si,…,SnIn serial number；
Step 5: each cluster C according to obtained in step 4iThe timeslice dividing sequence Cst of (1≤i≤k)i={ sti1,
sti2,…,stip(1≤i≤k), calculate stijThe mean center frequencies of interior all signals, average bandwidth, average signal-to-noise ratio and
Average signal strength, ifThen without calculating；
Step 6: according to the calculated result of step 3 cluster result obtained and step 5 draw signal flow diagram, specifically include with
Lower step:
Step 6.1): drawing signal flow diagram coordinate system, and wherein X-axis indicates that frequency, Y-axis indicate the time, and frequency range is radio
The collected frequency separation of signal data, time range are the period that radio-signal data is collected；Each cluster is drawn
A bars stream is made；
Step 6.2): to cluster Ci(1≤i≤k), according to stijThe value mFreq of the mean center frequency of interior signalij, in signal stream
Corresponding frequency and timeslice t in figure coordinate systemjInitial position graphical pointv A, according to being divided into timeslice tj+1Interior signal is averaged
Centre frequency mFreqi(j+1), corresponding frequency and timeslice t in signal flow diagram coordinate systemj+1Initial position, that is, timeslice tj's
Final position graphical pointv B, tie point A and point B, draw a line segment AB, to determine signal stream corresponding to the cluster in the time
Piece tjCenter；
If timeslice tjWithout subsequent time piece t continuous therewithj+1OrThen timeslice tjCorresponding signal stream is not drawn
System；
Step 6.3): according to the clustering to timeslice tjThe average bandwidth of interior signal determines the width of signal stream, and draws letter
Number stream left margin and right margin, left margin and right margin are symmetrical about line segment AB；
Step 6.4): according to the clustering to timeslice tjThe value and signal strength color of the average signal strength of interior signal are compiled
Code, filling left margin to the region between line segment AB；
Step 6.5): according to the clustering to timeslice tjThe average signal-to-noise ratio and signal-to-noise ratio color of interior signal encode, filling
Line segment AB is to right border area；Signal-to-noise ratio color coding refers to the ascending gradual change gray scale for being encoded to gradual change of snr value
Value；
Step 6.6): it repeats the above steps 6.2) to step 6.5), successively draws Ci(1≤i≤k) all timeslices, and it is all
Cluster.
2. the method for visualizing of radio-signal data according to claim 1, which is characterized in that the step 2 dissipates
In point diagram, a signal is indicated with the circle that a radius is R, according to the signal strength of the signal, is encoded in signal strength color
In find color filling corresponding to its value circle.
3. the method for visualizing of radio-signal data according to claim 2, which is characterized in that the signal strength face
Color encodes concrete methods of realizing are as follows: color mode uses RGB mode, converts the signal strength indication of (0, -140) dbm range to
(0, (5*255)/i) range, i.e., each signal strength indication s (dbm) in (0, -140) range is in (0, (5*255)/i) range pair
A unique value mindex is answered, obtaining mindex, passing through following manner obtains color corresponding to the signal strength indication later
Color (r, g, b):
If mindex >=0 and mindex≤255/i, enables
R=255；
G=map (mindex, 0,255/i, 0,255)；
B=0；
If mindex > 255/i and mindex≤255*2/i, enable
R=map (mindex, 255/i, 255*2/i, 0,255)；
G=255；
B=0；
If mindex > 255*2/i and mindex≤255*3/i, enable
R=0；
G=255
B=map (mindex, 255*2/i, 255*3/i, 0,255)；
If mindex > 3*255/i and mindex≤255*4/i, enable
R=0；
G=map (mindex, 255*3/i, 255*4/i, 0,255)；
B=255；
If mindex > 4*255/i and mindex≤255*5/i, enable
R=0；
G=0；
B=map (mindex, 255*4/i, 255*5/i, 0,255)；
Wherein map (x, x1, x2, x3, x4) function is to return to x of the value in (x1, x2) range to be transformed into (x3, x4) range
Respective value；In this way, make 0 to -140dbm signal strength indication be encoded to by red fade to yellow fade to again it is green
Color fades to cyan again and fades to blue again, finally fades to the coloration item of black.
4. the method for visualizing of radio-signal data according to claim 3, which is characterized in that the signal-to-noise ratio color
Encode concrete methods of realizing are as follows: use greyscale color mode, convert 0 to 255 ranges for the snr value that range is 0 to 100
Color value, i.e., a color value of corresponding (0, the 255) range of each snr value in (0,100) range, which is
It is encoded for the greyscale color of the signal-to-noise ratio；0 to 100 snr value is encoded to gradual change gray scale from white to black.
5. the method for visualizing of radio-signal data according to claim 4, which is characterized in that the step 3 uses
DBScan clustering algorithm clusters radio-signal data, calculates step 2 frequency-bandwidth scatter plot Euclidean distanceAs two signal S in clustering algorithmi,SjThe distance between (1≤i, j≤n, i ≠ j).
6. the method for visualizing of radio-signal data according to claim 5, which is characterized in that in the step 3, make
Radio-signal data is clustered with clustering algorithm, and the color of customized each cluster；Define the color of each cluster
When, the distribution situation of combining wireless electrical signal data signal strength, selection signal intensity coding be not used and with the cluster
The high color of interior signaling point Fill Color discrimination.
7. the method for visualizing of radio-signal data described according to claim 1~any one of 6, which is characterized in that institute
It states in step 4, the time span of the basis of design radio signal acquisition of p and sample rate selection.
8. the method for visualizing of radio-signal data described according to claim 1~any one of 6, which is characterized in that institute
It states in step 5, clusters CiTimeslice t is divided into (1≤i≤k)jThe mean center frequency of all signals in (1≤j≤p),
The calculation of average bandwidth, average signal-to-noise ratio and average signal strength is as follows:
9. the method for visualizing of radio-signal data described according to claim 1~any one of 6, which is characterized in that institute
It states in step 6, two rendering parameters: timeslice signaling point degree of rarefication e and continuous sparse time the piece number z, timeslice signal is set
Point degree of rarefication indicates that the minimum signal number in a timeslice, continuous sparse time the piece number indicate that signal number is less than timeslice
The continuous time the piece number of signal degree of rarefication is less than e if there is the signaling point in the sequential time slices more than or equal to z, that
This segment signal stream is not drawn.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610419842.XA CN105897488B (en) | 2016-06-13 | 2016-06-13 | A kind of method for visualizing of radio-signal data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610419842.XA CN105897488B (en) | 2016-06-13 | 2016-06-13 | A kind of method for visualizing of radio-signal data |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105897488A CN105897488A (en) | 2016-08-24 |
| CN105897488B true CN105897488B (en) | 2019-04-09 |
Family
ID=56729418
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610419842.XA Active CN105897488B (en) | 2016-06-13 | 2016-06-13 | A kind of method for visualizing of radio-signal data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105897488B (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106911407A (en) * | 2017-02-08 | 2017-06-30 | 广州新拓慧电子科技有限公司 | Spectrum signature visualization device and method for visualizing |
| CN108197647B (en) * | 2017-12-28 | 2021-08-31 | 中南大学 | Rapid clustering method for automobile starter endurance test data |
| CN109389172B (en) * | 2018-10-11 | 2022-05-20 | 中南大学 | Radio signal data clustering method based on non-parameter grid |
| CN109447163B (en) * | 2018-11-01 | 2022-03-22 | 中南大学 | Radar signal data-oriented moving object detection method |
| CN112016180B (en) * | 2019-05-28 | 2023-12-15 | 浙江德盛铁路器材股份有限公司 | Quality condition analysis method for railway track foundation equipment |
| CN113225145B (en) * | 2021-05-08 | 2022-01-28 | 中南大学 | Radio signal multi-dimensional time-varying feature visualization method |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103267905A (en) * | 2013-04-25 | 2013-08-28 | 北京航空航天大学 | Electromagnetic background noise extraction method based on hierarchy-clustering algorithm |
| CN103812577A (en) * | 2012-11-06 | 2014-05-21 | 西南交通大学 | Method for automatically identifying and learning abnormal radio signal type |
-
2016
- 2016-06-13 CN CN201610419842.XA patent/CN105897488B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103812577A (en) * | 2012-11-06 | 2014-05-21 | 西南交通大学 | Method for automatically identifying and learning abnormal radio signal type |
| CN103267905A (en) * | 2013-04-25 | 2013-08-28 | 北京航空航天大学 | Electromagnetic background noise extraction method based on hierarchy-clustering algorithm |
Non-Patent Citations (2)
| Title |
|---|
| 基于维度扩展的Radviz可视化聚类分析方法;周芳芳 等;《软件学报》;20160531;第27卷(第5期);第1127-1139页 |
| 无线电电磁环境可视化研究;王新民 等;《管理信息化》;20130531;第62-63页 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105897488A (en) | 2016-08-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105897488B (en) | A kind of method for visualizing of radio-signal data | |
| CN104158611A (en) | System and method of detecting interference of wireless signal based on spectral analysis | |
| CN103473550B (en) | Based on the leaf image dividing method of Lab space and local dynamic threshold | |
| CN109947238B (en) | Non-cooperative gesture recognition method based on WIFI | |
| CN108171973B (en) | Traffic travel mode identification method based on mobile phone grid data | |
| CN110033132A (en) | Tropical cyclone forecasting procedure based on depth targets detection and numerical weather forecast | |
| CN102571230A (en) | Distributed collaborative signal identification method based on blind estimation of higher order statistics and signal to noise ratio | |
| CN104660356B (en) | A kind of collaborative frequency spectrum sensing method of half-blindness with reliable false-alarm performance | |
| CN109474355A (en) | Adaptive noise THRESHOLD ESTIMATION and method for extracting signal based on spectrum monitoring data | |
| CN109284345B (en) | Electromagnetic spectrum display method, storage medium, terminal and system | |
| CN106879017A (en) | A kind of decision method and device of LTE network covering quality | |
| CN102882820A (en) | Digital modulation signal identifying method under non-gaussian noise in cognitive radio | |
| CN105916202A (en) | Probabilistic WiFi indoor positioning fingerprint database construction method | |
| CN102271022B (en) | Spectrum sensing method based on maximum generalized characteristic value | |
| CN105785437B (en) | A kind of method and device of the automatic discrimination in Unusual Seismic road | |
| CN105468922B (en) | A kind of live electromagnetic environment Signal online analysis and intimidation estimating method | |
| CN110460962A (en) | Rail traffic user identification method and device | |
| Zuo et al. | Detection and Analysis of Urban Area Hotspots Based on Cell Phone Traffic. | |
| CN104602271B (en) | The generation method and device of problem area in a kind of coverage optimization | |
| Zhou et al. | Computing the color complexity of images | |
| CN113225145B (en) | Radio signal multi-dimensional time-varying feature visualization method | |
| CN104022839B (en) | A kind of information fusion decision method being applied to cooperative spectrum sensing | |
| CN103036626A (en) | Wireless communication method based on cognitive radio cooperation users and threshold testing combined selection | |
| CN107889210A (en) | Building method for positioning user and system | |
| CN102711154B (en) | Regional analysis method for GSM downlink interference |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |