CN109613477A - A kind of TDOA location tracking method under complex environment - Google Patents
A kind of TDOA location tracking method under complex environment Download PDFInfo
- Publication number
- CN109613477A CN109613477A CN201910026517.0A CN201910026517A CN109613477A CN 109613477 A CN109613477 A CN 109613477A CN 201910026517 A CN201910026517 A CN 201910026517A CN 109613477 A CN109613477 A CN 109613477A
- Authority
- CN
- China
- Prior art keywords
- data
- base station
- array
- coordinate
- tdoa
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 60
- 238000004364 calculation method Methods 0.000 claims abstract description 9
- 238000004422 calculation algorithm Methods 0.000 claims description 15
- 239000011159 matrix material Substances 0.000 claims description 13
- 238000001914 filtration Methods 0.000 claims description 7
- 230000008569 process Effects 0.000 claims description 7
- 238000012216 screening Methods 0.000 claims description 4
- 238000006467 substitution reaction Methods 0.000 claims description 4
- 239000013598 vector Substances 0.000 claims description 4
- 230000008859 change Effects 0.000 claims description 3
- 238000012545 processing Methods 0.000 claims description 3
- 230000017105 transposition Effects 0.000 claims description 2
- 238000012360 testing method Methods 0.000 description 7
- 230000008901 benefit Effects 0.000 description 6
- 230000003247 decreasing effect Effects 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 230000033001 locomotion Effects 0.000 description 3
- 230000009466 transformation Effects 0.000 description 3
- 230000002159 abnormal effect Effects 0.000 description 2
- 230000001419 dependent effect Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 241001269238 Data Species 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000010924 continuous production Methods 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 238000012217 deletion Methods 0.000 description 1
- 230000037430 deletion Effects 0.000 description 1
- 238000009795 derivation Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 239000011521 glass Substances 0.000 description 1
- 230000004807 localization Effects 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 238000010187 selection method Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S5/00—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations
- G01S5/02—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations using radio waves
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S5/00—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations
- G01S5/02—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations using radio waves
- G01S5/0294—Trajectory determination or predictive filtering, e.g. target tracking or Kalman filtering
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Mobile Radio Communication Systems (AREA)
- Position Fixing By Use Of Radio Waves (AREA)
Abstract
The invention discloses the TDOA location tracking methods under a kind of complex environment, comprising: step 1, reads TDOA data；Step 2, data prediction is carried out, the biggish base station of error is removed；Step 3, remaining data of effective base station number more than or equal to 4 are resolved, calculation result is put into array P；Step 4, master base station is replaced, step 3 is repeated, until all base stations in the data have all served as a master base station；Step 5, if P is not sky, most suitable point is therefrom selected to be put into the array T for saving final positioning result, and empty array P；Step 6, if joined new point in T, average filter is carried out, filter result is put into track array F；Step 7, if being predicted using time and speed changing coordinates according to the data in F that new point is not added in T, prediction result being put into F and T simultaneously；Step 8, which is disposed, and returns to step 1, reads next data.
Description
Technical field
The present invention relates to indoor orientation methods, more particularly to the TDOA location tracking method under a kind of complex environment.
Background technique
With the continuous development of Internet of Things and mobile interchange technology, the position of more and more services and application dependent on user
How confidence breath is accurately positioned the big hot spot for just becoming current research to equipment.In an outdoor environment, global location
System (GPS) is higher with its precision, the strong and low-cost advantage of stability, achieves and is widely applied, but environment indoors
Under, since satellite-signal can not penetrate building, the use of GPS receives serious limitation.Moreover, compared to outdoor, it is indoor
The problems such as environment is influenced to become apparent by multipath effect, non line of sight transmission (NLOS), and there is also the dynamic changes of environment, this
The challenge of indoor accurate position is both increased a bit.Currently, domestic and international researchers have been proposed based on signal attenuation model,
TOA (Time-of-Arrival), TDOA (Time-Difference-of-Arrival) and AOA (Angle-of-Arrival)
Etc. principles indoor positioning solution, wherein TDOA does not need the advantage of strict time synchronization with it, has obtained extensive weight
Depending on and research.
In numerous TDOA localization methods, method (abbreviation Taylor sequence of the Wade H.Foy according to Taylor expansion proposition
Column method) simple, higher advantage of precision is achieved and is widely applied in the form of it.This method near given initial coordinate into
Row Taylor is unfolded and ignores the above component of second order, then by iterating to calculate the local least square method solution of error come successive optimization
Coordinate.Initial value of the precision of this method dependent on input, result is more accurate if initial value is close to true solution, but in reality
In the application of border, initial value is difficult to choose, and causes its precision poor, in some instances it may even be possible to can not restrain.And the characteristic of iteration needs it
Want biggish calculation amount.Document: Foy W H.Position-Location Solutions by Taylor-Series
Estimation[J].IEEE Transactions on Aerospace&Electronic Systems,2007,AES-12
(2):187-194.
In order to overcome based on Taylor expansion method existing for dependence initial value be iterated, not can guarantee convergence and
The big problem of time overhead, Y.T.Chan et al. propose the hyperbolic location algorithm (abbreviation Chan algorithm) with closed solutions.It should
Algorithm converts linear problem for nonlinear problem, then passes through weighted least-squares twice by introducing a variable of third
Method obtains final positioning result.Chan algorithm calculating speed is fast, as a result also more accurate, but the derivation of the algorithm is based on channel
Error is smaller and obeys that zero-mean gaussian is distributed it is assumed that the hypothesis is difficult to meet under true environment, therefore precision can be significant
Decline.Document: Chan Y T, Ho K C.A simple and efficient estimator for hyperbolic
location[J].IEEE Transactions on Signal Processing,2002,42(8):1905-1915.
Summary of the invention
Goal of the invention: classics TDOA location algorithm property under true environment such as Taylor sequence method, Chan algorithm are solved
It can be remarkably decreased, the problem of continuously smooth track can not be formed in tracing process, introduce and fix in Taylor and Chan algorithm
Elevation information, promotes the accuracy of positioning result, and the priori knowledge of data extending technology and coordinate is added, and being greatly decreased can not
The case where resolving, while by reasonably filtering and predicting strategy, enhance the slickness and continuity of track.It is above-mentioned in order to solve
Technical problem, the invention discloses the location tracking method based on TDOA under a kind of complex indoor environment, this method can be used for
In the applications such as storehouse management, locating guide, robot tracking, include the following steps:
Step 1, the TDOA data containing N number of (General N value is 4) above base station are read；
Step 2, data prediction is carried out according to historical data, removes the biggish base station of error；
Step 3, the data of N are more than or equal to for remaining effective base station number, using the Chan that elevation information is added and
Taylor sequence method is resolved, and calculation result is put into array P；
Step 4, master base station is replaced, step 3 is repeated, until all base stations in this TDOA data have all served as primary main base
It stands；
Step 5, if array P is not sky, most suitable point is therefrom selected to be put into the array T for saving final positioning result,
And empty array P；
Step 6, if joined new point in array T, average filter is carried out, filter result is put into track array F, is used for
Show real-time track；
Step 7, if that new point is not added in array T, according to the data in the array F of track using time and speed to working as
Preceding coordinate is predicted, prediction result is put into track array F and array T simultaneously；
Step 8, this TDOA data processing finishes, and returns to step 1, reads next TDOA data.
In step 1, the form of the TDOA data data of reading are as follows:
Data=(di1, di2, di3 ..., din),
Wherein n indicates the number of base station in this TDOA data, and i indicates master base station serial number, and 1≤i≤n, dij expression are worked as
Distance of the preceding coordinate to be asked to master base station and the base station to serial number j (difference of the distance of abbreviation base station j), 1≤j≤n, especially
Ground, as j=i, dij=0.The step requires base station number more than or equal to 4, i.e. and n >=4.
In step 2, the current difference DELTA data read between TDOA data and historical data is set are as follows:
Δ data=(Δ di1, Δ di2, Δ di3 ..., Δ din),
Wherein Δ dij is dij and the difference between moment preferable historical data before.Error threshold is set as t, if Δ
Dij > t just deletes dij from data.
Step 3 includes the following steps:
Step 3-1 carries out rough estimate to coordinate with the Chan algorithm for introducing height.TDOA data are carried out first simple
Transformation, make master base station become serial number 1 base station.Transformed data data_ form are as follows:
Data_=(d12, d13, d14 ..., d1n),
Wherein difference of d1j (2≤j≤n) the expression changing coordinates to distance between master base station and base station j.Set base station
Coordinate be (x1, y1, z1), (x2, y2, z2) ..., (xn, yn, zn), wherein (xi, yi, zi) indicates that the three-dimensional of base station i is sat
It marks, 1≤i≤n, known to the coordinate.The fixed height of the object to be positioned introduced is h, then improved Chan algorithm can be with
It indicates are as follows:
Ki=xi^2+yi^2+ (zi-h) ^2,
E=0.5* [d12^2+K1-K2；d13^2+K1-K3；…；D1n^2+K1-Kn],
Xij=xi-xj,
Yij=yi-yj,
G=[- x21 ,-y21, d12；-x31,-y31,d13；…；- xn1 ,-yn1, d1n],
Z_=inv (GT*inv(Q)*G)*GT* inv (Q) * E,
Ri=sqrt ((xi-Z_ (1)) ^2+ (yi-Z_ (2)) ^2+ (zi-h) ^2),
D=diag { r2, r3 ..., rn },
X=D*Q*D,
Z=inv (GT*inv(X)*G)*GT* inv (X) * E,
Wherein Ki, E, xij, yij, G, Z_, ri, D, X indicate the intermediate quantity introduced；^2 indicates that square operation, inv indicate square
The inversion operation of battle array, the T representing matrix transposition in the matrix upper right corner, Q indicate that the covariance matrix of error, sqrt indicate evolution fortune
Calculate, Z_ (i) indicate vector Z _ i-th of component (i=1,2,3), diag { r2, r3 ..., rn } indicates with r2, r3 ..., rn according to
The secondary diagonal matrix for diagonal entry.The Z finally acquired be containing there are three the vector of element, can be expressed as Z=(x, y,
R), wherein (x, y) i.e. required coordinate estimated value, r are distance of the coordinate to master base station.
The result that step 3-1 is obtained is iterated as the substitution of the initial value of Talor sequence method, obtains by step 3-2
More accurate coordinate.The step for equally introduce fixed height information, setting height h, initial value be (x0, y0), then improve
Taylor sequence method afterwards can indicate are as follows:
Ri=sqrt ((xi-x0) ^2+ (yi-y0) ^2+ (zi-h) ^2),
Rij=Ri-Rj,
H=[R21- (R2-R1)；R31-(R3-R1)；…；Rn1- (Rn-R1)],
Pi=(xi-x0)/Ri,
Qi=(yi-y0)/Ri,
M=[p1-p2, q1-q2；p1-p3,q1-q3；…；P1-pn, q1-qn],
[Δ x, Δ y]=inv (MT*inv(Q)*M)*MT* inv (Q) * H,
Wherein, Ri, Rij, H, pi, qi, M indicate the intermediate quantity introduced；Q is identical with step 1, is the covariance of error
Matrix；(Δ x, Δ y) are the grid deviation amounts found out in epicycle iteration, judge whether iteration restrains with it.If threshold parameter is
T, convergent Rule of judgment are as follows:
| Δ x |+| Δ y | < t,
If epicycle iteration does not restrain, with (x0+ Δ x, y0+ Δ y) replaces (x0, y0), repeats the above process, until
Convergence or the number of iterations reach the pre-set upper limit.If convergence, final calculated result is put into array P；If iteration time
Taylor sequence method does not still restrain number after reaching the upper limit, goes to step 3-3.
Step 3-3, if the Taylor iteration in step 3-2 does not restrain, with the array T for saving final positioning result
In the last one point (positioning is a continuous process, and array T is used to store the final positioning result of last time, wherein
The last one point be last moment positioning result, that is, the result that calculates of a upper TDOA data), i.e., it is the last at
Function resolve and pass through screening result as initial value, substitute into Taylor sequence method and be iterated, the convergence judgement of iteration and
Step 3-2 is identical, if convergence, is put into array P for result, is otherwise judged to resolving failure.
In step 4, it is known that the form of initial data be data=(di1, di2, di3 ..., din), it using base station i as
Master base station, to replace base station i to become master base station with base station j, then the data data ' after replacing master base station are as follows:
Data '=(di1-dij, di2-dij ..., din-dij).
In step 5, setting k-th of coordinate in array P, as Pk, current master base station is j, pushes away its TDOA number according to Pk is counter first
According to inversion formula are as follows:
Ui=sqrt ((xi-Pk (1)) ^2+ (yi-Pk (2)) ^2+ (zi-h) ^2),
Datak=(uj-u1, uj-u2 ..., uj-un),
Wherein i-th of component of Pk (i) indicates coordinate Pk, i value are 1 or 2, and ui indicates the distance between Pk to base station i,
Uj indicates that the distance between Pk to base station j, un indicate that the distance between Pk to base station n, h are identical with preceding step fixed high
Spend parameter.It is counter pushed away data datak after, calculate square of its Euclidean distance between initial data data, calculate public
Formula are as follows:
Wi=data (i)-datak (i),
Distk=w1^2+w2^2+ ...+wn^2,
Wherein wi indicates initial data data and the anti-difference for pushing away data datak on i-th of component, and data (i) is indicated
I-th of component of data, datak (i) indicate datak i-th of component, by above-mentioned formula calculating (dist1, dist2 ...,
Distm), m be array P in element number, take wherein the corresponding position of minimum value as a result, element in P is put, as this
The final positioning result of TDOA data, is put into array T.
In step 6, the window size of average filter is set as span, is set element number in array T and is then filtered public affairs as k
Formula are as follows:
F=(T (1)+T (2)+...+T (k))/k (k < span),
F=(T (k)+T (k-1)+...+T (k-span+1))/span (k >=span),
Wherein T (i) indicates that i-th of element of array T, 1≤i≤k, f are to obtain coordinate after filtering is handled, will
Array F is added in it.
In step 7, the size of prediction window is set as span, sets the element number in F as k, then the prediction of coordinate is public
Formula are as follows:
Si=sqrt ((F (k-i+1) (1)-F (k-i) (1)) ^2+ (F (k-i+1) (2)-F (k-i) (2)) ^2)
Vi=si/ti,
V=(v1+v2+ ...+vk)/k (k < span),
V=(v1+v2+ ...+vspan)/span (k >=span),
S=v*t,
G=(F (k) (2)-F (k-1) (2))/(F (k) (1)-F (k-1) (1)),
Δ x_=s/sqrt (1+g^2),
Δ y_=s/sqrt ((g^2+1)/g^2),
X=F (k) (1)+sign (F (k) (1)-F (k-1) (1)) * Δ x,
Y=F (k) (2)+sign (F (k) (2)-F (k-1) (2)) * Δ y,
Wherein si, vi, v, s, g, Δ x_, Δ y_ indicate the intermediate quantity introduced, and F (i) indicates i-th of element in array F,
1≤i≤k, F (i) (j) indicate j-th of component of i-th of element in array F, j value 1 or 2, ti indicate the i-th close point and
Time interval between the close point of i+1, t indicate the time interval in current data and F between last 1 point, and sign is symbol
Number function, positive number 1, negative are -1；(x, y) finally acquired is to predict coordinate.
In step 8, this TDOA data have been disposed, and final positioning result is the last one coordinate in array T, filter
Wave result is the last one coordinate in array F, and array F is for showing real-time track.Need to read next TDOA data at this time,
Handle the information of subsequent time.
The utility model has the advantages that remarkable advantage of the invention is under complicated true environment, such as by serious reflection and screening
When gear interference, can be greatly decreased in TDOA positioning because data are of poor quality occur the case where can not resolving, and reach higher
Precision guarantee the continuity and slickness of track, be obviously improved the performance of positioning system while in target tracking.
Detailed description of the invention
The present invention is done with reference to the accompanying drawings and detailed description and is further illustrated, it is of the invention above-mentioned or
Otherwise advantage will become apparent.
Fig. 1 is overall flow figure of the invention.
Fig. 2 is the flow chart that Chan-Taylor is resolved in the present invention.
It is TDOA data sequence calculated track through the invention shown in Fig. 3 a.
Fig. 3 b is the calculated scatter plot of TDOA data Taylor sequence method.
Specific embodiment
The present invention will be further described with reference to the accompanying drawings and embodiments.
Fig. 1 is overall flow figure of the invention, includes 8 steps.
In step 1, therefore, to assure that it reads in the base station number that data include and is more than or equal to 4, it is otherwise very little containing information, it is subsequent
Process of solution will be unable to carry out.The form of the TDOA data of reading is data=(di1, di2, di3 ..., din), and dij expression is worked as
For preceding coordinate to be asked to the distance of master base station and the difference of the distance to base station j, n indicates base station number, the step require n >=4.
In step 2, needs to delete the obvious base station bigger than normal of some errors, influence subsequent solution to avoid these abnormal datas
Step is calculated, the quality of data is promoted.Processing method be take the nearest preferable historical data of moment quality first, and by the data with
Current TDOA data compare, and calculate the deviation delta dij on each component.Assuming that preset error threshold be t, then when
When Δ dij > t, the component is deleted from TDOA data.Threshold value t is usually set to twice of target moving distance, wherein it is mobile away from
From can be estimated according to target movement speed and time.For example movement speed is set as 1m/s, current time and last moment
Time difference be 0.2s, then t is set as 0.4m.In order to avoid the non-abnormal data of deletion, increase the robustness of positioning system,
Threshold value t can suitably be expanded under above-mentioned calculation.
In step 3, the number of base station in data is first determined whether, because step 2 may will be deleted several errors in data
Component corresponding to larger base station.If remaining base station number is less than 4, it is meant that subsequent resolving can not be carried out, need to return to step
Rapid 1, next data is read, Chan-Taylor resolving is otherwise carried out.The flow chart of this solution process is shown in Fig. 2, and it includes such as
Lower step:
Step 3-1 carries out rough estimate to coordinate with the Chan algorithm for introducing height.TDOA data are done simply first
Transformation makes master base station become the base station of serial number 1, and the transformation is convenient only for subsequent expression, will not change changing coordinates and arrive
The difference of each base station distance does not influence calculated result.Transformed data mode are as follows:
Data_=(d12, d13, d14 ..., d1n),
Wherein difference of d1j (2≤j≤n) the expression changing coordinates to distance between master base station and base station j.Assuming that base station
Coordinate be (x1, y1, z1), (x2, y2, z2) ..., (xn, yn, zn), wherein (xi, yi, zi) indicates that the three-dimensional of base station i is sat
It marks, 1≤i≤n, known to the coordinate.The fixed height of the object to be positioned introduced is h, then improved Chan algorithm can be with
It indicates are as follows:
Ki=xi^2+yi^2+ (zi-h) ^2,
E=0.5* [d12^2+K1-K2；d13^2+K1-K3；…；D1n^2+K1-Kn],
Xij=xi-xj,
Yij=yi-yj,
G=[- x21 ,-y21, d12；-x31,-y31,d13；…；- xn1 ,-yn1, d1n],
Z_=inv (GT*inv(Q)*G)*GT* inv (Q) * E,
Ri=sqrt ((xi-Z_ (1)) ^2+ (yi-Z_ (2)) ^2+ (zi-h) ^2),
D=diag { r2, r3 ..., rn },
X=D*Q*D,
Z=inv (GT*inv(X)*G)*GT* inv (X) * E,
Above-mentioned formula acquires Z=(x, y, r), wherein (x, y) i.e. required coordinate estimated value, r is the coordinate to master base station
Distance.The process has only carried out the first time weighted least square in Chan algorithm, can also using x, y and r it
Between relationship do second of weighted least square, but the test under true environment shows in most cases for the first time
It is accurate enough to estimate resulting result, then carries out second of estimation and not only will increase calculation amount, in some instances it may even be possible to make under precision
Drop, therefore the step requires only to carry out a least-squares estimation.
The result that step 3-1 is obtained is iterated as the substitution of the initial value of Talor sequence method, obtains by step 3-2
More accurate coordinate.The step for equally introduce fixed height information, it is assumed that be highly h, initial value be (x0, y0), then improve
Taylor sequence method afterwards can indicate are as follows:
Ri=sqrt ((xi-x0) ^2+ (yi-y0) ^2+ (zi-h) ^2),
Rij=Ri-Rj,
H=[R21- (R2-R1)；R31-(R3-R1)；…；Rn1- (Rn-R1)],
Pi=(xi-x0)/Ri,
Qi=(yi-y0)/Ri,
M=[p1-p2, q1-q2；p1-p3,q1-q3；…；P1-pn, q1-qn],
[Δ x, Δ y]=inv (MT*inv(Q)*M)*MT* inv (Q) * H,
Wherein Q is identical with step 1, is the covariance matrix of error；(Δ x, Δ y) are the seats found out in epicycle iteration
Departure is marked, if threshold parameter is t, t is usually set to 1cm, restrains if Δ x+ Δ y < t, otherwise uses (x0+ Δ x, y0+ Δ
Y) replace (x0, y0), repeat the above process, until convergence or the number of iterations reach the pre-set upper limit.If convergence, will be final
Calculated result is put into array P；If Taylor sequence method does not still restrain the number of iterations after reaching the upper limit, 3- is gone to step
3.Test under true environment shows that in the higher situation of the quality of data, Taylor sequence method generally only needs to change for 1 to 2 times
In generation, can restrain.
Step 3-3 is successfully resolved and is passed through with the last time if the Taylor iteration in step 3-2 is not restrained
The result of screening substitutes into Taylor sequence method and is iterated as initial value, and the convergence judgement of iteration is identical as step 3-2,
If convergence, is put into P for result, otherwise it is assumed that resolving failure.In the case where the quality of data is poor, which is able to ascend
A possibility that Taylor sequence method restrains, increases the success rate of resolving.
In step 4, it is known that the form of initial data be data=(di1, di2, di3 ..., din), it using base station i as
Master base station to replace base station i to become master base station with base station j, that is, calculates (dj1, dj2 ..., djn), can be by:
Djk=dj-dk=(di-dk)-(di-dj)=dik-dij,
Obtaining the replacement base station j is the data after master base station:
Data '=(di1-dij, di2-dij ..., din-dij),
Replacement master base station can effectively expand original TDOA data, and in the case where there is n base station, a TDOA data can
To be converted into n item, and successively resolved with the Chan-Taylor method in step 3, increase calculation result quantity and can
By property.The main purpose of the step is to eliminate accidental error, and handles the case where original master base station is blocked.
In step 5, by replacement base station and after resolving, the number of the point in array P is likely larger than 1 (be up to n),
Need to select positioning result of the point the most suitable as this TDOA data.Selection method is reversely to be calculated by coordinate in P
They answer corresponding TDOA data out, and calculate these and calculate the flat of the Euclidean distance between data and this TDOA data
Side, takes element in the wherein corresponding P of minimum value to be put into array T as final positioning result.Selection needs to empty number after completing
Group P, to store the calculation result of next TDOA data.
In step 6, the window size of average filter is set as span, it is assumed that element number is k in T, then filtering strategies are
As k < span, the average value of all elements in T is taken as a result, otherwise taking the average value of span nearest element.Window
Bigger, smooth effect is more obvious, but can also lose more information simultaneously.
In step 7, the size of prediction window is set as span, it is assumed that the element number in F is k, then the prediction side of coordinate
Method is the average speed for calculating nearest a period of time first, is the average speed between nearest k point as k<span, work as k>=
It is the average speed between nearest span point when span, then with the speed multiplied by a upper point to the time between the moment
Interval, the moving distance predicted.The distance is intercepted on the extended line of nearest two lines, what is obtained is prediction
Coordinate.
In step 8, this TDOA data have been disposed, and final positioning result is the last one coordinate in T, filtering knot
Fruit is the last one coordinate in F, and F has smoother characteristic relative to T, can be used for showing real-time track.It needs to read at this time
A TDOA data are removed, the information of subsequent time is handled, to form continuous path.
Embodiment
In order to verify the validity of proposition method, place is disposed in actual environment and is tested.Test site is one
There is one block of glass in the room of 5m*7m or so, the room upper left corner, and neighbouring signal can occur significantly to reflect.Place surrounding is total to portion
8 base stations are affixed one's name to, the height of base station is probably in 3m or so.Tester walks several weeks along edge in the venue, motion profile
Close to rectangle.This TDOA data sequence acquired is calculated in the present invention as test data, wherein each step
Realization and parameter detail it is as follows:
Step 1, the TDOA data containing 4 and the above base station are read；
Step 2, the biggish base station of error is removed according to historical data, the threshold value t of error is set as 1m in the step；
Step 3, Chan-Taylor resolving is carried out, the fixed height h in Chan algorithm and Taylor sequence method is set as
It is 100 that 2m (fixed height of object to be positioned), error co-variance matrix Q, which are disposed as diagonal line, pair that remaining element is 500
Claim matrix.For judging that convergent threshold value t is set as 1cm in Taylor sequence method, the iteration upper limit is set as 25.
Step 4, it replaces master base station and repeats step 3；
Step 5, it if P is not sky, therefrom selects most suitable point and empties array P；
Step 6, if joined new point in T, average filter is carried out, the window size span of filtering is set as 25；
Step 7, if that new point is not added in T, changing coordinates are carried out in advance using time and speed according to the data in F
It surveys, the window size span of prediction is set as 5；
Step 8, which is disposed, and returns to step 1, reads next data.
It is this TDOA data sequence calculated track through the invention shown in Fig. 3 a, Fig. 3 b is data Taylor
The calculated scatter plot of sequence method.It is not difficult to find out that the calculated point of Taylor method is more dispersed, if connected in chronological order
It connects to form trajectory diagram, stability is poor, and there are serious hopping phenomenons.The especially upper left corner, glass-reflected lead to this part number
According to second-rate, Taylor sequence method calculated result error in these data is significantly increased, therefore does not have in reality
The ability used under the scene of border.And the track that the present invention obtains is then very close to real trace, track remain it is continuous and compared with
To be smooth, positioning accuracy is also relatively high, in 10~20cm or so.The data in the upper left corner through the invention in proposition optimization plan
Slightly, there is not target deviation and Loss, convincingly demonstrated the validity of method.
Above-mentioned test be only it is representative primary in repeatedly test, other environment, track under test can
Similar effect is obtained, the precision of positioning and the smooth continuity of track are significantly better than conventional method.
The present invention provides the TDOA location tracking method under a kind of complex environment, the method for implementing the technical solution
It is many with approach, the above is only a preferred embodiment of the present invention, it is noted that for the ordinary skill of the art
For personnel, various improvements and modifications may be made without departing from the principle of the present invention, these improvements and modifications
It should be regarded as protection scope of the present invention.All undefined components in this embodiment can be implemented in the prior art.
Claims (9)
1. the TDOA location tracking method under a kind of complex environment, which comprises the steps of:
Step 1, the TDOA data for containing N number of above base station are read；
Step 2, data prediction is carried out, the biggish base station of error is removed, remaining base station is effective base station；
Step 3, remaining data of effective base station number more than or equal to N are resolved, calculation result is put into array P；
Step 4, master base station is replaced, step 3 is repeated, until all base stations in this TDOA data have all served as a master base station；
Step 5, if array P is not sky, most suitable point is therefrom selected to be put into the array T for saving final positioning result, and clear
Empty array P；
Step 6, if joined new point in array T, average filter is carried out, filter result is put into track array F, for showing
Real-time track；
Step 7, it if that new point is not added in array T, is sat using time and speed to current according to the data in the array F of track
Mark is predicted, prediction result is put into track array F and array T simultaneously；
Step 8, this TDOA data processing finishes, and returns to step 1, reads next TDOA data.
2. the method according to claim 1, wherein in step 1, the form of the TDOA data data of reading are as follows:
Data=(di1, di2, di3 ..., din),
Wherein n indicates the number of base station in this TDOA data, and i indicates master base station serial number, 1≤i≤n, dij indicate currently to
Seek the difference of distance of the coordinate to the distance of master base station and to the base station of serial number j, 1≤j≤n, as j=i, dij=0.
3. according to the method described in claim 2, it is characterized in that, setting current reading TDOA data and history number in step 2
Difference DELTA data between are as follows:
Δ data=(Δ di1, Δ di2, Δ di3 ..., Δ din),
Wherein Δ dij is dij and the difference between moment preferable historical data before, sets error threshold as t, if Δ dij >
T just deletes dij from data.
4. according to the method described in claim 3, it is characterized in that, step 3 includes the following steps:
Step 3-1 converts TDOA data, and master base station is made to become the base station of serial number 1, transformed data data_ shape
Formula are as follows:
Data_=(d12, d13, d14 ..., d1n),
Wherein d1j indicates changing coordinates to the difference of distance between master base station and base station j, and 2≤j≤n sets the coordinate of base station
For (x1, y1, z1), (x2, y2, z2) ..., (xn, yn, zn), wherein the three-dimensional coordinate of (xi, yi, zi) expression base station i, 1≤i
≤ n, known to the coordinate；Using the Chan algorithm for introducing height, the fixed height of the object to be positioned of introducing is h, then after improving
Chan algorithmic notation are as follows:
Ki=xi^2+yi^2+ (zi-h) ^2,
E=0.5* [d12^2+K1-K2；d13^2+K1-K3；…；D1n^2+K1-Kn],
Xij=xi-xj,
Yij=yi-yj,
G=[- x21 ,-y21, d12；-x31,-y31,d13；…；- xn1 ,-yn1, d1n],
Z_=inv (GT*inv(Q)*G)*GT* inv (Q) * E,
Ri=sqrt ((xi-Z_ (1)) ^2+ (yi-Z_ (2)) ^2+ (zi-h) ^2),
D=diag { r2, r3 ..., rn },
X=D*Q*D,
Z=inv (GT*inv(X)*G)*GT* inv (X) * E,
Wherein Ki, E, xij, yij, G, Z_, ri, D, X indicate the intermediate quantity introduced；^2 indicates square operation, inv representing matrix
Inversion operation, the T representing matrix transposition in the matrix upper right corner, Q indicate that the covariance matrix of error, sqrt indicate extracting operation, Z_
(i) indicate vector Z _ i-th of component, diag { r2, r3 ..., rn } indicates with r2, r3 ..., rn to be followed successively by diagonal entry
Diagonal matrix；The Z finally acquired be expressed as Z=(x, y, r) containing there are three the vectors of element, wherein (x, y) i.e. required by seat
Estimated value is marked, r is distance of the coordinate to master base station；
The result that step 3-1 is obtained is iterated as the substitution of the initial value of Talor sequence method, obtains more smart by step 3-2
True coordinate: introducing fixed height information, setting height h, and coordinate initial value is (x0, y0), then improved Taylor sequence
Column method indicates are as follows:
Ri=sqrt ((xi-x0) ^2+ (yi-y0) ^2+ (zi-h) ^2),
Rij=Ri-Rj,
H=[R21- (R2-R1)；R31-(R3-R1)；…；Rn1- (Rn-R1)],
Pi=(xi-x0)/Ri,
Qi=(yi-y0)/Ri,
M=[p1-p2, q1-q2；p1-p3,q1-q3；…；P1-pn, q1-qn],
[Δ x, Δ y]=inv (MT*inv(Q)*M)*MT* inv (Q) * H,
Wherein, Ri, Rij, H, pi, qi, M indicate the intermediate quantity introduced；(Δ x, Δ y) are the grid deviations found out in epicycle iteration
Amount, judges whether iteration restrains with it；If threshold parameter is t, convergent Rule of judgment are as follows:
| Δ x |+| Δ y | < t,
If epicycle iteration does not restrain, with (x0+ Δ x, y0+ Δ y) replaces (x0, y0), repeats the above process, until convergence
Or the number of iterations reaches the pre-set upper limit；If convergence, final calculated result is put into array P；If the number of iterations reaches
Taylor sequence method does not still restrain after to the upper limit, goes to step 3-3；
Step 3-3, if the Taylor iteration in step 3-2 does not restrain, in the array T for saving final positioning result
The last one point, i.e., it is the last successfully to resolve and initial value is used as by the result of screening, substitution Taylor sequence method into
The convergence judgement of row iteration, iteration is identical as step 3-2, if convergence, puts array for result and enter P, be otherwise judged to resolving
Failure.
5. method as claimed in claim 4, which is characterized in that in step 4, it is known that the form of initial data data is data=
(di1, di2, di3 ..., din), it is using base station i as master base station, to replace base station i to become master base station with base station j, then more
Data data ' behind change owner base station are as follows:
Data '=(di1-dij, di2-dij ..., din-dij).
6. according to the method described in claim 5, it is characterized in that, in step 5, k-th of coordinate is set in array P as Pk, when
Preceding master base station is j, pushes away its TDOA data datak, inversion formula according to Pk is counter are as follows:
Ui=sqrt ((xi-Pk (1)) ^2+ (yi-Pk (2)) ^2+ (zi-h) ^2),
Datak=(uj-u1, uj-u2 ..., uj-un),
Wherein i-th of component of Pk (i) indicates coordinate Pk, i value are 1 or 2, and ui indicates the distance between Pk to base station i, uj table
Show the distance between Pk to base station j, un indicates the distance between Pk to base station n, it is counter pushed away data datak after, calculate it
Square of Euclidean distance, calculation formula between initial data data are as follows:
Wi=data (i)-datak (i),
Distk=w1^2+w2^2+ ...+wn^2,
Wherein wi indicates initial data data and the anti-difference for pushing away data datak on i-th of component, and data (i) indicates data
I-th of component, datak (i) indicate datak i-th of component, by above-mentioned formula calculating (dist1, dist2 ...,
Distm), m is the number of element in array P, element in the wherein corresponding array P of minimum value is taken, as this TDOA number
According to final positioning result, be put into array T.
7. according to the method described in claim 6, it is characterized in that, in step 6, set the window size of average filter as
Span, sets in array T element number as k, then Filtering Formula are as follows:
F=(T (1)+T (2)+...+T (k))/k (k < span),
F=(T (k)+T (k-1)+...+T (k-span+1))/span (k >=span),
Wherein T (i) indicates i-th of element of array T, and 1≤i≤k, f are the coordinate obtained after filtering is handled, by it
Array F is added.
8. the method according to the description of claim 7 is characterized in that set the size of prediction window as span in step 7, if
Determining the element number in F is k, then the predictor formula of coordinate are as follows:
Si=sqrt ((F (k-i+1) (1)-F (k-i) (1)) ^2+ (F (k-i+1) (2)-F (k-i) (2)) ^2)
Vi=si/ti,
V=(v1+v2+ ...+vk)/k (k < span),
V=(v1+v2+ ...+vspan)/span (k >=span),
S=v*t,
G=(F (k) (2)-F (k-1) (2))/(F (k) (1)-F (k-1) (1)),
Δ x_=s/sqrt (1+g^2),
Δ y_=s/sqrt ((g^2+1)/g^2),
X=F (k) (1)+sign (F (k) (1)-F (k-1) (1)) * Δ x_,
Y=F (k) (2)+sign (F (k) (2)-F (k-1) (2)) * Δ y_,
Wherein si, vi, v, s, g, Δ x_, Δ y_ indicate introduce intermediate quantity, F (i) indicate array F in i-th of element, 1≤
I≤k, F (i) (j) indicate that j-th of component of i-th of element in array F, j value 1 or 2, ti indicate the i-th close point and i+1
Time interval between close point, t indicate the time interval in current data and F between last 1 point, and sign is symbol letter
Number, positive number 1, negative are -1；(x, y) finally acquired is to predict coordinate.
9. according to the method described in claim 8, it is characterized in that, this TDOA data have been disposed, most in step 8
Whole positioning result is the last one coordinate in array T, and filter result is the last one coordinate in array F, and array F is for showing reality
When track, read next TDOA data at this time, handle the information of subsequent time.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910026517.0A CN109613477B (en) | 2019-01-11 | 2019-01-11 | TDOA (time difference of arrival) positioning tracking method in complex environment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910026517.0A CN109613477B (en) | 2019-01-11 | 2019-01-11 | TDOA (time difference of arrival) positioning tracking method in complex environment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109613477A true CN109613477A (en) | 2019-04-12 |
| CN109613477B CN109613477B (en) | 2020-08-04 |
Family
ID=66015784
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910026517.0A Active CN109613477B (en) | 2019-01-11 | 2019-01-11 | TDOA (time difference of arrival) positioning tracking method in complex environment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109613477B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110850363A (en) * | 2019-10-22 | 2020-02-28 | 南京大学 | Method for carrying out dynamic filtering optimization based on real-time positioning track data |
| CN111444467A (en) * | 2020-04-21 | 2020-07-24 | 南京大学 | Method for local linear interpolation and prediction based on real-time positioning track data |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101052214A (en) * | 2006-04-24 | 2007-10-10 | 华为技术有限公司 | Method for positioning mobile station and positioning auxiliary device for realizing said method |
| WO2008152410A1 (en) * | 2007-06-15 | 2008-12-18 | University Of Plymouth | Method and apparatus for determining the speed and orientation of networked mobile stations |
| JP2009198435A (en) * | 2008-02-25 | 2009-09-03 | Mitsubishi Electric Corp | Positioning device and positioning method for unknown transmission station |
| CN103901397A (en) * | 2014-03-13 | 2014-07-02 | 中国民用航空总局第二研究所 | Choosing method for multi-point positioning location solution in complex scene environment |
| CN104471433A (en) * | 2012-12-05 | 2015-03-25 | 三菱电机株式会社 | Positioning and tracking device |
| CN105137391A (en) * | 2015-09-17 | 2015-12-09 | 中国矿业大学(北京) | TDOA-based CSS (chirp spread spectrum) precise positioning method |
| CN107295636A (en) * | 2017-07-19 | 2017-10-24 | 成都恒高科技有限公司 | A kind of mobile base station positioner, location equipment and method positioned based on TDOA |
| CN108112071A (en) * | 2016-11-11 | 2018-06-01 | 中兴通讯股份有限公司 | Localization method, locating base station, location-server and alignment system |
| CN108738132A (en) * | 2017-04-24 | 2018-11-02 | 温州市鹿城区中津先进科技研究院 | A kind of three base station movement communication positioning methods based on TDOA |
-
2019
- 2019-01-11 CN CN201910026517.0A patent/CN109613477B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101052214A (en) * | 2006-04-24 | 2007-10-10 | 华为技术有限公司 | Method for positioning mobile station and positioning auxiliary device for realizing said method |
| WO2008152410A1 (en) * | 2007-06-15 | 2008-12-18 | University Of Plymouth | Method and apparatus for determining the speed and orientation of networked mobile stations |
| JP2009198435A (en) * | 2008-02-25 | 2009-09-03 | Mitsubishi Electric Corp | Positioning device and positioning method for unknown transmission station |
| CN104471433A (en) * | 2012-12-05 | 2015-03-25 | 三菱电机株式会社 | Positioning and tracking device |
| CN103901397A (en) * | 2014-03-13 | 2014-07-02 | 中国民用航空总局第二研究所 | Choosing method for multi-point positioning location solution in complex scene environment |
| CN105137391A (en) * | 2015-09-17 | 2015-12-09 | 中国矿业大学(北京) | TDOA-based CSS (chirp spread spectrum) precise positioning method |
| CN108112071A (en) * | 2016-11-11 | 2018-06-01 | 中兴通讯股份有限公司 | Localization method, locating base station, location-server and alignment system |
| CN108738132A (en) * | 2017-04-24 | 2018-11-02 | 温州市鹿城区中津先进科技研究院 | A kind of three base station movement communication positioning methods based on TDOA |
| CN107295636A (en) * | 2017-07-19 | 2017-10-24 | 成都恒高科技有限公司 | A kind of mobile base station positioner, location equipment and method positioned based on TDOA |
Non-Patent Citations (1)
| Title |
|---|
| 赵峰等: ""基于RSSI加权数据融合的TDOA定位算法"", 《桂林电子科技大学学报》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110850363A (en) * | 2019-10-22 | 2020-02-28 | 南京大学 | Method for carrying out dynamic filtering optimization based on real-time positioning track data |
| CN111444467A (en) * | 2020-04-21 | 2020-07-24 | 南京大学 | Method for local linear interpolation and prediction based on real-time positioning track data |
| CN111444467B (en) * | 2020-04-21 | 2022-03-15 | 南京大学 | Method for local linear interpolation and prediction based on real-time positioning track data |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109613477B (en) | 2020-08-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112533149B (en) | Moving target positioning algorithm based on UWB mobile node | |
| Wann et al. | Hybrid TDOA/AOA indoor positioning and tracking using extended Kalman filters | |
| CN102131290B (en) | WLAN (Wireless Local Area Network) indoor neighbourhood matching positioning method based on autocorrelation filtering | |
| CN109511085B (en) | UWB fingerprint positioning method based on MeanShift and weighted k nearest neighbor algorithm | |
| CN108307301B (en) | Indoor positioning method based on RSSI ranging and track similarity | |
| CN102088769B (en) | Wireless location method for directly estimating and eliminating non-line-of-sight (NLOS) error | |
| CN110933630A (en) | Indoor three-dimensional positioning method and device based on ultra-wideband communication | |
| CN108872934B (en) | Indoor three-dimensional positioning method based on non-line-of-sight error suppression | |
| CN110850363B (en) | Method for carrying out dynamic filtering optimization based on real-time positioning track data | |
| CN104080165A (en) | Indoor wireless sensor network positioning method based on TDOA | |
| CN106793085A (en) | Fingerprint positioning method based on normality assumption inspection | |
| CN105759274B (en) | A kind of typhoon pays close attention to area's radar precipitation estimating and measuring method | |
| CN106199500A (en) | Fingerprint characteristic localization method and device | |
| CN111722180A (en) | Kalman filtering-based indoor pedestrian positioning method, device and system | |
| CN104754735A (en) | Construction method of position fingerprint database and positioning method based on position fingerprint database | |
| CN109613477A (en) | A kind of TDOA location tracking method under complex environment | |
| CN102880673A (en) | Indoor positioning method | |
| Yin et al. | UWB-based indoor high precision localization system with robust unscented Kalman filter | |
| Lee et al. | Construction of an indoor positioning system for home IoT applications | |
| El Gemayel et al. | Error analysis of a low cost TDoA sensor network | |
| Kuxdorf-Alkirata et al. | Reliable and low-cost indoor localization based on bluetooth low energy | |
| Arai et al. | Color radiomap interpolation for efficient fingerprint wifi-based indoor location estimation | |
| CN111444467B (en) | Method for local linear interpolation and prediction based on real-time positioning track data | |
| CN106772357A (en) | AI PHD wave filters under signal to noise ratio unknown condition | |
| Prieto et al. | NLOS mitigation based on range estimation error characterization in an RTT-based IEEE 802.11, indoor location system |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |