CN109283489A - A kind of UWB fine positioning method, apparatus, equipment and computer readable storage medium - Google Patents
A kind of UWB fine positioning method, apparatus, equipment and computer readable storage medium Download PDFInfo
- Publication number
- CN109283489A CN109283489A CN201811445684.0A CN201811445684A CN109283489A CN 109283489 A CN109283489 A CN 109283489A CN 201811445684 A CN201811445684 A CN 201811445684A CN 109283489 A CN109283489 A CN 109283489A
- Authority
- CN
- China
- Prior art keywords
- solution
- error
- uwb
- optimal observation
- estimation
- 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
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
- G01S5/0294—Trajectory determination or predictive filtering, e.g. target tracking or Kalman filtering
-
- 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
- G01S5/0009—Transmission of position information to remote stations
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Navigation (AREA)
Abstract
This application discloses a kind of UWB fine positioning method, apparatus, equipment and computer readable storage mediums, motion information is made full use of by Kalman filtering, balance optimal observation solution and estimation solution, obtain kalman estimate solution, Kalman filtering is compensated for because of modeling and uncertain noise bring error by BP neural network simultaneously, optimal final estimation solution is obtained, keeps positioning result more accurate, solving existing UWB precise positioning method there is technical issues that.
Description
Technical field
This application involves wireless location technology field more particularly to a kind of UWB fine positioning method, apparatus, equipment and calculating
Machine readable storage medium storing program for executing.
Background technique
Electric power field work management is electricity safety production and its important link, and the environment of electric power field work is multiple
It is miscellaneous, and there is risk, it is understood that there may be charging equipment and venture operation etc. are accidentally touched in electric power field work personnel's mistakenly entering into charged region
Risk, therefore, it is necessary to carry out strong supervision to the staff in electric power field work.Although grid company is by utilizing physics
The technological means such as fence, electrical verification tool, near point alarm device are prevented, but due to the unreachable position of site safety measure, still
The risk hidden danger that field operation can not so be solved, supervises the staff of electric power field work, needs to carry out essence to it
Determining position, existing GPS positioning technology error is big and unstable, it is unable to satisfy security requirement, and utilize UWB location technology benefit
It with precise positioning algorithm locating personnel position, learns its action trail, can be improved positioning accuracy, to improve safety management water
It is flat.But UWB precise positioning method commonly used at present is the letters such as the TDOA/TOA/AOA obtained based on UWB signal measurement
It ceases simultaneous equations to solve, and gained disaggregation is obtained into optimization solution by Kalman filter, Kalman filter is in filtering
In due to model and uncertain noise factor so that gained optimization solution still have large error.
Summary of the invention
The embodiment of the present application provides a kind of UWB fine positioning method, apparatus, equipment and computer readable storage medium, solution
Existing UWB precise positioning method of having determined there is technical issues that.
In view of this, the application first aspect provides a kind of UWB fine positioning method, comprising:
The optimal observation solution for obtaining the whole sampling instants measured in preset time period according to UWB positioning system, obtains institute
State the optimal observation disaggregation in preset time period;
The motion state equation of Kalman filter, initialized card Thalmann filter are determined according to the optimal observation disaggregation
After parameter, the optimal observation disaggregation is inputted into the Kalman filter, is exported after obtaining the Kalman filter filtering
Estimation disaggregation;
The estimation disaggregation is divided into training sample set and test sample collection, by the training sample set and the test specimens
This collection carries out network training by preset BP neural network respectively, with Kalman filtering one-step prediction solution and Kalman Filter Estimation
The first error of solution, kalman gain, optimal observation solve with the second error of Kalman Filter Estimation solution Kalman's filter to input
It is output that wave estimation, which is solved with the third error of Theory Solution, when training error is less than preset error value or the number of iterations greater than preset
When the number of iterations, terminate network training, completes BP neural network model construction;
New optimal observation solution and new estimation solution are obtained, by the BP neural network model, exports Kalman filtering
The third error, the difference for doing the third error to the new estimation solution adjusts, obtains finally estimating to solve.
Preferably, the motion state equation that Kalman filter is determined according to the optimal observation disaggregation, specifically:
The x coordinate and y-coordinate for extracting the optimal observation solution, when finding out each sampling on the direction x and the direction y respectively
The speed at quarter, if 90% or more the speed is in [1.1m/s, 1.95m/s], it is determined that the optimal observation solution is at the uniform velocity
Walking states, if 90% or more the speed is in [35m/s, 78m/s], it is determined that the optimal observation solution is at the uniform velocity to drive
Otherwise state calculates the acceleration of the optimal observation solution.
Preferably, the preset BP neural network is made of input layer, hidden layer and output layer, the nerve of the input layer
First number is 3, and the neuron number of the hidden layer is 6, and the neuron number of the output layer is 1, learning rate 0.035.
Preferably, the first activation primitive of the hidden layer is logsig function, the second activation primitive of the output layer
For tansig function.
Preferably, the preset error value is 0.0001.
Preferably, the preset the number of iterations is 2000.
The application second aspect provides a kind of UWB accurate positioning device, comprising:
Optimal observation solves module, for obtaining the whole sampling instants measured in preset time period according to UWB positioning system
Optimal observation solution, obtain the optimal observation disaggregation in the preset time period;
Estimation solution module, for determining the motion state equation of Kalman filter according to the optimal observation disaggregation, just
After beginningization Kalman filter parameter, the optimal observation disaggregation is inputted into Kalman filter, obtains Kalman filter filter
The estimation disaggregation exported after wave;
BP modeling module, for the estimation disaggregation to be divided into training sample set and test sample collection, by the trained sample
This collection and the test sample collection carry out network training by preset BP neural network respectively, with Kalman filtering one-step prediction solution
It solves with the first error of Kalman Filter Estimation solution, kalman gain, optimal observation and is missed with the second of Kalman Filter Estimation solution
Difference is input, and the third error of Kalman Filter Estimation solution and Theory Solution is output, when training error is less than preset error value, or
When the number of iterations is greater than preset the number of iterations, terminate network training, completes BP neural network model construction;
Test module, it is defeated by the BP neural network model for obtaining new optimal observation solution and new estimation solution
The third error of Kalman filtering out, the difference for doing the third error to the new estimation solution adjust, and obtain final
Estimation solution.
The application third aspect provides a kind of UWB fine positioning equipment, and the equipment includes processor and memory:
Said program code is transferred to the processor for storing program code by the memory;
The processor is used for the fine positioning side UWB according to the instruction execution first aspect in said program code
Method.
The application fourth aspect provides a kind of computer readable storage medium, and the computer readable storage medium is used for
Program code is stored, said program code is for executing UWB fine positioning method described in first aspect.
The 5th aspect of the application provides a kind of computer program product including instruction, when it runs on computers
When, so that the computer executes UWB fine positioning method described in first invention.
As can be seen from the above technical solutions, the embodiment of the present application has the advantage that
In the application, a kind of UWB fine positioning method is provided, comprising: obtain in preset time period according to UWB positioning system
The optimal observation solution of the whole sampling instants measured obtains the optimal observation disaggregation in preset time period;It is solved according to optimal observation
Collect the motion state equation for determining Kalman filter, after initialized card Thalmann filter parameter, by the input of optimal observation disaggregation
Kalman filter obtains the estimation disaggregation exported after Kalman filter filtering;Will estimation disaggregation be divided into training sample set and
Training sample set and test sample collection are carried out network training by preset BP neural network respectively, with karr by test sample collection
The first error of graceful filtering one-step prediction solution and Kalman Filter Estimation solution, kalman gain, optimal observation solution are filtered with Kalman
Second error of wave estimation solution is to input, and the third error of Kalman Filter Estimation solution and Theory Solution is output, works as training error
When being greater than preset the number of iterations less than preset error value or the number of iterations, terminate network training, completes BP neural network model structure
It builds;New optimal observation solution and new estimation solution are obtained, by BP neural network model, the third for exporting Kalman filtering is missed
Difference, the difference for doing third error to new estimation solution adjust, and obtain finally estimating solution.Method provided by the present application, passes through karr
Graceful filtering makes full use of motion information, balances optimal observation solution and estimation solution, obtains kalman estimate solution, while passing through BP nerve
Network building out Kalman filtering obtains optimal final estimation and solves, make to position because of modeling and uncertain noise bring error
As a result more accurate, solving existing UWB precise positioning method there is technical issues that.
Detailed description of the invention
Fig. 1 is the flow diagram of one of the embodiment of the present application UWB fine positioning method;
Fig. 2 is another flow diagram of one of the embodiment of the present application UWB fine positioning method;
Fig. 3 is the BP neural network training algorithm flow diagram in the embodiment of the present application;
Fig. 4 is the structural schematic diagram of one of the embodiment of the present application UWB accurate positioning device.
Specific embodiment
In order to make those skilled in the art more fully understand application scheme, below in conjunction in the embodiment of the present application
Attached drawing, the technical scheme in the embodiment of the application is clearly and completely described, it is clear that described embodiment is only this
Apply for a part of the embodiment, instead of all the embodiments.Based on the embodiment in the application, those of ordinary skill in the art exist
Every other embodiment obtained under the premise of creative work is not made, shall fall in the protection scope of this application.
In order to make it easy to understand, referring to Fig. 1, a kind of UWB fine positioning method in the embodiment of the present application, comprising:
Step 101, the optimal observation for obtaining the whole sampling instants measured in preset time period according to UWB positioning system
Solution, obtains the optimal observation disaggregation in preset time period.
It should be noted that UWB positioning system includes three base stations (B1, B2, B3) and a shifting in the embodiment of the present application
Moving-target label, the configuration software to be matched by ultra-wideband positioning system, establishes space coordinates, if B1 coordinate is (x1,
y1), B2 coordinate is (x2,y2), B3 coordinate is (x3,y3), mobile target labels are (x, y).According to distance calculation formula, have:
Then there is matrix equation:
Dematrix equation can acquire unique solutionWherein, d21For the distance of base station B1 and base station B2, d31For base station B1 and
The distance of base station B3, d1Distance for base station B1 away from mobile tag can acquire the optimal observation solution of each sampling instant accordingly,
Obtain the optimal observation disaggregation in the preset time period.
Step 102, the motion state equation that Kalman filter is determined according to optimal observation disaggregation, the filter of initialized card Germania
After wave device parameter, optimal observation disaggregation is inputted into Kalman filter, obtains the estimation solution exported after Kalman filter filtering
Collection.
It should be noted that in the embodiment of the present application, when carrying out Kalman filter parameter initialization, by starting point coordinate
As the initial value of state vector, it is initial that the error co-variance matrix of optimal observation solution and true value can be initialized as 0 or random
Change, influence of the initialization result to optimal estimation solution is little, and only influences preceding filter value several times, process noise error covariance square
Battle array, observation noise error co-variance matrix, can be in Kalman filtering process according to filter result appropriate adjustment, theoretically, mistake
The value of journey noise error covariance matrix shows that more greatly more trusting observation solves, and optimal estimation solution and observation solution are more close to observation is made an uproar
The value of sound error co-variance matrix is bigger, show more trust observation, optimal estimation solution and observation solution more close to.By optimal observation
Input of the disaggregation as Kalman filter exports the estimation disaggregation for Kalman filter.
Estimation disaggregation is divided into training sample set and test sample collection by step 103, by training sample set and test sample collection
Network training is carried out by preset BP neural network respectively, with Kalman filtering one-step prediction solution and Kalman Filter Estimation solution
First error, kalman gain, optimal observation are solved with the second error of Kalman Filter Estimation solution to input, and Kalman filtering is estimated
It is output that meter, which is solved with the third error of Theory Solution, when training error is less than preset error value or the number of iterations greater than preset iteration
When number, terminate network training, completes BP neural network model construction.
It should be noted that estimation disaggregation is divided into training sample set and test sample collection, this Shen in the embodiment of the present application
Please embodiment estimation solution concentrate, press Kalman filtering sequence, since first Kalman Filter Estimation value choose 80 instruct
Practice sample, wherein training sample input is the first error of Kalman filtering one-step prediction solution and Kalman Filter Estimation solution, card
It is input, Kalman Filter Estimation solution and theory that Germania gain, optimal observation, which are solved with the second error of Kalman Filter Estimation solution,
The third error of solution is output.When training error is less than preset error value or the number of iterations greater than preset the number of iterations, the two is full
When the one of condition of foot, training terminates, and output error obtains final BP neural network model.
Step 104 obtains new optimal observation solution and new estimation solution, by BP neural network model, exports Kalman
The third error of filtering, the difference for doing third error to new estimation solution adjust, and obtain finally estimating solution.
It should be noted that in the embodiment of the present application, after step 103 obtains BP neural network model, subsequent need
The optimal observation solution and new estimation solution of acquisition are exported into the third error of Kalman filtering by BP neural network model,
The difference adjustment that third error is done to new estimation solution, obtains finally estimating solution, does not need reselection training sample again
Network training is carried out with test sample, it is, of course, understood that when the input condition of training sample changes,
It can be with appropriate adjustment BP neural network model, to adapt to different test samples.
The UWB localization method provided in the embodiment of the present application, is acquired first within preset time period using UWB positioning system
The optimal observation disaggregation of whole sampling instants, obtains the optimal observation disaggregation in this time, then analyzes in the period
Approximate motion state, the i.e. state equation of Kalman filter are obtained to optimal observation disaggregation, initialized card Thalmann filter
Optimal observation disaggregation is inputted Kalman filter by parameter, and output is Kalman Filter Estimation disaggregation, is again based on Kalman
Filtering estimation solution set analysis obtains the input and output of training sample, training BP neural network, finally, according to newly generated Kalman
Filter value is inputted, and after BP neural network, the error of Kalman filtering value and theoretical value, newly generated karr can be obtained
Graceful filter value and error carry out difference operation and obtain finally estimating to solve.
In the embodiment of the present application, a kind of UWB fine positioning method is provided, comprising: it is fixed according to UWB in preset time period to obtain
The optimal observation solution for whole sampling instants that position system measures, obtains the optimal observation disaggregation in preset time period;According to optimal
Observation disaggregation determines the motion state equation of Kalman filter, after initialized card Thalmann filter parameter, by optimal observation solution
Collection input Kalman filter, obtains the estimation disaggregation exported after Kalman filter filtering;Estimation disaggregation is divided into trained sample
Training sample set and test sample collection are carried out network training by preset BP neural network respectively by this collection and test sample collection,
With the first error of Kalman filtering one-step prediction solution and Kalman Filter Estimation solution, kalman gain, optimal observation solution and card
Second error of Kalman Filtering estimation solution is to input, and the third error of Kalman Filter Estimation solution and Theory Solution is output, works as instruction
When white silk error is less than preset error value or the number of iterations greater than preset the number of iterations, terminate network training, completes BP neural network
Model construction;New optimal observation solution and new estimation solution are obtained, by BP neural network model, exports the of Kalman filtering
Three errors, the difference for doing third error to new estimation solution adjust, and obtain finally estimating solution.Method provided by the present application, passes through
Kalman filtering makes full use of motion information, balances optimal observation solution and estimation solution, obtains kalman estimate solution, while passing through BP
Neural network compensates for Kalman filtering because of modeling and uncertain noise bring error, obtains optimal final estimation solution, makes
Positioning result is more accurate, and solving existing UWB precise positioning method there is technical issues that.
In order to make it easy to understand, referring to Fig. 2, another kind UWB fine positioning method in the embodiment of the present application, comprising:
Step 201, the optimal observation for obtaining the whole sampling instants measured in preset time period according to UWB positioning system
Solution, obtains the optimal observation disaggregation in preset time period.
It should be noted that the step 101 in step 201 and previous embodiment is consistent, no longer gone to live in the household of one's in-laws on getting married in detail herein
It states.
Step 202, the x coordinate and y-coordinate for extracting optimal observation solution, when finding out each sampling on the direction x and the direction y respectively
The speed at quarter, if 90% or more speed is in [1.1m/s, 1.95m/s], it is determined that optimal observation solution is at the uniform velocity walking states,
If 90% or more speed is in [35m/s, 78m/s], it is determined that optimal observation solution is state of at the uniform velocity driving, and otherwise, is calculated most
The acceleration of excellent observation solution.
It should be noted that in the embodiment of the present application, the selection rule of the motion state equation of Kalman filter can be with
It is:
The x coordinate and y-coordinate for extracting optimal observation solution, find out the speed of each sampling instant on the direction x and the direction y respectively
Degree, if 90% or more speed is in [1.1m/s, 1.95m/s], it is determined that optimal observation solution is at the uniform velocity walking states, if 90%
Above speed is in [35m/s, 78m/s], it is determined that otherwise optimal observation solution calculates optimal observation for state of at the uniform velocity driving
The acceleration of solution is then uniformly accelerated motion when acceleration is basically unchanged;If adjacent two o'clock acceleration changes, it is considered as
Present mode is the mode for calculating current acceleration in real time and being modified.Above-mentioned each one kinematical equation of correspondence of each case,
That is the state equation of Kalman filter.
After step 203, initialized card Thalmann filter parameter, optimal observation disaggregation is inputted into Kalman filter, is obtained
The estimation disaggregation exported after Kalman filter filtering.
Estimation disaggregation is divided into training sample set and test sample collection by step 204, by training sample set and test sample collection
Network training is carried out by preset BP neural network respectively, with Kalman filtering one-step prediction solution and Kalman Filter Estimation solution
First error, kalman gain, optimal observation are solved with the second error of Kalman Filter Estimation solution to input, and Kalman filtering is estimated
It is output that meter, which is solved with the third error of Theory Solution, when training error is less than preset error value or the number of iterations greater than preset iteration
When number, terminate network training, completes BP neural network model construction.
Further, preset BP neural network is made of input layer, hidden layer and output layer, the neuron number of input layer
It is 3, the neuron number of hidden layer is 6, and the neuron number of output layer is 1, learning rate 0.035.
Further, the first activation primitive of hidden layer is logsig function, and the second activation primitive of output layer is
Tansig function.
Further, preset error value is 0.0001.
Further, preset the number of iterations is 2000.
It should be noted that estimation disaggregation is divided into training sample set and test sample collection, the embodiment of the present application is being estimated
Solution concentrates 80 training samples of selection, and wherein training sample input is Kalman filtering one-step prediction solution and Kalman Filter Estimation
The first error of solution, kalman gain, optimal observation solve with the second error of Kalman Filter Estimation solution Kalman's filter to input
Wave estimation solution and the third error of Theory Solution are output.When training error is less than preset error value or the number of iterations greater than preset
The number of iterations, when the two meets one of condition, training terminates, and output error obtains final BP neural network model.
BP neural network training algorithm flow diagram is as shown in figure 3, training sample and the specific network of test sample instruction
Practice process are as follows:
Normalized is done to outputting and inputting for training sample;
It the weight of random initializtion BP neural network and is biased to lesser value, carries out first this training;
If input vector is X=(x1, x2, x3), the weight of hidden layer is Wih, and the first dimension is 3 × 6, is biased to Bh,
Second dimension is 6, is Who in the power of output layer, and third dimension is 6, is biased to Bo, and fourth dimension number is 1.It is hidden in propagated forward
The output for hiding layer is ho=logsig (Wih*X+Bh), and the output of output layer is yo=tansig (Who*ho+Bo);
Calculating target functionValue, k be training sample number, if training miss
Difference is less than 0.0001, then training is completed, and if more than 0.0001, then by error back propagation, finds out the mistake of the hidden layer of output layer
Difference, to obtain the adjustment amount of weight and biasing;
DWho=(yt-yo) * (1-yo) * ho;
DWih=who* (yt-yo) * (1-ho) * ho*X;
DBh=Who* (yt-yo) * (1-ho) * ho;
DBo=(yt-yo) * (1-yo);
Into next iteration:
Who=Who+ μ * dWho, Wih=Wih+ μ * dWih, Bh=Bh+ μ * dBh, Bo=Bo+ μ * dBo at this time, wherein μ
For learning rate, during the application is implemented, μ=0.035 is chosen.Until training error is greater than less than 0.0001 or the number of iterations
2000, then training terminates.
The network of test sample instructs process, and the input and output of test sample and training sample do consistent normalized,
When training error is met the requirements, training error microwave 0.0001 in the embodiment of the present application then completes BP neural network mould
The building of type can come into operation BP neural network, otherwise adjust network parameter, re -training.
Step 205 obtains new optimal observation solution and new estimation solution, by BP neural network model, exports Kalman
The third error of filtering, the difference for doing third error to new estimation solution adjust, and obtain finally estimating solution.
It should be noted that the step 104 in step 205 and preceding embodiment is consistent, no longer it is described in detail herein.
In the embodiment of the present application, a kind of UWB accurate positioning device is additionally provided, comprising:
Optimal observation solves module 401, when being sampled in preset time period according to the whole that UWB positioning system measures for obtaining
The optimal observation solution carved, obtains the optimal observation disaggregation in preset time period.
Estimation solution module 402, for determining the motion state equation of Kalman filter according to optimal observation disaggregation, initially
After changing Kalman filter parameter, optimal observation disaggregation is inputted into Kalman filter, is obtained defeated after Kalman filter filters
Estimation disaggregation out.
BP modeling module 403, for that will estimate that disaggregation is divided into training sample set and test sample collection, by training sample set and
Test sample collection carries out network training by preset BP neural network respectively, is filtered with Kalman filtering one-step prediction solution and Kalman
The first error of wave estimation solution, kalman gain, optimal observation solves and the second error of Kalman Filter Estimation solution is input, card
Kalman Filtering estimates that the third error of solution and Theory Solution is output, when training error is big less than preset error value or the number of iterations
When preset the number of iterations, terminate network training, completes BP neural network model construction.
Test module 404, it is defeated by BP neural network model for obtaining new optimal observation solution and new estimation solution
The third error of Kalman filtering out, the difference for doing third error to new estimation solution adjust, and obtain finally estimating solution.
In the embodiment of the present application, a kind of UWB fine positioning equipment is additionally provided, equipment includes processor and memory:
Program code is transferred to processor for storing program code by memory;
Processor is used for according to the UWB fine positioning method in the instruction execution preceding method embodiment in program code.
In the embodiment of the present application, a kind of computer readable storage medium is additionally provided, computer readable storage medium is used for
Program code is stored, program code is used to execute the UWB fine positioning method in preceding method embodiment.
In the embodiment of the present application, a kind of computer program product including instruction is additionally provided, when it is transported on computers
When row, so that computer executes the UWB fine positioning method in preceding method embodiment.
The description of the present application and term " first " in above-mentioned attached drawing, " second ", " third ", " the 4th " etc. are (if deposited
) it is to be used to distinguish similar objects, without being used to describe a particular order or precedence order.It should be understood that use in this way
Data are interchangeable under appropriate circumstances, so that embodiments herein described herein for example can be in addition to illustrating herein
Or the sequence other than those of description is implemented.In addition, term " includes " and " having " and their any deformation, it is intended that
Cover it is non-exclusive include, for example, containing the process, method, system, product or equipment of a series of steps or units need not limit
In step or unit those of is clearly listed, but may include be not clearly listed or for these process, methods, produce
The other step or units of product or equipment inherently.
In several embodiments provided herein, it should be understood that disclosed device and method can pass through it
Its mode is realized.For example, the apparatus embodiments described above are merely exemplary, for example, the division of the unit, only
Only a kind of logical function partition, there may be another division manner in actual implementation, such as multiple units or components can be tied
Another system is closed or is desirably integrated into, or some features can be ignored or not executed.Another point, it is shown or discussed
Mutual coupling, direct-coupling or communication connection can be through some interfaces, the INDIRECT COUPLING or logical of device or unit
Letter connection can be electrical property, mechanical or other forms.
The unit as illustrated by the separation member may or may not be physically separated, aobvious as unit
The component shown may or may not be physical unit, it can and it is in one place, or may be distributed over multiple
In network unit.It can select some or all of unit therein according to the actual needs to realize the mesh of this embodiment scheme
's.
It, can also be in addition, each functional unit in each embodiment of the application can integrate in one processing unit
It is that each unit physically exists alone, can also be integrated in one unit with two or more units.Above-mentioned integrated list
Member both can take the form of hardware realization, can also realize in the form of software functional units.
If the integrated unit is realized in the form of SFU software functional unit and sells or use as independent product
When, it can store in a computer readable storage medium.Based on this understanding, the technical solution of the application is substantially
The all or part of the part that contributes to existing technology or the technical solution can be in the form of software products in other words
It embodies, which is stored in a storage medium, including some instructions are used so that a computer
Equipment (can be personal computer, server or the network equipment etc.) executes the complete of each embodiment the method for the application
Portion or part steps.And storage medium above-mentioned includes: USB flash disk, mobile hard disk, read-only memory (full name in English: Read-Only
Memory, english abbreviation: ROM), random access memory (full name in English: Random Access Memory, english abbreviation:
RAM), the various media that can store program code such as magnetic or disk.
The above, above embodiments are only to illustrate the technical solution of the application, rather than its limitations;Although referring to before
Embodiment is stated the application is described in detail, those skilled in the art should understand that: it still can be to preceding
Technical solution documented by each embodiment is stated to modify or equivalent replacement of some of the technical features;And these
It modifies or replaces, the spirit and scope of each embodiment technical solution of the application that it does not separate the essence of the corresponding technical solution.
Claims (10)
1. a kind of UWB fine positioning method characterized by comprising
The optimal observation solution for obtaining the whole sampling instants measured in preset time period according to UWB positioning system, obtains described pre-
Set the optimal observation disaggregation in the period;
The motion state equation of Kalman filter, initialized card Thalmann filter parameter are determined according to the optimal observation disaggregation
Afterwards, the optimal observation disaggregation is inputted into the Kalman filter, obtain exporting after the Kalman filter filters estimates
Count disaggregation;
The estimation disaggregation is divided into training sample set and test sample collection, by the training sample set and the test sample collection
Network training is carried out by preset BP neural network respectively, with Kalman filtering one-step prediction solution and Kalman Filter Estimation solution
First error, kalman gain, optimal observation are solved with the second error of Kalman Filter Estimation solution to input, and Kalman filtering is estimated
It is output that meter, which is solved with the third error of Theory Solution, when training error is less than preset error value or the number of iterations greater than preset iteration
When number, terminate network training, completes BP neural network model construction;
It obtains new optimal observation solution and new estimation solution and the institute of Kalman filtering is exported by the BP neural network model
Third error is stated, the difference for doing the third error to the new estimation solution adjusts, and obtains finally estimating solution.
2. UWB fine positioning method according to claim 1, which is characterized in that described true according to the optimal observation disaggregation
Determine the motion state equation of Kalman filter, specifically:
The x coordinate and y-coordinate for extracting the optimal observation solution, find out each sampling instant on the direction x and the direction y respectively
Speed, if 90% or more the speed is in [1.1m/s, 1.95m/s], it is determined that the optimal observation solution is at the uniform velocity to walk
State, if 90% or more the speed is in [35m/s, 78m/s], it is determined that the optimal observation solution is shape of at the uniform velocity driving
Otherwise state calculates the acceleration of the optimal observation solution.
3. UWB fine positioning method according to claim 1, which is characterized in that the preset BP neural network by input layer,
Hidden layer and output layer are constituted, and the neuron number of the input layer is 3, and the neuron number of the hidden layer is 6, described defeated
The neuron number of layer is 1 out, learning rate 0.035.
4. UWB fine positioning method according to claim 3, which is characterized in that the first activation primitive of the hidden layer is
Logsig function, the second activation primitive of the output layer are tansig function.
5. UWB fine positioning method according to claim 1, which is characterized in that the preset error value is 0.0001.
6. UWB fine positioning method according to claim 1, which is characterized in that the preset the number of iterations is 2000.
7. a kind of UWB accurate positioning device characterized by comprising
Optimal observation solves module, for obtaining the whole sampling instants measured in preset time period according to UWB positioning system most
Excellent observation solution, obtains the optimal observation disaggregation in the preset time period;
Estimation solution module, for determining the motion state equation of Kalman filter, initialization according to the optimal observation disaggregation
After Kalman filter parameter, the optimal observation disaggregation is inputted into Kalman filter, after obtaining Kalman filter filtering
The estimation disaggregation of output;
BP modeling module, for the estimation disaggregation to be divided into training sample set and test sample collection, by the training sample set
Network training is carried out by preset BP neural network respectively with the test sample collection, with Kalman filtering one-step prediction solution and card
The first error of Kalman Filtering estimation solution, kalman gain, optimal observation solves and the second error of Kalman Filter Estimation solution is
The third error of input, Kalman Filter Estimation solution and Theory Solution is output, when training error is less than preset error value or iteration
When number is greater than preset the number of iterations, terminate network training, completes BP neural network model construction;
Test module passes through the BP neural network model, output card for obtaining new optimal observation solution and new estimation solution
The third error of Kalman Filtering, the difference for doing the third error to the new estimation solution adjust, are finally estimated
Solution.
8. a kind of UWB fine positioning equipment, which is characterized in that the equipment includes processor and memory:
Said program code is transferred to the processor for storing program code by the memory;
The processor is used for fixed according to the instruction execution UWB essence described in any one of claims 1-6 in said program code
Position method.
9. a kind of computer readable storage medium, which is characterized in that the computer readable storage medium is for storing program generation
Code, said program code require the described in any item UWB fine positioning methods of 1-6 for perform claim.
10. a kind of computer program product including instruction, which is characterized in that when run on a computer, so that described
Computer perform claim requires the described in any item UWB fine positioning methods of 1-6.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811445684.0A CN109283489B (en) | 2018-11-29 | 2018-11-29 | UWB (ultra wide band) precise positioning method, device, equipment and computer readable storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811445684.0A CN109283489B (en) | 2018-11-29 | 2018-11-29 | UWB (ultra wide band) precise positioning method, device, equipment and computer readable storage medium |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109283489A true CN109283489A (en) | 2019-01-29 |
CN109283489B CN109283489B (en) | 2021-04-23 |
Family
ID=65173107
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811445684.0A Active CN109283489B (en) | 2018-11-29 | 2018-11-29 | UWB (ultra wide band) precise positioning method, device, equipment and computer readable storage medium |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109283489B (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109874101A (en) * | 2019-04-15 | 2019-06-11 | 东北大学 | A kind of Indoor Positioning System Using Ultra Wideband Radio and method based on Kalman filtering |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105182288A (en) * | 2015-09-15 | 2015-12-23 | 北京航空航天大学 | Indoor-positioning-system-based RSSI Kalman filtering method |
CN106251696A (en) * | 2016-09-30 | 2016-12-21 | 深圳市金溢科技股份有限公司 | A kind of parking lot intelligent management server, system, method and UWB position label |
CN106793077A (en) * | 2017-01-05 | 2017-05-31 | 重庆邮电大学 | The UWB localization methods and system of dynamic object in a kind of self adaptation room |
CN106908759A (en) * | 2017-01-23 | 2017-06-30 | 南京航空航天大学 | A kind of indoor pedestrian navigation method based on UWB technology |
US20180067191A1 (en) * | 2015-06-16 | 2018-03-08 | Michael Hamilton | Location Estimation System |
-
2018
- 2018-11-29 CN CN201811445684.0A patent/CN109283489B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20180067191A1 (en) * | 2015-06-16 | 2018-03-08 | Michael Hamilton | Location Estimation System |
CN105182288A (en) * | 2015-09-15 | 2015-12-23 | 北京航空航天大学 | Indoor-positioning-system-based RSSI Kalman filtering method |
CN106251696A (en) * | 2016-09-30 | 2016-12-21 | 深圳市金溢科技股份有限公司 | A kind of parking lot intelligent management server, system, method and UWB position label |
CN106793077A (en) * | 2017-01-05 | 2017-05-31 | 重庆邮电大学 | The UWB localization methods and system of dynamic object in a kind of self adaptation room |
CN106908759A (en) * | 2017-01-23 | 2017-06-30 | 南京航空航天大学 | A kind of indoor pedestrian navigation method based on UWB technology |
Non-Patent Citations (1)
Title |
---|
徐爱功等: "RBF神经网络辅助的UWB/INS组合导航算法", 《导航定位学报》 * |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109874101A (en) * | 2019-04-15 | 2019-06-11 | 东北大学 | A kind of Indoor Positioning System Using Ultra Wideband Radio and method based on Kalman filtering |
Also Published As
Publication number | Publication date |
---|---|
CN109283489B (en) | 2021-04-23 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Ma et al. | Fusion of RSS and phase shift using the Kalman filter for RFID tracking | |
CN109509210A (en) | Barrier tracking and device | |
Brook et al. | Collaborative grasp planning with multiple object representations | |
CN107635204A (en) | A kind of indoor fusion and positioning method and device of motor behavior auxiliary, storage medium | |
CN107084714B (en) | A kind of multi-robot Cooperation object localization method based on RoboCup3D | |
CN106125925A (en) | Method is arrested based on gesture and voice-operated intelligence | |
CN101221238B (en) | Dynamic deviation estimation method based on gauss average value mobile registration | |
CN104199022B (en) | Target modal estimation based near-space hypersonic velocity target tracking method | |
CN109521757A (en) | Static-obstacle thing recognition methods and device | |
Tsardoulias et al. | Critical rays scan match SLAM | |
CN106612495B (en) | A kind of indoor orientation method and system based on propagation loss study | |
CN110276783A (en) | A kind of multi-object tracking method, device and computer system | |
CN103426179B (en) | A kind of method for tracking target based on mean shift multiple features fusion and device | |
Qian et al. | Grasp pose detection with affordance-based task constraint learning in single-view point clouds | |
CN103529424B (en) | RFID (radio frequency identification) and UKF (unscented Kalman filter) based method for rapidly tracking indoor target | |
CN109461184A (en) | A kind of crawl point automatic positioning method of robot arm crawl object | |
CN105354860A (en) | Box particle filtering based extension target CBMeMBer tracking method | |
CN106952293A (en) | A kind of method for tracking target based on nonparametric on-line talking | |
Narayana et al. | Intelligent visual object tracking with particle filter based on Modified Grey Wolf Optimizer | |
CN105844217A (en) | Multi-target tracking method based on measure-driven target birth intensity PHD (MDTBI-PHD) | |
CN106403953B (en) | A method of for underwater independent navigation and positioning | |
CN109448023A (en) | A kind of satellite video Small object method for real time tracking of combination space confidence map and track estimation | |
Chen et al. | Multi-scale bio-inspired place recognition | |
CN108470189A (en) | Multiple domain radiation source information fusion method based on antithesis similarity model | |
CN106054167A (en) | Intensity filter-based multi-extended target tracking method |
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 |