CN113419211A - Radio frequency positioning method, electronic equipment and storage medium - Google Patents
Radio frequency positioning method, electronic equipment and storage medium Download PDFInfo
- Publication number
- CN113419211A CN113419211A CN202110614429.XA CN202110614429A CN113419211A CN 113419211 A CN113419211 A CN 113419211A CN 202110614429 A CN202110614429 A CN 202110614429A CN 113419211 A CN113419211 A CN 113419211A
- Authority
- CN
- China
- Prior art keywords
- radio frequency
- frequency measurement
- line
- sight
- measurement data
- 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 80
- 238000003860 storage Methods 0.000 title claims abstract description 15
- 238000005259 measurement Methods 0.000 claims abstract description 159
- 238000012549 training Methods 0.000 claims abstract description 111
- 238000002372 labelling Methods 0.000 claims abstract description 24
- 238000009826 distribution Methods 0.000 claims description 12
- 238000004590 computer program Methods 0.000 claims description 10
- 238000001914 filtration Methods 0.000 claims description 9
- 238000004458 analytical method Methods 0.000 claims description 3
- 238000007781 pre-processing Methods 0.000 claims description 3
- 230000008569 process Effects 0.000 description 13
- 238000004891 communication Methods 0.000 description 9
- 230000008859 change Effects 0.000 description 8
- 230000003044 adaptive effect Effects 0.000 description 7
- 238000005516 engineering process Methods 0.000 description 7
- 238000012544 monitoring process Methods 0.000 description 7
- 238000010276 construction Methods 0.000 description 6
- 238000010586 diagram Methods 0.000 description 6
- 238000010801 machine learning Methods 0.000 description 6
- 230000000474 nursing effect Effects 0.000 description 5
- 230000009467 reduction Effects 0.000 description 5
- 238000004364 calculation method Methods 0.000 description 4
- 238000012706 support-vector machine Methods 0.000 description 4
- 238000013145 classification model Methods 0.000 description 3
- 230000033001 locomotion Effects 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 230000009429 distress Effects 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- 230000002238 attenuated effect Effects 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 230000009545 invasion Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 230000007704 transition Effects 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
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
- G01S5/0278—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 involving statistical or probabilistic considerations
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
- G06N20/10—Machine learning using kernel methods, e.g. support vector machines [SVM]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- Computational Linguistics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Health & Medical Sciences (AREA)
- Evolutionary Biology (AREA)
- Biophysics (AREA)
- Medical Informatics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Probability & Statistics with Applications (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Position Fixing By Use Of Radio Waves (AREA)
Abstract
The invention provides a radio frequency positioning method, electronic equipment and a storage medium, wherein the method comprises the following steps: inputting the radio frequency measurement data into a trained non-line-of-sight identification model to obtain a judgment result of the radio frequency measurement data; when the judgment result of the radio frequency measurement data is non-line-of-sight data, inputting the radio frequency measurement data into a trained non-line-of-sight calibration model to obtain calibrated radio frequency measurement data; the trained non-line-of-sight identification model is obtained by training according to radio frequency measurement sample data carrying a discrimination result label; the trained non-line-of-sight calibration model is obtained by training according to radio frequency measurement sample data carrying a distance tag; the label of the radio frequency measurement sample data is obtained based on radio frequency environment self-training learning. The problems that in the prior art, radio frequency signals are easily interfered by non-line-of-sight shielding, so that the positioning accuracy is reduced, even the positioning cannot be performed, and the cost is increased and the mobility is reduced due to the fact that manual data acquisition and labeling are needed in the traditional supervised learning are effectively solved.
Description
Technical Field
The present invention relates to the field of communication positioning technologies, and in particular, to a radio frequency positioning method, an electronic device, and a storage medium.
Background
With the development of wireless communication technology and internet of things technology, location-aware services have become a basic requirement for many emerging applications. For a long time, the satellite positioning technology has been widely applied to outdoor environments by virtue of the advantages of large coverage, high precision, high reliability and the like, but satellite signals are easily shielded by obstacles and seriously attenuated in environments such as indoor and urban canyons, so that the positioning precision is sharply reduced, and even positioning cannot be performed. Therefore, in recent years, many non-satellite positioning means have been developed, and these technologies provide positioning, navigation and other additional services in the whole process by acquiring high-precision and reliable position information of a user in an indoor complex environment, mainly including methods based on radio frequency communication, inertial navigation, visual navigation, and the like.
At present, an indoor positioning technology based on radio frequency signals develops rapidly, but a large number of buildings, walls, furniture and other shelters exist in an actual indoor environment, so that the signals are interfered by non-line-of-sight (NLOS) in a propagation process, a measurement result has large errors, and robust positioning under a complex scene is difficult to realize. In order to solve the influence of non-line-of-sight propagation on a positioning system, the existing method focuses on non-line-of-sight identification and calibration, for example, NLOS measurement is abandoned through a non-line-of-sight identification method, and only LOS measurement with higher precision is used for positioning; or firstly calibrating NLOS measurement by a non-line-of-sight calibration method, and then carrying out fusion positioning. In recent years, a non-line-of-sight identification and calibration method based on machine learning provides a more flexible non-parametric framework, and the model performance is remarkably improved. The method is based on a supervised learning theory, needs to manually acquire and label a large amount of data in a positioning area in advance to construct a training data set, has the obvious defects of time and labor consumption and high cost, needs to acquire and label data again when the environment is dynamically changed, such as increase and decrease and movement of furniture, is difficult to update and maintain a database, and has low model mobility. Therefore, how to realize automatic data acquisition and labeling, reduce the construction and update cost of the database, quickly obtain a model adaptive to the current environment, improve the precision and robustness of the whole positioning system, and are the problems to be solved by the current high-precision indoor positioning technology based on radio frequency signals.
Disclosure of Invention
The invention provides a radio frequency positioning method, electronic equipment and a storage medium, which are used for solving the problem that the identification of card types cannot be well realized in the prior art.
The invention provides a radio frequency positioning method, which comprises the following steps:
inputting the radio frequency measurement data into a trained non-line-of-sight identification model to obtain a judgment result of the radio frequency measurement data;
when the judgment result of the radio frequency measurement data is non-line-of-sight data, inputting the radio frequency measurement data into a trained non-line-of-sight calibration model to obtain calibrated radio frequency measurement data;
the trained non-line-of-sight recognition model is obtained by training according to radio frequency measurement sample data carrying a discrimination result label; the trained non-line-of-sight calibration model is obtained by training according to radio frequency measurement sample data carrying a distance tag; the label of the radio frequency measurement sample data is obtained based on radio frequency environment self-training learning.
According to an embodiment of the present invention, before the step of inputting the radio frequency measurement data into the trained non-line-of-sight recognition model, the method further includes:
analyzing the map information to obtain a grid map; wherein the weight of each grid point in the grid-point map is used for approximating the continuous probability distribution of the user state;
updating the weight of each grid point based on the pedestrian dead reckoning information acquired by the mobile terminal to obtain the updated grid point weight;
matching and aligning the radio frequency measurement data with grid points obtained by the pedestrian dead reckoning information based on a timestamp matching principle to obtain grid points corresponding to each group of radio frequency measurement data;
obtaining a distinguishing result label and a distance label of each group of radio frequency measurement data according to the lattice points and the map information corresponding to each group of radio frequency measurement data, and determining the radio frequency measurement data carrying the distinguishing result label and the radio frequency measurement data carrying the distance label;
and respectively carrying out sample quantity expansion on the radio frequency measurement data carrying the identification result label and the radio frequency measurement data carrying the distance label according to the updated lattice point weight to obtain a first training data set and a second training data set.
According to an embodiment of the present invention, after the step of obtaining the first training data set and the second training data set, the method further includes:
s31, training a preset model based on the first training data set and the second training data set respectively to obtain a non-line-of-sight recognition model and a non-line-of-sight calibration model;
s32, performing radio frequency measurement identification on the first training data set through the non-line-of-sight identification model, and performing radio frequency measurement calibration on the second training data set through the non-line-of-sight calibration model to obtain auxiliary positioning information;
s33, updating according to the weight of each grid point of the auxiliary positioning information to obtain a new grid point weight;
s34, according to the new lattice point weight, re-labeling the radio frequency measurement data in the first training data set and the second training data set to obtain a first training enhanced data set and a second training enhanced data set;
and S35, retraining the preset model according to the first training enhanced data set and the second training enhanced data set respectively, and repeating the steps S31 to S35 until preset conditions are met to obtain a trained non-line-of-sight recognition model and a trained non-line-of-sight calibration model.
According to the radio frequency positioning method provided by the embodiment of the invention, the step of analyzing the map information to obtain the lattice map specifically comprises the following steps:
carrying out uniform discretization pretreatment on the map information to obtain an initial grid map;
and performing connectivity analysis on the initial grid-point map, and excluding positions which cannot be reached to obtain a final grid-point map.
According to the radio frequency positioning method provided by the embodiment of the invention, the step of updating the weight of each grid point based on the pedestrian dead reckoning information acquired by the mobile terminal to obtain the updated grid point weight specifically comprises the following steps:
obtaining the initialization weight of each grid point according to the grid-point map;
obtaining the prior weight of each grid point in a state prediction equation of Bayesian filtering according to the pedestrian dead reckoning information;
and updating the weight of each grid point according to the prior weight of each grid point to obtain the updated grid point weight.
According to the radio frequency positioning method provided by the embodiment of the invention, the uniform discretization preprocessing process comprises the following specific steps:
wherein p iskAnd thetakRespectively representing the position and the course angle of the user at the k step, and jointly forming a state variable of the system, zkRepresenting the radio frequency measurements taken by the user at step k, the observed variables constituting the system, p (p)k,θk|z1：k) Representing the posterior probability distribution of the k-th round state,
indicates that the k-th wheel state is (i)p,ia) The posterior weight of (a).
According to the radio frequency positioning method provided by the embodiment of the invention, the state prediction equation of the Bayesian filtering specifically comprises the following steps:
wherein,
representing derived location grid points based on indoor map information
Is transferred to
Probability of (m)kThe pedestrian dead reckoning information of the k step comprises the step length l of the k stepkAnd heading angle variation amount Δ θk，
Indicates that the state of the k-1 th round is (i)p,ia) The posterior weight of (a) is calculated,
indicates that the k-th wheel state is (i)p，ia) A priori weight of.
More specifically, the method further comprises: after the step of obtaining calibrated radio frequency measurement data, the method further comprises:
obtaining posterior weights of each lattice point in a state updating equation of Bayesian filtering according to the calibrated radio frequency measurement data, and obtaining the final positioning of the user according to the posterior weights;
wherein the state update equation is:
wherein, the user positioning result is:
wherein,
is a system observation equation which represents the kth step of the user at a position grid point
To obtain an observed variable zkThe probability of (c).
An embodiment of the present invention further provides a radio frequency positioning apparatus, including:
the processing module is used for inputting the radio frequency measurement data into the trained non-line-of-sight identification model to obtain a judgment result of the radio frequency measurement data, and when the judgment result of the radio frequency measurement data is the non-line-of-sight data, the radio frequency measurement data is input into the trained non-line-of-sight calibration model to obtain calibrated radio frequency measurement data;
the positioning module is used for analyzing map information, obtaining the prior weight of each grid point based on pedestrian dead reckoning information obtained by the mobile terminal, obtaining the posterior weight of each grid point based on calibrated radio frequency measurement data, and obtaining the final positioning result of the user according to the posterior weight;
the trained non-line-of-sight recognition model is obtained by training according to radio frequency measurement sample data carrying a discrimination result label; the trained non-line-of-sight calibration model is obtained by training according to radio frequency measurement sample data carrying a distance tag; the label of the radio frequency measurement sample data is obtained based on radio frequency environment self-training learning.
The invention also provides an electronic device comprising a memory, a processor and a computer program stored on the memory and operable on the processor, wherein the processor implements the steps of the radio frequency positioning method as described in any one of the above when executing the program.
The invention also provides a non-transitory computer-readable storage medium having stored thereon a computer program which, when executed by a processor, performs the steps of the radio frequency location method as described in any of the above.
The invention provides a radio frequency positioning method, electronic equipment and storage medium, which can improve the accuracy of radio frequency positioning by identifying and calibrating radio frequency measurement data through a trained non-line-of-sight identification model and a trained non-line-of-sight calibration model, effectively solves the problems that in the prior art, the positioning precision is reduced or even the radio frequency signals are not positioned due to the interference of non-line-of-sight shielding, and the cost is increased and the mobility is reduced due to the fact that the traditional supervision learning needs to artificially acquire and label data, can avoid manually acquiring and labeling a large amount of data in a positioning area in advance by providing a positioning method based on radio frequency environment self-training learning, directly utilizes position estimation to realize automatic data acquisition and labeling, can also acquire a large amount of training data again through a short-time positioning process when the environment is dynamically changed, and reduces the construction and updating costs of a database, improving the mobility of the model; in addition, the positioning method based on the radio frequency environment self-training learning continuously improves the label quality by using the updated position estimation to re-label data in an iteration mode, can quickly obtain a model adaptive to the current environment, further obtains more reliable auxiliary positioning information to calibrate the user positioning result, and improves the precision and robustness of the whole positioning system.
Drawings
In order to more clearly illustrate the technical solutions of the present invention or the prior art, the drawings needed for the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and those skilled in the art can also obtain other drawings according to the drawings without creative efforts.
Fig. 1 is a schematic flow chart of a radio frequency positioning method according to an embodiment of the present invention;
fig. 2 is a flowchart of a positioning method for radio frequency environment self-training learning according to an embodiment of the present invention;
fig. 3 is a schematic structural diagram of a radio frequency positioning device according to an embodiment of the present invention;
fig. 4 is a second schematic structural diagram of an rf positioning device according to an embodiment of the present invention;
fig. 5 is a schematic physical structure diagram of an electronic device provided in the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings, and it is obvious that the described embodiments are some, but not all embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The application of the rf positioning method described in the present invention includes but is not limited to the following scenarios:
the positioning method can be applied to hospital patient monitoring and logistics management. A plurality of base stations are deployed in a hospital, a positioning bracelet is worn for a special patient needing real-time monitoring, the position track of the patient can be checked in real time by a background, the moving range of the patient can be limited by setting an electronic fence, and the patient can be positioned and rescued in time when an emergency occurs; the positioning chest card and the positioning work card are worn by security personnel, nursing personnel and the like of a hospital, the positions and the movement tracks of the security personnel, the nursing personnel and the like can be checked in real time by the background, wireless attendance checking, electronic roll calling and the like are realized, and the management efficiency is improved.
The positioning method can be applied to nursing of personnel in nursing homes. A plurality of basic stations are deployed in a nursing home, wear the location label for the old man, location bracelet etc, can master old man's positional information and historical orbit information in real time, the safety of protection old man, and can implement various behavior monitoring in the region, including overtime monitoring, motionless monitoring etc, go up and down stairs to the old man, walk and tumble, get into the condition in danger area and in time report an emergency and ask for help the signal through the urgent distress call button to managers, managers can in time obtain old man's distress call information and accurate position, promptly help.
The positioning method can be applied to prison personnel management. A plurality of base stations are deployed in a prison, a prisoner wears an anti-dismantling positioning bracelet, the position, the moving track, the personnel distribution condition and the like of each area of the prisoner in the prison are monitored in real time, a system background can set forbidden zone invasion, separation supervision alarm and the like, and the prisoner is intelligently supervised; the positioning worker plate is worn for the prison police, the position of the prison police is grasped in real time, the patrol route of the prison police is intelligently planned, the prison police in the prison is timely reminded of going to an emergency area, and the management work of the prison is more intelligent.
The positioning method can be applied to factory personnel management. A plurality of base stations are deployed in a factory, positioning labels are combined with worker cards, bracelets, helmets and the like, all-weather all-day-long positioning monitoring is carried out on all workers in the factory, the positions of the workers can be checked in real time on a monitoring display platform, whether the workers are on duty or not is determined, the positions of the workers and a distribution heat map are counted, automatic attendance is achieved, the phenomena of worker off duty, empty duty and the like are intelligently identified, the working time length of the workers is counted in real time, the working time utilization rate of the workers is analyzed, and labor consumption is reduced.
The positioning method can be applied to the aspect of warehouse logistics. A plurality of base stations are deployed in a warehouse, and positioning worker cards are worn by people, so that the position information of the people can be accurately displayed in a logistics storage control center, the historical track condition of the people can be inquired and replayed, the activity operation track of the people can be tracked, the operation flow can be optimized, the production efficiency is improved, the fine management of the people is realized, and the accurate management and control are realized to lean on production and reasonable scheduling and arrangement.
Fig. 1 is a schematic flow chart of a radio frequency positioning method according to an embodiment of the present invention, as shown in fig. 1, including:
step S1, inputting the radio frequency measurement data into the trained non-line-of-sight recognition model to obtain the discrimination result of the radio frequency measurement data;
specifically, the rf measurement data described in the present invention may refer to the rf measurement data detected by the positioning target through the rf measurement sensing device in an indoor environment, and the rf measurement data includes, but is not limited to, a signal arrival timestamp, a signal sampling waveform, a measurement distance between the base station and the user equipment, a location of the base station, and the like.
The radio frequency measurement described in the present invention uses wireless signal types including, but not limited to, ultra wide band signals, 5G signals, bluetooth signals, Wi-Fi signals, etc.
Because a great number of shelters exist in the indoor environment, the radio frequency signal is easily interfered by the NLOS in the propagation process, and the final measurement result has great errors.
If the judgment result of the radio frequency measurement data is line-of-sight data, it indicates that the equipment measurement data is not interfered by the NLOS, that is, there is no shielding between the positioning target and the base station, and the positioning result is not affected by the shielding object, so that the radio frequency measurement result does not need to be calibrated.
If the judgment result of the radio frequency measurement data is non-line-of-sight data, it indicates that the equipment measurement data is interfered by the NLOS, and there is a high possibility that the positioning is inaccurate, and further calibration is needed.
Step S2, when the judgment result of the radio frequency measurement data is non-line-of-sight data, inputting the radio frequency measurement data into a trained non-line-of-sight calibration model to obtain calibrated radio frequency measurement data;
the trained non-line-of-sight recognition model is obtained by training according to radio frequency measurement sample data carrying a discrimination result label; the trained non-line-of-sight calibration model is obtained by training according to radio frequency measurement sample data carrying a distance tag; the label of the radio frequency measurement sample data is obtained based on radio frequency environment self-training learning.
When the discrimination result of the radio frequency measurement data described in the invention is non-line-of-sight data, it indicates that shielding exists between the base station and the positioning target, so that the radio frequency measurement data needs to be calibrated through a trained non-line-of-sight calibration model, thereby obtaining calibrated radio frequency measurement data.
The discrimination result tag described in the present invention specifically includes a non-line-of-sight data tag and a line-of-sight data tag.
The distance tag described in the present invention specifically refers to distance data information between a base station and a positioning target.
The trained non-line-of-sight recognition model and the trained non-line-of-sight calibration model described in the embodiment of the invention are self-trained and learned, and data is re-labeled by using updated position estimation in an iterative manner, so that the quality of the label is continuously improved, and a model adaptive to the current environment can be quickly obtained, namely the training data and the label are automatically updated, and a large amount of labels do not need to be labeled manually.
The invention identifies and calibrates the radio frequency measurement data through the trained non-line-of-sight identification model and the trained non-line-of-sight calibration model, thereby improving the accuracy of radio frequency positioning, and the application effectively solves the problems that in the prior art, the radio frequency signal is easy to be blocked and interfered by non-line-of-sight so as to cause the reduction of positioning precision and even the incapability of positioning, and the problem that the traditional supervised learning needs to manually collect and label data, which leads to the increase of cost and the reduction of mobility, the positioning method based on the radio frequency environment self-training learning can avoid manually acquiring and labeling a large amount of data in a positioning area in advance, directly utilizes position estimation to realize automatic data acquisition and labeling, when the environment changes dynamically, a large amount of training data can be obtained again through a short-time positioning process, the construction and updating cost of the database is reduced, and the model mobility is improved; in addition, the positioning method based on the radio frequency environment self-training learning continuously improves the label quality by using the updated position estimation to re-label data in an iteration mode, can quickly obtain a model adaptive to the current environment, further obtains more reliable auxiliary positioning information to calibrate the user positioning result, and improves the precision and robustness of the whole positioning system.
Based on any of the above embodiments, before the step of inputting the radio frequency measurement data into the trained non-line-of-sight recognition model, the method further includes:
analyzing the map information to obtain a grid map; wherein the weight of each grid point in the grid-point map is used for approximating the continuous probability distribution of the user state;
updating the weight of each grid point based on the pedestrian dead reckoning information acquired by the mobile terminal to obtain the updated grid point weight;
matching and aligning the radio frequency measurement data with grid points obtained by the pedestrian dead reckoning information based on a timestamp matching principle to obtain grid points corresponding to each group of radio frequency measurement data;
obtaining a distinguishing result label and a distance label of each group of radio frequency measurement data according to the lattice points and the map information corresponding to each group of radio frequency measurement data, and determining the radio frequency measurement data carrying the distinguishing result label and the radio frequency measurement data carrying the distance label;
and respectively carrying out sample quantity expansion on the radio frequency measurement data carrying the identification result label and the radio frequency measurement data carrying the distance label according to the updated lattice point weight to obtain a first training data set and a second training data set.
Specifically, the map information described in the present invention is available from a builder, and includes specific information on a wall or other obstruction and space.
The human track reckoning information described in the invention can be obtained by using an accelerometer, a gyroscope, a magnetic sensor and other inertial sensors which are arranged in the mobile phone.
The weight of each grid point in the grid-point map can represent the possibility that the user is at the grid point position.
The timestamp matching principle described in the embodiment of the invention is to match the radio frequency measurement data with the base station timestamp with the pedestrian dead reckoning information with the mobile phone timestamp according to the timestamp.
The invention obtains the pedestrian dead reckoning information obtained by the mobile terminal, and obtains the weight change information on different grid points according to the map information and the pedestrian dead reckoning information. Specifically, the prior weight of the state is obtained in a state prediction equation of bayesian filtering, where the state prediction equation is:
wherein,
representing derived location grid points based on indoor map information
Is transferred to
Probability of (m)kThe pedestrian dead reckoning information of the k step comprises the step length l of the k stepkAnd heading angle variation amount Δ θk，
Indicates that the state of the k-1 th round is (i)p,ia) The posterior weight of (a) is calculated,
indicates that the k-th wheel state is (i)p，ia) A priori weight of.
The transition probability between two location grid points is determined by the connectivity between the grid points, taking into account the map information. Specifically, if the lattice point is located
And
is in communication with each other
If the position lattice point
And
are not communicated with each other, then
And according to a timestamp matching principle, matching and aligning the acquired radio frequency measurement data and each step of the user, and automatically labeling the data by combining the map information, the position of the base station, the positions of the grid points and the weight change information on different grid points.
To make full use of weight change information at different grid points, location weights are selected
(in the first iteration are
) The maximum M grid points are used as candidate user positions, and the position weights are normalized to obtain corresponding probabilities { prob1,prob2,…,probM}. For a piece of data matched at the current moment, defining a sample expansion factor C, and copying the data according to the probability to obtain C multiplied by probiAnd allocating labels to the samples, wherein the labels are obtained by calculation according to the positions of corresponding candidate users, and the label contents include but are not limited to a discrimination result label, a distance label and the like:
1) and (4) judging a result label: connecting by taking the position of the base station and the candidate user position as end points, and marking as non-line-of-sight if the line segment passes through the obstacle in the indoor map information
If the line segment does not pass through the obstacle in the indoor map information, marking as the sight distance
2) Distance label: using the position of the base station and the connection length of the candidate user position as the distance label
Wherein
Indicates the currently connected base station bkThe position of (a).
For example, taking M ═ 2 and C ═ 10, the probability of obtaining the candidate user position is { prob1＝0.7,prob20.3, for a piece of data matched at the current time, C × prob is copied1Copy Cxprob 7 times and compute label according to the first candidate user position1And calculating the label according to the position of the second candidate user 3 times to finally obtain 10 samples. The method can expand the number of samples, effectively stores weight change information on different lattice points, and improves the accuracy of the labeled content.
And finally, constructing a training data set, wherein the training data set is composed of the collected radio frequency measurement data and the calculated labels, and a first training data set and a second training data set are obtained.
According to the invention, through the connectivity relation and the new person track calculation information on the lattice map, automatic data acquisition and marking are realized by using position estimation, and when the environment is dynamically changed, a large amount of training data can be obtained again through a short-time positioning process, so that the construction and updating cost of a database is reduced, and the model mobility is improved.
Based on any of the above embodiments, after the step of obtaining the first training data set and the second training data set, the method further comprises:
s31, training a preset model based on the first training data set and the second training data set respectively to obtain a non-line-of-sight recognition model and a non-line-of-sight calibration model;
s32, performing radio frequency measurement identification on the first training data set through the non-line-of-sight identification model, and performing radio frequency measurement calibration on the second training data set through the non-line-of-sight calibration model to obtain auxiliary positioning information;
s33, updating according to the weight of each grid point of the auxiliary positioning information to obtain a new grid point weight;
s34, according to the new lattice point weight, re-labeling the radio frequency measurement data in the first training data set and the second training data set to obtain a first training enhanced data set and a second training enhanced data set;
and S35, retraining the preset model according to the first training enhanced data set and the second training enhanced data set respectively, and repeating the steps S31 to S35 until preset conditions are met to obtain a trained non-line-of-sight recognition model and a trained non-line-of-sight calibration model.
Specifically, in the invention, effective features related to non-line-of-sight are extracted from the acquired radio frequency measurement data. In an embodiment of the present invention, the performing dimension reduction on the sampled waveform data r (t) of the received UWB signal, and extracting a statistic related to non-line-of-sight as an input feature from the processed waveform data r (t) of the received UWB signal includes:
1) received signal energy:
εr＝∫T|r(t)|2dt
2) maximum amplitude:
rmax＝maxt|r(t)|
3) rise time:
trise＝tH-tL
wherein,
tL＝min{t：|r(t)|≥ασn}
tH＝min{t：|r(t)|≥βrmax}
4) average excess delay:
5) root mean square delay spread:
6) kurtosis:
wherein,
the above 6-dimensional characteristics and the measured distance between the base station and the UE
Together forming an input feature vector x for the machine learning model.
And then selects the appropriate machine learning method. In one embodiment of the invention, a Support Vector Machine (SVM) model is adopted, which has the advantages of simple implementation, less adjustable parameters and good generalization performance. According to the first training data set
Training a classification model to obtain a non-line-of-sight recognition model, based on the second training data set
The regression model is trained to derive a non-line-of-sight calibration model. The non-linear transformation of the input features being included in
And may be implemented by a kernel function.
The classification model based on SVM is:
solving an optimization problem:
the final non-line-of-sight recognition model is obtained as follows:
the regression model based on SVM is:
solving an optimization problem:
the final non-line-of-sight calibration model is obtained as follows:
identifying and calibrating radio frequency measurement data through a non-line-of-sight identification model and a non-line-of-sight calibration model, reducing non-line-of-sight errors, and acquiring auxiliary positioning information, and obtaining posterior weight of a state in a state updating equation of Bayesian filtering, wherein the state updating equation is as follows:
wherein,
is a system observation equation which represents the kth step of the user at a position grid point
To obtain an observed variable zkThe probability of (c).
In one embodiment of the present invention, only non-line-of-sight data is corrected, taking into account that UWB can achieve accurate distance estimates under line-of-sight conditions. Specifically, for a received UWB signal waveform, whether the received UWB signal waveform is non-line-of-sight data or not is judged according to a non-line-of-sight identification model, if the received UWB signal waveform is the non-line-of-sight data, the error is continuously corrected by adopting a non-line-of-sight calibration model, and corrected distance measurement is obtained; otherwise, the original distance measurement is still used. Thus, the system observation equation can be written as:
wherein,
representing the original distance measurement between the user at step k and the base station that was determined to be line-of-sight,
the corrected distance measurement between the user at the k-th step and the base station determined to be not in line of sight is shown. It can be assumed that the distance measurements follow a Gaussian distribution with the mean being the current location grid
And the connected base station bkDistance between them
Estimating the user position by using the posterior weight of the state after the updating of the auxiliary positioning information:
and returning to the step S31 of reconstructing the training data set, and iteratively executing the learning and positioning process, namely, re-labeling the radio frequency measurement data by using the updated position estimation so as to obtain a more accurate label, improve the model prediction performance, enhance the non-line-of-sight error calibration capability, and iteratively correct the user positioning result. After several iterations, the system can quickly obtain a model adaptive to the current environment, and the precision and robustness of the whole positioning system are improved.
The positioning system and the method based on the radio frequency environment self-training learning effectively solve the problems that in the prior art, radio frequency signals are easily interfered by non-line-of-sight shielding, so that the positioning accuracy is reduced, even the positioning is impossible, and the cost is increased and the mobility is reduced due to the fact that the traditional supervised learning needs to artificially acquire and label data; the positioning method based on the radio frequency environment self-training learning can avoid manually acquiring and labeling a large amount of data in a positioning area in advance, realize automatic data acquisition and labeling by directly utilizing position estimation, and can also obtain a large amount of training data again through a short-time positioning process when the environment is dynamically changed, so that the construction and updating cost of a database is reduced, and the model mobility is improved; in addition, the positioning method based on the radio frequency environment self-training learning continuously improves the label quality by using the updated position estimation and re-labeling data in an iteration mode, can quickly obtain a model adaptive to the current environment, accordingly obtains more reliable auxiliary positioning information to correct the positioning result of the user, and improves the precision and robustness of the whole positioning system.
Based on any of the above embodiments, the step of analyzing the map information to obtain the grid-point map specifically includes:
carrying out uniform discretization pretreatment on the map information to obtain an initial grid map;
and performing connectivity analysis on the initial grid-point map, and excluding positions which cannot be reached to obtain a final grid-point map.
Specifically, map information is analyzed, uniform discretization preprocessing is carried out to obtain a grid map, and connectivity among grid points is established to represent the limitation of obstacles in the map on the movement of a user and exclude positions which cannot be reached. Specifically, if the connecting line between the two position lattice points is intersected with the wall body, the connecting line is not communicated; considering that the step length of one step of walking of the pedestrian is not too large, if the length of the connecting line between the two position grid points exceeds the threshold value dmaxIt is not connected. Meanwhile, the continuous probability distribution of the user state is expressed as the weight on the discrete grid point so as to improve the calculation efficiency of the system, and the discrete approximation process is expressed as:
wherein p iskAnd thetakRespectively representing the position and the course angle of the user at the k step, and jointly forming a state variable of the system, zkRepresenting the radio frequency measurements taken by the user at step k, the observed variables constituting the system, p (p)k，θk|z1：k) Representing the posterior probability distribution of the k-th round state,
indicates that the k-th wheel state is (i)p，ia) The posterior weight of (a).
In the embodiment of the invention, in the process of constructing the grid-point map, the connectivity among the grid points is fully considered, so that the model can be better trained.
Fig. 2 is a flowchart of a positioning method for radio frequency environment self-training learning according to an embodiment of the present invention, as shown in fig. 2, including:
s101, analyzing map information, carrying out uniform discretization pretreatment to obtain a grid-point map, establishing connectivity among grid points, and expressing continuous probability distribution as weights on different grid points;
s102, acquiring pedestrian dead reckoning information obtained by the mobile terminal, and obtaining weight change information on different grid points according to the map information and the pedestrian dead reckoning information;
s103, firstly, matching and aligning the collected radio frequency measurement data and each step of the user according to a timestamp matching principle, and then automatically labeling data by combining the map information, the position of the base station, the positions of the grid points and weight change information on different grid points to construct a training data set;
and S104, completing model self-training by adopting a machine learning method according to the training data set obtained in the S103, and obtaining a non-line-of-sight recognition and calibration model.
S105, identifying and calibrating the radio frequency measurement data according to the non-line-of-sight identification model and the non-line-of-sight calibration model obtained in the S104, reducing non-line-of-sight errors, acquiring auxiliary positioning information and updating weights on different grid points;
s106, obtaining a user positioning result according to the updated weights on different grid points, re-labeling the radio frequency measurement data by using the updated weight information to obtain a better and accurate label, enhancing the non-line-of-sight error calibration capability of the model, returning to the step S103, iterating the learning and positioning process, and correcting the user positioning result.
Fig. 3 is a schematic structural diagram of a radio frequency positioning device according to an embodiment of the present invention, as shown in fig. 3, including: a processing module 310 and a positioning module 320; the processing module 310 is configured to input radio frequency measurement data into a trained non-line-of-sight identification model to obtain a discrimination result of the radio frequency measurement data, and input the radio frequency measurement data into a trained non-line-of-sight calibration model when the discrimination result of the radio frequency measurement data is non-line-of-sight data to obtain calibrated radio frequency measurement data; the positioning module 320 is configured to analyze map information, obtain a prior weight of each grid point based on pedestrian dead reckoning information obtained by the mobile terminal, obtain a posterior weight of each grid point based on calibrated radio frequency measurement data, and obtain a final positioning result of the user according to the posterior weight; the trained non-line-of-sight recognition model is obtained by training according to radio frequency measurement sample data carrying a discrimination result label; the trained non-line-of-sight calibration model is obtained by training according to radio frequency measurement sample data carrying a distance tag; the label of the radio frequency measurement sample data is obtained based on radio frequency environment self-training learning.
Fig. 4 is a second schematic structural diagram of a radio frequency positioning device provided in an embodiment of the present invention, as shown in fig. 4, the radio frequency positioning device in the embodiment of the present invention further includes a map module, an acquisition module, a learning module, and a positioning module, wherein:
and the map module is used for analyzing the map information to obtain a grid map and representing the continuous probability distribution of the user state as the weight on different grid points.
The acquisition module is used for acquiring the pedestrian dead reckoning information obtained by the mobile terminal and obtaining the weight change information on different grid points according to the map information and the pedestrian dead reckoning information.
The learning module is used for realizing automatic acquisition and labeling of radio frequency measurement, constructing a training data set and obtaining a self-training non-line-of-sight identification and calibration model based on a machine learning method.
The positioning module is used for recognizing and correcting radio frequency measurement data based on the self-training non-line-of-sight recognition model and the non-line-of-sight calibration model, acquiring auxiliary positioning information to update weights on different grid points, and then fusing the indoor map information, the pedestrian dead reckoning information and the auxiliary positioning information through Bayesian filtering to complete indoor positioning.
The learning module comprises a data acquisition module, a data labeling module and a model training module, wherein:
the data acquisition module is used for automatically receiving radio frequency signals communicated between the base station and the user equipment and acquiring radio frequency measurement data from the radio frequency signals, wherein the radio frequency measurement data comprises but is not limited to a signal arrival timestamp, a signal sampling waveform, a measurement distance between the base station and the user equipment, a position of the base station and the like.
The data labeling module is used for automatically labeling the radio frequency measurement data so as to construct a training data set, labeling contents include but are not limited to a discrimination result label, a distance label and the like, and the labeling contents can be obtained by calculation based on the map information, the position of the base station, the positions of the lattice points and weight change information on different lattice points.
And the model training module is used for finishing model training based on the training data set, training the classification model to obtain a non-line-of-sight recognition model by adopting a machine learning method, and training the regression model to obtain a non-line-of-sight calibration model.
The embodiment of the invention identifies and calibrates the radio frequency measurement data through the trained non-line-of-sight identification model and the trained non-line-of-sight calibration model, thereby improving the accuracy of radio frequency positioning, and the application effectively solves the problems that in the prior art, the radio frequency signal is easy to be blocked and interfered by non-line-of-sight so as to cause the reduction of positioning precision and even the incapability of positioning, and the problem that the traditional supervised learning needs to manually collect and label data, which leads to the increase of cost and the reduction of mobility, the positioning method based on the radio frequency environment self-training learning can avoid manually acquiring and labeling a large amount of data in a positioning area in advance, directly utilizes position estimation to realize automatic data acquisition and labeling, when the environment changes dynamically, a large amount of training data can be obtained again through a short-time positioning process, the construction and updating cost of the database is reduced, and the model mobility is improved; in addition, the positioning method based on the radio frequency environment self-training learning continuously improves the label quality by using the updated position estimation to re-label data in an iteration mode, can quickly obtain a model adaptive to the current environment, further obtains more reliable auxiliary positioning information to calibrate the user positioning result, and improves the precision and robustness of the whole positioning system.
Fig. 5 is a schematic physical structure diagram of an electronic device provided in the present invention, and as shown in fig. 5, the electronic device may include: a processor (processor)510, a communication Interface (Communications Interface)520, a memory (memory)530 and a communication bus 540, wherein the processor 510, the communication Interface 520 and the memory 530 communicate with each other via the communication bus 540. Processor 510 may invoke logic instructions in memory 530 to perform a radio frequency location method comprising: inputting the radio frequency measurement data into a trained non-line-of-sight identification model to obtain a judgment result of the radio frequency measurement data; when the judgment result of the radio frequency measurement data is non-line-of-sight data, inputting the radio frequency measurement data into a trained non-line-of-sight calibration model to obtain calibrated radio frequency measurement data; the trained non-line-of-sight recognition model is obtained by training according to radio frequency measurement sample data carrying a discrimination result label; the trained non-line-of-sight calibration model is obtained by training according to radio frequency measurement sample data carrying a distance tag; the label of the radio frequency measurement sample data is obtained based on radio frequency environment self-training learning.
Furthermore, the logic instructions in the memory 530 may be implemented in the form of software functional units and stored in a computer readable storage medium when the software functional units are sold or used as independent products. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
In another aspect, the present invention also provides a computer program product, the computer program product comprising a computer program stored on a non-transitory computer-readable storage medium, the computer program comprising program instructions, which when executed by a computer, enable the computer to perform the radio frequency positioning method provided by the above methods, the method comprising: inputting the radio frequency measurement data into a trained non-line-of-sight identification model to obtain a judgment result of the radio frequency measurement data; when the judgment result of the radio frequency measurement data is non-line-of-sight data, inputting the radio frequency measurement data into a trained non-line-of-sight calibration model to obtain calibrated radio frequency measurement data; the trained non-line-of-sight recognition model is obtained by training according to radio frequency measurement sample data carrying a discrimination result label; the trained non-line-of-sight calibration model is obtained by training according to radio frequency measurement sample data carrying a distance tag; the label of the radio frequency measurement sample data is obtained based on radio frequency environment self-training learning.
In yet another aspect, the present invention further provides a non-transitory computer-readable storage medium, on which a computer program is stored, the computer program being implemented by a processor to perform the radio frequency positioning method provided in the foregoing embodiments, the method including: inputting the radio frequency measurement data into a trained non-line-of-sight identification model to obtain a judgment result of the radio frequency measurement data; when the judgment result of the radio frequency measurement data is non-line-of-sight data, inputting the radio frequency measurement data into a trained non-line-of-sight calibration model to obtain calibrated radio frequency measurement data; the trained non-line-of-sight recognition model is obtained by training according to radio frequency measurement sample data carrying a discrimination result label; the trained non-line-of-sight calibration model is obtained by training according to radio frequency measurement sample data carrying a distance tag; the label of the radio frequency measurement sample data is obtained based on radio frequency environment self-training learning.
The above-described embodiments of the apparatus are merely illustrative, and the units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.
Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware. With this understanding in mind, the above-described technical solutions may be embodied in the form of a software product, which can be stored in a computer-readable storage medium such as ROM/RAM, magnetic disk, optical disk, etc., and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the methods described in the embodiments or some parts of the embodiments.
Finally, it should be noted that: the above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (10)
1. A radio frequency location method, comprising:
inputting the radio frequency measurement data into a trained non-line-of-sight identification model to obtain a judgment result of the radio frequency measurement data;
when the judgment result of the radio frequency measurement data is non-line-of-sight data, inputting the radio frequency measurement data into a trained non-line-of-sight calibration model to obtain calibrated radio frequency measurement data;
the trained non-line-of-sight recognition model is obtained by training according to radio frequency measurement sample data carrying a discrimination result label; the trained non-line-of-sight calibration model is obtained by training according to radio frequency measurement sample data carrying a distance tag; the label of the radio frequency measurement sample data is obtained based on radio frequency environment self-training learning.
2. The radio frequency location method of claim 1, wherein prior to the step of inputting radio frequency measurement data into the trained non-line-of-sight recognition model, the method further comprises:
analyzing the map information to obtain a grid map; wherein the weight of each grid point in the grid-point map is used for approximating the continuous probability distribution of the user state;
updating the weight of each grid point based on the pedestrian dead reckoning information acquired by the mobile terminal to obtain the updated grid point weight;
matching and aligning the radio frequency measurement data with grid points obtained by the pedestrian dead reckoning information based on a timestamp matching principle to obtain grid points corresponding to each group of radio frequency measurement data;
obtaining a distinguishing result label and a distance label of each group of radio frequency measurement data according to the lattice points and the map information corresponding to each group of radio frequency measurement data, and determining the radio frequency measurement data carrying the distinguishing result label and the radio frequency measurement data carrying the distance label;
and respectively carrying out sample quantity expansion on the radio frequency measurement data carrying the identification result label and the radio frequency measurement data carrying the distance label according to the updated lattice point weight to obtain a first training data set and a second training data set.
3. A radio frequency location method according to claim 2, wherein after the step of obtaining the first training data set and the second training data set, the method further comprises:
s31, training a preset model based on the first training data set and the second training data set respectively to obtain a non-line-of-sight recognition model and a non-line-of-sight calibration model;
s32, performing radio frequency measurement identification on the first training data set through the non-line-of-sight identification model, and performing radio frequency measurement calibration on the second training data set through the non-line-of-sight calibration model to obtain auxiliary positioning information;
s33, updating according to the weight of each grid point of the auxiliary positioning information to obtain a new grid point weight;
s34, according to the new lattice point weight, re-labeling the radio frequency measurement data in the first training data set and the second training data set to obtain a first training enhanced data set and a second training enhanced data set;
and S35, retraining the preset model according to the first training enhanced data set and the second training enhanced data set respectively, and repeating the steps S31 to S35 until preset conditions are met to obtain a trained non-line-of-sight recognition model and a trained non-line-of-sight calibration model.
4. The radio frequency positioning method according to claim 2, wherein the step of analyzing the map information to obtain a grid-point map specifically comprises:
carrying out uniform discretization pretreatment on the map information to obtain an initial grid map;
and performing connectivity analysis on the initial grid-point map, and excluding positions which cannot be reached to obtain a final grid-point map.
5. The radio frequency positioning method according to claim 2, wherein the step of updating the weight of each grid point based on the pedestrian dead reckoning information obtained by the mobile terminal to obtain an updated grid point weight specifically comprises:
obtaining the initialization weight of each grid point according to the grid-point map;
obtaining the prior weight of each grid point in a state prediction equation of Bayesian filtering according to the pedestrian dead reckoning information;
and updating the weight of each grid point according to the prior weight of each grid point to obtain the updated grid point weight.
6. The radio frequency positioning method according to claim 4, wherein the step of performing uniform discretization preprocessing specifically comprises:
wherein p iskAnd thetakRespectively representing the position and the course angle of the user at the k step, and jointly forming a state variable of the system, zkRepresenting the radio frequency measurements taken by the user at step k, the observed variables constituting the system, p (p)k,θk|z1：k) Representing the posterior probability distribution of the k-th round state,
indicates that the k-th wheel state is (i)p,ia) The posterior weight of (a).
7. The radio frequency positioning method according to claim 5, wherein the Bayesian filtering state prediction equation specifically comprises:
wherein,
representing derived location grid points based on indoor map information
Is transferred to
Probability of (m)kThe pedestrian dead reckoning information of the k step comprises the step length l of the k stepkAnd heading angle variation amount Δ θk，
Indicates that the state of the k-1 th round is (i)p,ia) The posterior weight of (a) is calculated,
indicates that the k-th wheel state is (i)p,ia) A priori weight of.
8. The radio frequency location method of claim 1, further comprising: after the step of obtaining calibrated radio frequency measurement data, the method further comprises:
according to the calibrated radio frequency measurement data and a Bayesian filtering state updating equation, obtaining posterior weights of all lattice points, and according to the posterior weights, obtaining the final positioning of the user;
wherein the state update equation is:
wherein, the user positioning result is:
wherein,
is a system observation equation which represents the kth step of the user at a position grid point
To obtain an observed variable zkThe probability of (c).
9. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the steps of the radio frequency location method according to any of claims 1 to 8 are implemented when the processor executes the program.
10. A non-transitory computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the radio frequency location method according to any one of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110614429.XA CN113419211B (en) | 2021-06-02 | 2021-06-02 | Radio frequency positioning method, electronic equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110614429.XA CN113419211B (en) | 2021-06-02 | 2021-06-02 | Radio frequency positioning method, electronic equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113419211A true CN113419211A (en) | 2021-09-21 |
| CN113419211B CN113419211B (en) | 2024-07-16 |
Family
ID=77713616
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110614429.XA Active CN113419211B (en) | 2021-06-02 | 2021-06-02 | Radio frequency positioning method, electronic equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113419211B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115134746A (en) * | 2022-07-06 | 2022-09-30 | 国能龙源电气有限公司 | Method and system for establishing interference identification model library of UWB positioning system |
| CN115237790A (en) * | 2022-08-01 | 2022-10-25 | 青岛柯锐思德电子科技有限公司 | NLOS signal identification and acquisition method and system of UWB |
| CN118509990A (en) * | 2024-07-19 | 2024-08-16 | 中国电信股份有限公司 | Indoor positioning method, indoor positioning device, computer equipment, readable storage medium and program product |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20110177786A1 (en) * | 2009-03-30 | 2011-07-21 | Stefano Marano | Method and system for identification and mitigation of errors in non-line-of-sight distance estimation |
| CN109188410A (en) * | 2018-08-27 | 2019-01-11 | 清华大学 | A kind of range calibration method, device and equipment under non line of sight scene |
| CN109302690A (en) * | 2018-09-30 | 2019-02-01 | 桂林电子科技大学 | A kind of non line of sight indoor orientation method based on optimization Kalman filtering |
| CN110334788A (en) * | 2019-07-08 | 2019-10-15 | 北京信息科技大学 | Distributed multi-antenna reader positioning system and its method based on deep learning |
-
2021
- 2021-06-02 CN CN202110614429.XA patent/CN113419211B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20110177786A1 (en) * | 2009-03-30 | 2011-07-21 | Stefano Marano | Method and system for identification and mitigation of errors in non-line-of-sight distance estimation |
| CN109188410A (en) * | 2018-08-27 | 2019-01-11 | 清华大学 | A kind of range calibration method, device and equipment under non line of sight scene |
| CN109302690A (en) * | 2018-09-30 | 2019-02-01 | 桂林电子科技大学 | A kind of non line of sight indoor orientation method based on optimization Kalman filtering |
| CN110334788A (en) * | 2019-07-08 | 2019-10-15 | 北京信息科技大学 | Distributed multi-antenna reader positioning system and its method based on deep learning |
Non-Patent Citations (2)
| Title |
|---|
| 周牧 等: "基于Wi-Fi即时定位与映射像素模板匹配的室内运动地图构建与定位", 电子与信息学报, vol. 40, no. 05, 31 May 2018 (2018-05-31), pages 1050 - 1058 * |
| 张熠 等: "一种结合众包的WIFI指纹数据库自构建及更新算法", 小型微型计算机系统, no. 11, 15 November 2017 (2017-11-15), pages 2458 - 2464 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115134746A (en) * | 2022-07-06 | 2022-09-30 | 国能龙源电气有限公司 | Method and system for establishing interference identification model library of UWB positioning system |
| CN115237790A (en) * | 2022-08-01 | 2022-10-25 | 青岛柯锐思德电子科技有限公司 | NLOS signal identification and acquisition method and system of UWB |
| CN115237790B (en) * | 2022-08-01 | 2023-04-28 | 青岛柯锐思德电子科技有限公司 | UWB NLOS signal identification and acquisition method and system |
| CN118509990A (en) * | 2024-07-19 | 2024-08-16 | 中国电信股份有限公司 | Indoor positioning method, indoor positioning device, computer equipment, readable storage medium and program product |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113419211B (en) | 2024-07-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Rao et al. | Real-time monitoring of construction sites: Sensors, methods, and applications | |
| CN111919476B (en) | Indoor positioning method, server and positioning system | |
| Park et al. | Self-corrective knowledge-based hybrid tracking system using BIM and multimodal sensors | |
| CN106714110B (en) | Wi-Fi position fingerprint map automatic construction method and system | |
| US20220172594A1 (en) | Tracking and alert method and system for worker productivity and safety | |
| Seco et al. | Improving RFID-based indoor positioning accuracy using Gaussian processes | |
| EP2992355B1 (en) | Scalable real-time location detection based on overlapping neural networks | |
| EP3052957B1 (en) | Enhanced indoor localization method | |
| CN113419211B (en) | Radio frequency positioning method, electronic equipment and storage medium | |
| CN103176164B (en) | Based on the Multi-Target Passive Tracking method of wireless sensor network | |
| WO2017121168A1 (en) | Cluster-based magnetic positioning method, device and system | |
| CN105636201B (en) | Indoor orientation method based on sparse signal fingerprint database | |
| CN109212476B (en) | RFID indoor positioning algorithm based on DDPG | |
| CN104185270A (en) | Indoor positioning method, system and positioning platform | |
| CN103491627B (en) | A kind of closely real-time accurate positioning method of integrated many algorithms | |
| CN109164411A (en) | A kind of personnel positioning method based on multi-data fusion | |
| WO2015134448A1 (en) | Real-time location detection using exclusion zones | |
| CN103298156A (en) | Passive multi-target detecting and tracking method based on wireless sensor networks | |
| CN106231552A (en) | A kind of noncooperative target localization method based on hybrid mode | |
| Seco et al. | RFID-based centralized cooperative localization in indoor environments | |
| US11519734B2 (en) | Facility management based on cumulative device location data | |
| CN114758364B (en) | Industrial Internet of things scene fusion positioning method and system based on deep learning | |
| Zheng et al. | Adaptive device-free passive localization coping with dynamic target speed | |
| Seco et al. | Autocalibration of a wireless positioning network with a FastSLAM algorithm | |
| CN114245298B (en) | Indoor personnel marking and positioning method based on pressure sensing |
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 |