KR101157103B1 - Apparatus and Method for estimating position based on self organization algorithm, and Recording medium thereof - Google Patents
Apparatus and Method for estimating position based on self organization algorithm, and Recording medium thereof Download PDFInfo
- Publication number
- KR101157103B1 KR101157103B1 KR1020100091015A KR20100091015A KR101157103B1 KR 101157103 B1 KR101157103 B1 KR 101157103B1 KR 1020100091015 A KR1020100091015 A KR 1020100091015A KR 20100091015 A KR20100091015 A KR 20100091015A KR 101157103 B1 KR101157103 B1 KR 101157103B1
- Authority
- KR
- South Korea
- Prior art keywords
- node
- self
- learning
- signal
- winner
- Prior art date
Links
Images
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Software Systems (AREA)
- Radar Systems Or Details Thereof (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Data Mining & Analysis (AREA)
- Automation & Control Theory (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Medical Informatics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Remote Sensing (AREA)
- Radar, Positioning & Navigation (AREA)
Abstract
Disclosed are a position measuring apparatus and method based on self-organization. Self-organization-based position measuring apparatus according to an embodiment of the present invention comprises a sensor unit consisting of a plurality of sensors for detecting a signal from a signal generator, such as an active tag attached to the positioning object; Quantify the magnitude and frequency of the signal into the input data of the self-regulation map, but smooth the quantized signal after continuous reception while the object is moving to increase the stability of the input data and refer to the spatial coordinates for self-organization learning. An input data generation module for constructing learning data consisting of input data quantifying a signal transmitted from the learned learning position and the referenced actual spatial coordinates; A learning DB for storing the learning data for self-organizing learning; Self-organization learning reflecting the structure of signal sensing space Creates a node weight for mapping the input data to spatial coordinates for each node on the grid by setting a grid structure of the same or lower dimension as the number of signals, and self-organizing learning In the step, while repeatedly presenting the learning data from the learning DB, it is determined the winner node and the next winner nodes based on the self-organization map learning method for each iterative learning step, and update the node weights of the winner node and the next winner node. It stores the weight of each node based on the learning result and determines the winner node and next-order nodes in response to the input data in the actual position measurement execution step, and measures the actual spatial position from the spatial coordinates corresponding to the winner-node and next-order nodes. Including self-organizing mapping module for coordinate mapping According to embodiments of the present invention, it is possible to increase the reliability of the measurement coordinates by learning the weight of each node by using the learning data collected in the field structure of the position measurement area set to be the movement of water interference minimized.
Description
TECHNICAL FIELD The present invention relates to real-time location tracking, and more particularly, to a location measuring apparatus and method based on self-organization.
Real-time locating systems (RTLS), which have been conventionally implemented, have serious problems of location tracking inaccuracy due to signal distortion of a measurement environment. The main reason is that it is not easy to obtain a reliable signal pattern by the interference effect caused by the movement of a person or obstacle and the structure causing reflection, absorption, and distortion of the signal. In addition, the conventional trigonometric algorithm, which is widely used, does not provide a real-time solution by requiring a large amount of calculation when the number of sensors is large, when the position measurement object moves, and when tracking a large number of objects. have. In particular, in order to apply a tracking algorithm in three-dimensional space, a complex algorithm such as a quadrilateral method has to be applied.
Active Tag, which is widely used in RTLS, communicates with readers to execute commands and to inform one's own information at regular intervals. The reader wirelessly communicates with the access point (AP) that matches the tag and the sink. The reader measures the information and signal strength values transmitted from the tag and transmits the signal to the matching AP.
The syncing AP transmits commands from the server to the reader or sends information from the reader to the server. The server stores the collected tag information and analyzes it to calculate the location of the tag. Therefore, there is a problem in that time synchronization between devices such as tags or access points needs to be precisely corrected. In this case, if the real time calculation is not supported, a synchronization imbalance due to stream accumulation occurs.
When measuring the location inside the warehouse, if there are metal obstacles such as racks in the middle of transmitting and receiving radio signals, the transmission and reception will not be performed properly or the signal will be distorted. To reduce this error, location tracking systems use various filtering algorithms. The Kalman filter, the most representative algorithm, is an efficient calculation method that tracks the direction of time with the noise motion equation in real time, but this alone has a limitation in correcting the calculated position coordinates along with the error distance information.
The first technical problem to be achieved by the present invention is to collect the training data in the state structure of the location measurement space is set in the learning phase to build a self-organizing map for the location measurement to increase the reliability when measuring the actual space location It is to provide a position measuring device based on self-organization capable of position measurement.
The second technical problem to be achieved by the present invention is to provide a position measuring method based on self-organization applied to the position measuring apparatus.
A third technical object of the present invention is to provide a recording medium that can be read by a computer system as a medium on which a program for executing the self-organization-based positioning method in a computer system is recorded.
In order to achieve the first technical problem, a position measuring apparatus based on self-organization according to an embodiment of the present invention includes a plurality of sensors for detecting a signal transmitted from a signal generating apparatus such as an active tag attached to a position measuring object. Sensor unit consisting of; The size and frequency of the signal are quantified by the input data of the self-regulation map, and the training data consists of the input data obtained by quantifying the signal transmitted from the learning position to which the spatial coordinates are referred for self-organization learning and the reference actual spatial coordinates. An input data generation module for constructing a; A learning DB for storing the learning data for self-organizing learning; Self-organization learning reflecting the structure of signal sensing space Creates a node weight for mapping the input data to spatial coordinates for each node on the grid by setting a grid structure of the same or lower dimension as the number of signals, and self-organizing learning In the step, while repeatedly presenting the learning data from the learning DB, it is determined the winner node and the next winner nodes based on the self-organization map learning method for each iterative learning step, and update the node weights of the winner node and the next winner node. It stores the weight of each node based on the learning result and determines the winner node and next-order nodes in response to the input data in the actual position measurement execution step, and measures the actual spatial position from the spatial coordinates corresponding to the winner-node and next-order nodes. Including self-organizing mapping module for coordinate mapping . Preferably, the input data generation module may increase the stability of the input data by smoothing and quantifying the signals continuously received when the object is moving.
In order to achieve the second technical problem, the self-organization based position measurement method according to an embodiment of the present invention, quantified by the size and frequency of the signal from the signal transmitted from the reference position for the self-organization map learning Calculating learning data consisting of one input data and actual spatial coordinates of the referenced position; Set up a grid structure of the same or lower dimension as the number of signals in the signal space to generate node weights for mapping the input data to the spatial coordinates for each node on the grid, and iteratively present the learning data repeatedly and indirectly. The self-organization map which determines the winner node and the next winner node based on the self-organization map learning method at each stage, updates the node weight of the winner node and the next winner node, and constructs a self-organization map for location tracking corresponding to the learning result. Learning phase; In response to the input data quantified according to the signal pattern generated from the signal generator in the actual position measurement execution step, determine the winner node and next-order nodes of the self-organizing map for location tracking, and the corresponding position coordinates of the winner node and next-order nodes. Measuring from the actual space position. In addition, when tracking the trajectory of the moving object, the input signal smoothing step may be performed to smoothly convert the received signals into input data in order to increase continuous positioning accuracy on the trajectory.
According to the embodiments of the present invention, by collecting the learning data of the structure of the location measurement site and the interference of the moving object is minimized and learning at the learning stage, it is possible to increase the reliability and to quickly measure the location and to significantly reduce the real-time calculation amount. Even three-dimensional position tracking is possible. In addition, the accuracy can be further improved by modifying the arrangement of the nodes or by combining the position coordinates of the winner node and the next winner node as the weight using the activation level of the winner node and the next winner node in the actual position tracking step.
1 illustrates an example of a self-organizing map for position measurement.
2A illustrates a side mapping side of a position measuring device based on self-organization according to an embodiment of the present invention.
FIG. 2B illustrates a process in which a setting device for conveniently setting a reference position when generating learning data is added.
3A and 3B illustrate a process of generating input data for each node in an indoor space.
4 illustrates a side calculation side of an output layer based on self-organization according to an embodiment of the present invention.
Hereinafter, with reference to the drawings will be described a preferred embodiment of the present invention. However, embodiments of the present invention illustrated below may be modified in various other forms, and the scope of the present invention is not limited to the embodiments described below.
Embodiments of the present invention construct a learning DB using learning data with the structure of the location measurement space and the interference of moving objects minimized, and construct a self-organizing map for spatial location tracking based on the self-organizing map learning method. In the execution step, Winner and Winner nodes can be determined from the input data quantified by the given signal, and the measurement position can be directly obtained from the corresponding coordinates, or accuracy can be improved by combining by weight adjustment. Provide a method. In addition, when tracking the trajectory of a moving position measurement object, a method of smoothing an input signal of a continuous position is provided to increase the accuracy of the measurement.
A neural network emulates the characteristics of a brain function by a computer, and artificial neurons (nodes) that form a network by synapse coupling change the strength of synapses through learning, and refer to a general model having problem solving ability. Self-Organizing Map (SOM) is a structure in which the information processing system voluntarily remodels and changes the organization in the system based on past experiences and input of information from the outside in order to enhance processing functions. It is a kind of neural network modeled as. Self-organizing maps are learned by self-learning, and usually generate maps with the same or lower dimensions (usually two-dimensional) as the number of input signals. This map is based on a discrete representation of the spatial arrangement of nodes forming the competing layer and preserves the topological properties of the input space. Like other neural networks, SOM operates in two modes: learning and execution. Learning is the process of creating maps from input data, also called vector quantization. In the execution process, when new input data is given, a result of grouping or mapping the neural network according to the learning result is output.
1 illustrates an example of a neural network model based on a self-organization map for location tracking.
In FIG. 1, node i ( i = 1 ,..., I ) is a node of an input layer to which input data is input, and node j ( j = 1 ,..., J ) is a node of a competition layer having a grid structure. W ji denotes the connection strength connecting each node with the input layer and each node in the contention layer. In neural networks, the connection strength is the node weight (node Also called weights). In general, a self-organizing map refers to a neural network composed of input strength and contention layer, and connection strength between input layer and contention layer. Presented. Up to three nodes in the output layer correspond to spatial coordinates (x, y, z). Each node in the contention layer is connected to every node in the input and output layers. In the neural network model based on the self-organizing map for location tracking, when the input data is formed from the signal pattern detected in the space and input to the input layer, the node weight of the competitive layer having the node weight between the input layer and the competition layer is the most similar to the input data. The operation starts by selecting as the winner node and then obtaining the next winner nodes having similar node weights. The similarity between the node weight and the input data between the input layer and the node of the contention layer is the activation level of the contention layer node, and the winner node is the node with the highest level of activation, followed by the nodes of the higher layer. Select with nodes. When the winner node and the next winner node are determined, the position coordinates can be output based on the node weights between the winner node and the next winner node and the output node, and the error is reduced by weighted combination based on the activation level of the winner node and the next winner node. Can be.
Embodiments of the present invention require a learning process. To do this, set the grid structure to reflect the signal space and define points on the grid as nodes. In the learning process, we first initialize all the node weights of the neural network. In the learning phase, the active tag or signal generator is moved in space to generate signals (magnetic fields, RF, infrared rays, ultrasonic waves, etc.) at each reference position, and are detected by the sensor unit and converted into quantified input data such as signal strength and frequency. When the input data is n-dimensional, the input data is represented by P = (p 1 , p 2 ,…, p i ,… p n ) and combined with the reference position (x, y, z) to learn data D = (P , x, y, z) are constructed and stored in the learning DB.
In the above-described embodiment, the learning is repeated until the learning is stabilized. In each repetitive learning step, the learning data D = (P, x, y, z) is randomly selected from the learning DB, and for each jitter node j Similarity S j (P) between the node weight vector from the input layer and the corresponding input data P is obtained. The node with the highest similarity among the nodes in the contention layer is called the winner node and is denoted by the node j * . For each contention node node j, the node weight W ji from the input layer and the node weight W kj to the output layer are updated as shown in
Where t is the iterative learning step and S j (P) is the similarity level between the node weight vector of node j and P. N (j * , j) is a function of neighboring relationship between Winner node j * and node j and decreases as the distance between two nodes increases in the competitive grid structure.N (j * , j * ) = 1 Has a value. Also, α (t, S j (P), N (j * , j)) and β (t, S j (P), N (j * , j)) are similar to the level of repetition and node j. The node weight learning rate determined by the neighbor function value of j and the winner node.
The learning step ends when the node weight converges or when the preset learning time is exceeded.
2A illustrates a position learning aspect of a position measuring apparatus based on self-organization according to an embodiment of the present invention.
The active tag 210 transmits a signal by embedding a battery in itself.
The sensor units 221-224 are configured of a plurality of sensors for measuring the strength of the signal transmitted from the active tag 210. Although the sensor units 221-224 may be configured with only three sensors according to spaces, the case in which four sensors are used in FIG. 2 will be described as an example.
The input data generation module 231 quantifies the signal strength and the frequency of the signal measured by the
Provided that Sig s (1) = Sig (1). Where t is the iterative learning step, Sig (t) is the received signal in the iterative learning step t, Sig s (t) is the signal that is actually applied by smoothing, and λ is the reflectance of the most recent signal. Have
The input data generation module 231 configures the learning data in combination with the input data and the reference position on the referenced real space for learning, and stores the learning data in the learning DB 240.
The self-
Another embodiment of the present invention may further include a position calculator (not shown) for measuring the position of the actual moving object by using the updated node weight. The position calculator determines winners and next winners based on the input data quantifying the strength of the signals measured in real time by the sensor units 221-224 and the frequency of each signal, and uses the winners and the next winners. Calculate the position coordinates of the object.
2B shows an example of a setting device for facilitating a learning step.
The setting device of FIG. 2B is a device in which a sensor and an application are combined with a signal generating device such as an active tag and a laser sensor for referring to a position in a real space. This learning apparatus is not limited to a predefined reference point in space, but can be freely selected to easily select a plurality of signal transmitting positions as a learning point. The setting device 211 is a reference position setting device capable of easily grasping its own position by measuring a distance to the reflective wall by firing a laser beam in all directions in a space having a reflective wall everywhere.
3A and 3B illustrate a process of generating input data from a signal transmitted from a node.
Figure 3a shows the concept of using the estimated distance between the signal strength and the signal arrival time, such as the signal and the receiving sensor as input data.
3b is a concept of receiving a plurality of signals in sequence instead of one from a sensor for consecutively transmitted signals, accumulating the estimated distances between the originating position and the receiving sensor as a histogram and then utilizing the number of signals for each estimated distance section as input data. However, the number of input nodes in the self-organizing map increases in proportion to the number of histogram sections. However, it has the advantage of enabling precise learning. For example, if one element of the input data of the self-organizing map is the distance between the originating device and the specific sensor, and this is a conventional concept of inputting to a specific node i of the input layer of the self-organizing map, the histogram-type input layer structure is the specific type. The number of input nodes (node i1, node i2, node i3,…, node im when the number of histogram intervals is m) for the sensor is used, and each input node has the number of signals in the corresponding interval. Is input. In FIG. 3B, when m = 5 and a histogram section is given as a 5m section, a 10m section, a 15m section, a 20m section, and a 25m section, five input node structures corresponding to the specific sensor are illustrated.
4 illustrates the concept of estimating the actual position from the winner node j * and the next-order nodes j ', j'',j''' of the competition layer and a combination thereof when estimating the position of the actual object.
As an example of the present invention, when input data P is given, it is assumed that winner node j * and next-order nodes j ', j'',j''' are determined, and the activation level of these nodes is S j * (P), S j (P), S j '' (P), S j ''' (P), and the node weight from these nodes to the output node k is W kj *, W kj' , W kj '' , W kj ''' can be estimated by Equation 3 by their linear combination.
An algorithm such as Equation 3 may also be used in the above-described position calculator (not shown).
The invention can be implemented via software. Preferably, a program for executing a self-organization based position measuring method according to embodiments of the present invention may be provided by recording a program for executing in a computer on a computer-readable recording medium. When implemented in software, the constituent means of the present invention are code segments that perform the necessary work. The program or code segments may be stored on a processor readable medium or transmitted by a computer data signal coupled with a carrier on a transmission medium or network.
A computer-readable recording medium includes all kinds of recording apparatuses in which data that can be read by a computer system is stored. Examples of the computer readable recording medium include ROM, RAM, CD-ROM, DVD 占 ROM, DVD-RAM, magnetic tape, floppy disk, hard disk, optical data storage, and the like. The computer readable recording medium can also be distributed over network coupled computer devices so that the computer readable code is stored and executed in a distributed fashion.
Although the present invention has been described with reference to one embodiment shown in the drawings, this is merely exemplary, and it will be understood by those skilled in the art that various modifications and variations may be made therefrom. And, such modifications should be considered to be within the technical protection scope of the present invention. Therefore, the true technical protection scope of the present invention will be defined by the technical spirit of the appended claims.
Claims (10)
Input data is calculated by quantifying at least one of the measured signal strength or the frequency of each signal, generating node weights of the self-organizing map for location tracking for each node of the grid structure reflecting the location measurement space, An input data generation module for constructing learning data by combining reference positions in space;
Whenever the training data is randomly presented, a self-organization map for location tracking is generated by determining a winner node and a next winner node based on a self-organization map learning method and updating the node weights of the winner node and the next winner node. A self organizing mapping module; And
Self-organization based position measurement device comprising a learning database for storing the training data.
The input data generation module
When the signal generator is in motion, using the measured signal and the signal smoothed in the previous iterative learning step, a smoothed signal of the current iterative learning step is obtained and converted into the input data. Based positioning device.
The self-organization map for location tracking
And an input layer, a contention layer, and an output layer for mapping the input data to spatial coordinates.
The sensor unit
A device for measuring position based on self-organization, characterized by measuring the intensity of at least one of a magnetic field, an RF signal, an infrared ray or an ultrasonic wave.
Input data is calculated by quantifying at least one of the measured signal strength or the frequency of each signal, generating node weights of the self-organizing map for location tracking for each node of the grid structure reflecting the location measurement space, An input data generation module for constructing learning data by combining reference positions in space;
Whenever the training data is randomly presented, a self-organization map for location tracking is generated by determining a winner node and a next winner node based on a self-organization map learning method and updating the node weights of the winner node and the next winner node. A self organizing mapping module;
A learning database for storing the learning data; And
The input unit quantifying the strength of the signal and the frequency of each signal measured in real time by the sensor unit is applied to the self-organization map for location tracking to determine the winner node and the next winner node, and the winner node and the next winner node. And a position calculator for calculating a current position coordinate of the signal generator by using the self-organization based position measurement device.
The sensor unit
And measuring the intensity of at least one of a magnetic field, an RF signal, an infrared ray, or an ultrasonic wave transmitted from the signal generator in real time.
Calculating, by the position measuring device, input data by quantifying at least one of the intensity of the measured signal or the frequency of each signal;
Generating, by the positioning device, node weights of the self-organizing map for location tracking for each node of the grid structure reflecting the positioning space, and combining learning data with reference data in the space; And
When the position measurement device randomly presents the learning data, the position measurement device determines the winner node and the next winner nodes based on the self-organization map learning method, and updates the node weights of the winner nodes and the next winner nodes. Creating a Self-Organization Map
Including, self-organization based position measurement method.
The step of calculating the input data
When the signal generator is in motion, using the measured signal and the signal smoothed in the previous iteration learning step to obtain a smoothed signal of the current iteration learning step, characterized in that the step of converting to the input data, magnetic Location measurement based on organization.
The position measuring device calculates input data by quantifying at least one of the measured signal strength or the frequency of each signal, and generates node weights of the self-organizing map for position tracking for each node of the grid structure reflecting the position measuring space. Constructing learning data by combining input data and a reference position in space;
When the position measurement device randomly presents the learning data, the position measurement device determines the winner node and the next winner nodes based on the self-organization map learning method, and updates the node weights of the winner nodes and the next winner nodes. Generating a self-organizing map;
Determining, by the position measuring device, a winner node and a next winner node by applying input data quantifying the strength of a signal measured in real time by the sensor unit and the frequency of each signal to the self-organizing map for location tracking; And
And calculating, by the position measuring device, the current position coordinates of the signal generator using the winner node and the next winner nodes.
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| KR20090087707 | 2009-09-16 | ||
| KR1020090087707 | 2009-09-16 |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| KR20110030397A KR20110030397A (en) | 2011-03-23 |
| KR101157103B1 true KR101157103B1 (en) | 2012-06-22 |
Family
ID=43936068
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| KR1020100091015A KR101157103B1 (en) | 2009-09-16 | 2010-09-16 | Apparatus and Method for estimating position based on self organization algorithm, and Recording medium thereof |
Country Status (1)
| Country | Link |
|---|---|
| KR (1) | KR101157103B1 (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20150060456A (en) * | 2013-11-26 | 2015-06-03 | 한국전자통신연구원 | Apparatus and method for generating radio fingerprint map |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111061273B (en) * | 2019-12-26 | 2023-06-06 | 航天时代(青岛)海洋装备科技发展有限公司 | Autonomous obstacle avoidance fusion method and system for unmanned ship |
| CN113093162B (en) * | 2021-04-14 | 2022-04-01 | 国能智慧科技发展(江苏)有限公司 | Personnel trajectory tracking system based on AIOT and video linkage |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2006065703A (en) * | 2004-08-30 | 2006-03-09 | Inst Of Systems Information Technologies Kyushu | Self-position estimation device, self-position estimation method, program capable of executing self-position estimation method by computer, and recording medium recording program |
| KR100801246B1 (en) * | 2006-12-26 | 2008-02-04 | 부산대학교 산학협력단 | System and method for managing of tag locations |
| KR20090022829A (en) * | 2007-08-31 | 2009-03-04 | 이커스텍(주) | Real-time position recognition method and system in a ubiquitous environment |
| KR20090088516A (en) * | 2008-02-15 | 2009-08-20 | 한국과학기술연구원 | Method for self-localization of a robot based on object recognition and environment information around the recognized object |
-
2010
- 2010-09-16 KR KR1020100091015A patent/KR101157103B1/en active IP Right Grant
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2006065703A (en) * | 2004-08-30 | 2006-03-09 | Inst Of Systems Information Technologies Kyushu | Self-position estimation device, self-position estimation method, program capable of executing self-position estimation method by computer, and recording medium recording program |
| KR100801246B1 (en) * | 2006-12-26 | 2008-02-04 | 부산대학교 산학협력단 | System and method for managing of tag locations |
| KR20090022829A (en) * | 2007-08-31 | 2009-03-04 | 이커스텍(주) | Real-time position recognition method and system in a ubiquitous environment |
| KR20090088516A (en) * | 2008-02-15 | 2009-08-20 | 한국과학기술연구원 | Method for self-localization of a robot based on object recognition and environment information around the recognized object |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20150060456A (en) * | 2013-11-26 | 2015-06-03 | 한국전자통신연구원 | Apparatus and method for generating radio fingerprint map |
| KR101694521B1 (en) * | 2013-11-26 | 2017-01-10 | 한국전자통신연구원 | Apparatus and method for generating radio fingerprint map |
Also Published As
| Publication number | Publication date |
|---|---|
| KR20110030397A (en) | 2011-03-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Yang et al. | Efficient object localization using sparsely distributed passive RFID tags | |
| Ma et al. | Fusion of RSS and phase shift using the Kalman filter for RFID tracking | |
| CN103093453B (en) | Calculating can change posture and/or the shape of entity | |
| US10466348B2 (en) | Position acquisition method and apparatus | |
| WO2017121168A1 (en) | Cluster-based magnetic positioning method, device and system | |
| Sun et al. | Device-free wireless localization using artificial neural networks in wireless sensor networks | |
| Xu et al. | AdaRF: Adaptive RFID-based indoor localization using deep learning enhanced holography | |
| Soltani et al. | Enhancing Cluster-based RFID Tag Localization using artificial neural networks and virtual reference tags | |
| Srinivasan et al. | A survey of sensory data boundary estimation, covering and tracking techniques using collaborating sensors | |
| Meyer-Delius et al. | Using artificial landmarks to reduce the ambiguity in the environment of a mobile robot | |
| Alonso et al. | Enhanced WiFi localization system based on Soft Computing techniques to deal with small-scale variations in wireless sensors | |
| KR101157103B1 (en) | Apparatus and Method for estimating position based on self organization algorithm, and Recording medium thereof | |
| Zhou et al. | Application of backpropagation neural networks to both stages of fingerprinting based WIPS | |
| Ali et al. | A deep learning model implementation based on rssi fingerprinting for lora-based indoor localization | |
| Abdel-Hadi | Efficient artificial intelligence-based localization algorithms for Wireless Sensor Networks | |
| Suneetha et al. | A Wireless Detector Network for Three-Dimensional Positioning Using Artificial Neural Networks | |
| Shenoy et al. | A lightweight ANN based robust localization technique for rapid deployment of autonomous systems | |
| Shareef et al. | Localization using extended Kalman filters in wireless sensor networks | |
| Azenha et al. | A neural network approach for radio frequency based indoors localization | |
| Bhatia et al. | Different localization techniques for real time location sensing using passive RFID | |
| JP6921332B2 (en) | Systems and methods that use multiple modality for occupancy perception | |
| Gai et al. | Multi-group localization problem of service robots based on hybrid external localization algorithm with application to shopping mall environment | |
| CN114423077B (en) | Intelligent super-surface-based 4D endogenous safety indoor positioning method and positioning system thereof | |
| Yang et al. | SLAM algorithm for 2D object trajectory tracking based on RFID passive tags | |
| Weerasinghe et al. | Towards IoT; Comparison of RSS based indoor localization using supervised learning and trilateration in WSN |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| A201 | Request for examination | ||
| E902 | Notification of reason for refusal | ||
| E701 | Decision to grant or registration of patent right | ||
| GRNT | Written decision to grant | ||
| FPAY | Annual fee payment |
Payment date: 20150602 Year of fee payment: 4 |
|
| FPAY | Annual fee payment |
Payment date: 20160602 Year of fee payment: 5 |
|
| FPAY | Annual fee payment |
Payment date: 20170601 Year of fee payment: 6 |
|
| FPAY | Annual fee payment |
Payment date: 20180531 Year of fee payment: 7 |
|
| FPAY | Annual fee payment |
Payment date: 20190530 Year of fee payment: 8 |