Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Indoor 3D positioning method based on generalized regression neural network

A generalized regression and neural network technology, applied in the field of indoor 3D positioning based on generalized regression neural network, can solve the problems of inflexible coordination between global and local search capabilities, optimization capabilities and efficiency to be improved, and low positioning accuracy. Achieve the effect of solving over-fitting phenomenon, solving poor prediction effect and high positioning accuracy

Pending Publication Date: 2022-03-11
XUZHOU COLLEGE OF INDAL TECH
View PDF1 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, the above four improvement methods do not improve the location update strategy, and the coordination of global and local search capabilities is not flexible enough, and the optimization ability and optimization efficiency need to be improved
Therefore, it is known that the positioning method based on relative position in the prior art is easily disturbed by the environment, and the positioning accuracy is not high

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Indoor 3D positioning method based on generalized regression neural network
  • Indoor 3D positioning method based on generalized regression neural network
  • Indoor 3D positioning method based on generalized regression neural network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0100] Refer Figure 1 - Figure 16 The present invention provides an indoor 3D positioning method based on a generalized regression neural network, including the following steps:

[0101] Step 1: Get the intensity indication value of the receiving signal of the test site, determine the range of the test site according to the intensity indication value.

[0102] Step 2: Collect the three-dimensional coordinate sample data of the test nodes in the test site, train and establish a generalized regression neural network model.

[0103] Step 3: Optimize the gray wolf algorithm, the improved gray wolf optimization algorithm IGWO adjusts the flow chart of smoothing factor Image 6 As shown, the smooth factor is adjusted using the optimized gray wolf algorithm to obtain the step of optimal optimization smooth factor σ, including:

[0104] Step 301: Using the good point set method to initialize the gray wolf population of gray wolf algorithm, obtain the steps of initial group distribution of ...

Embodiment 2

[0149] In the traditional gray wolf algorithm, the initial wolf population is randomly generated, the random distribution of population makes the initial individual cannot traverse the entire search space, causing the loss of better solution, may also fall into local optimal value, lowered The optimization ability of the algorithm. The initial population having a uniform distribution can improve the optimization effect of the algorithm to some extent. The present invention adopts a combination of a combination method to initialize the gray wolf population to obtain a more uniform and stable initial population distribution.

[0150] The number of initial groups in the wolf group is N. The dimension of the search space is D.

[0151] rim d = E d 1 ≤ D ≤ D (25)

[0152] The Dimension of the Item of the Centralized Point Point is expressed as:

[0153]

[0154] Put the best point to the initial population of the gray wolf, the search space is 1 dimension, then the value of the 1-dim...

Embodiment 3

[0162] The experiment is selected in a 12m × 10m office ring, based on the ZigBee positioning system, prepare 5 Corporate Nodes (APs), and placed in four corners and center positions of the test field ceiling, the data used to receive the positioning node There are various obstacles in the test site. The test area is divided into 120 small areas, each small area size of 1 m × 1m, and the center position of the small area is used as the test point. The tester carries the positioning node with you, from the (0,0) position, traversing each test small area. At each test point, the positioning node sends 100 packets to the 5 trick nodes, and forms [RSSI 1 , RSSI 2 , RSSI 3 , RSSI 4 , RSSI 5 (x i Y i ,z i )] Sample data, i = 1, 2, 3, ..., 120, as sample set data. 100 sets of data from 120 sets of sample data as a training sample set, and the remaining 20 groups are set as a test sample set. The RSSI value of the training sample set is used as the input of the network model. Training the...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses an indoor 3D positioning method based on a generalized regression neural network, relates to the technical field of 3D positioning, and obtains a better fitness value and a faster convergence speed through the improvement of population initialization, the logarithmic adjustment of control parameters and the introduction of a fitness weight through multiple iterations. Meanwhile, the method is introduced into a generalized regression neural network model, an optimal generalized regression neural network model is established by utilizing the value of an optimized smoothing factor sigma, and the three-dimensional position of a node to be measured is accurately positioned. According to the indoor 3D positioning method based on the generalized regression neural network provided by the invention, the improved grey wolf algorithm IGWO is introduced to optimize the smoothing factor sigma of the generalized regression neural network, the poor prediction effect and overfitting phenomena caused by improper manual parameter selection are effectively solved, the prediction precision of the model is improved, and the positioning accuracy of the model is improved. And the positioning precision is higher and the optimization speed is faster.

Description

Technical field [0001] The present invention relates to the field of 3D positioning techniques, and in particular, to a chamber 3D positioning method based on a generalized regression neural network. Background technique [0002] In order to achieve more precise indoor positioning, the artificial intelligence theory of artificial neural network is used in the indoor positioning of the artificial intelligence theory of artificial neural networks. Among them, the generalized regression Neural Network (GRNN) is strong, and the data fitting capacity and learning speed are also better, and it is very suitable for indoor positioning. Its model structure is simple, and the predictive performance is largely dependent on the value of the parameter σ, but the subjective impact is large when σ is valued, and a method of optimizing σ values ​​is required to achieve high-precision positioning. [0003] Greywolfoptimizer, GWO is an emerging smart optimization algorithm with strong convergence ...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Applications(China)
IPC IPC(8): H04W4/02H04W4/021H04W4/33G06N3/00G06N3/04G06N3/08
CPCH04W4/021H04W4/33H04W4/023G06N3/006G06N3/04G06N3/08
Inventor 高媛
Owner XUZHOU COLLEGE OF INDAL TECH
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com