CN111854745A - Clock prediction method based on Internet of things indoor positioning - Google Patents

Clock prediction method based on Internet of things indoor positioning Download PDF

Info

Publication number
CN111854745A
CN111854745A CN202010703571.7A CN202010703571A CN111854745A CN 111854745 A CN111854745 A CN 111854745A CN 202010703571 A CN202010703571 A CN 202010703571A CN 111854745 A CN111854745 A CN 111854745A
Authority
CN
China
Prior art keywords
clock
internet
things
parameters
neural network
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.)
Withdrawn
Application number
CN202010703571.7A
Other languages
Chinese (zh)
Inventor
王永贵
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Guangzhou Daoyuan Information Technology Co Ltd
Original Assignee
Guangzhou Daoyuan Information Technology Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Guangzhou Daoyuan Information Technology Co Ltd filed Critical Guangzhou Daoyuan Information Technology Co Ltd
Priority to CN202010703571.7A priority Critical patent/CN111854745A/en
Publication of CN111854745A publication Critical patent/CN111854745A/en
Withdrawn legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/10—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 by using measurements of speed or acceleration
    • G01C21/12—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 by using measurements of speed or acceleration executed aboard the object being navigated; Dead reckoning
    • G01C21/16—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 by using measurements of speed or acceleration executed aboard the object being navigated; Dead reckoning by integrating acceleration or speed, i.e. inertial navigation
    • G01C21/165—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 by using measurements of speed or acceleration executed aboard the object being navigated; Dead reckoning by integrating acceleration or speed, i.e. inertial navigation combined with non-inertial navigation instruments
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/20—Instruments for performing navigational calculations
    • G01C21/206—Instruments for performing navigational calculations specially adapted for indoor navigation
    • 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
    • G01S19/00—Satellite radio beacon positioning systems; Determining position, velocity or attitude using signals transmitted by such systems
    • G01S19/38—Determining a navigation solution using signals transmitted by a satellite radio beacon positioning system
    • G01S19/39—Determining a navigation solution using signals transmitted by a satellite radio beacon positioning system the satellite radio beacon positioning system transmitting time-stamped messages, e.g. GPS [Global Positioning System], GLONASS [Global Orbiting Navigation Satellite System] or GALILEO
    • G01S19/42—Determining position
    • G01S19/45—Determining position by combining measurements of signals from the satellite radio beacon positioning system with a supplementary measurement
    • G01S19/47—Determining position by combining measurements of signals from the satellite radio beacon positioning system with a supplementary measurement the supplementary measurement being an inertial measurement, e.g. tightly coupled inertial
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • 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
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • 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
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/048—Activation functions
    • 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
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00—Network arrangements or protocols for supporting network services or applications
    • H04L67/01—Protocols
    • H04L67/12—Protocols specially adapted for proprietary or special-purpose networking environments, e.g. medical networks, sensor networks, networks in vehicles or remote metering networks

Landscapes

  • Engineering & Computer Science (AREA)
  • Remote Sensing (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Artificial Intelligence (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Computing Systems (AREA)
  • Evolutionary Computation (AREA)
  • General Health & Medical Sciences (AREA)
  • General Engineering & Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Software Systems (AREA)
  • Mathematical Physics (AREA)
  • Computational Linguistics (AREA)
  • Biomedical Technology (AREA)
  • Molecular Biology (AREA)
  • Biophysics (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Automation & Control Theory (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Evolutionary Biology (AREA)
  • Medical Informatics (AREA)
  • Signal Processing (AREA)
  • Position Fixing By Use Of Radio Waves (AREA)

Abstract

The invention belongs to the technical field of position location, and particularly relates to a clock prediction method based on internet of things indoor location. The system comprises an Internet of things receiving module, a clock data processing module and a clock data processing module, wherein the Internet of things receiving module is used for collecting satellite navigation parameters and observation parameters of a global navigation satellite system, and an Internet of things receiving device is sent to the clock data processing module through an Internet communication system; the clock data processing module extracts satellite navigation parameters and observation parameters to perform parameter preprocessing, and determines system predicted orbits and the like after eliminating a global navigation satellite system clock through the positioning parameters of the broadcast ephemeris. According to the clock prediction method based on the Internet of things indoor positioning, the convolutional neural network is fully utilized to carry out more accurate prediction of the nonlinear time sequence characteristic through learning of a large number of samples of clock prediction data of the Internet of things positioning system, and the positioning precision of the Internet of things system is really improved.

Description

Clock prediction method based on Internet of things indoor positioning
Technical Field
The invention belongs to the technical field of position location, and particularly relates to a clock prediction method based on internet of things indoor location.
Background
Nowadays, science and technology rapidly develop, each field shows a trend of vigorous development, mobile communication is advancing from '4G' to '5G', and intelligent tools such as intelligent robots and unmanned vehicles are popularized, so that the life of people is more colorful. "location" as a kind of service is also beginning to be integrated into people's lives, and has prompted the generation of many location-dependent services, such as GPS positioning-based related services, including outdoor driving navigation, indoor shopping mall inquiry, etc., which involve the corner falling and touching aspects of life, and no positioning technology is utilized. Under the strong demands of promotion and living of positioning technology, the construction of complete and accurate outdoor and indoor positioning systems is urgent and becomes a research hotspot of modern scholars. The rapid development of indoor positioning is hindered by various reasons, mainly because the indoor environment is different from the outdoor environment, the indoor arrangement is complex, the transmission of signals is hindered by furniture, walls and various decorations, the signals are refracted, reflected and even buried, and various physical characteristics of the transmitted signals are seriously damaged. Therefore, it is necessary to solve the problems of inaccurate indoor positioning or even incapability of indoor positioning due to complicated indoor arrangement. The existing indoor positioning technology based on Bluetooth, WiFi, inertial navigation and the like is widely applied to indoor positioning, but the positioning accuracy is not very high. The ultra-wideband positioning technology has a bandwidth of several hundred megahertz, is unique in a plurality of wireless positioning technologies by virtue of super-strong wall penetrating capability, anti-interference capability, multi-path resolution capability and the like, and becomes a competitive technology for indoor positioning.
Disclosure of Invention
The invention aims to provide a clock prediction method based on the internet of things indoor positioning, which utilizes different types of clock synchronization reference networks based on a global navigation satellite system to detect a nonlinear time sequence according to the learning of a convolutional neural network on detection samples in the internet of things, utilizes clock parameters after task processing to realize the intelligent prediction of an indoor intelligent equipment satellite clock, and finally assists an internet of things user to obtain a high-level positioning result.
The purpose of the invention is realized as follows:
a clock prediction method based on Internet of things indoor positioning comprises the following steps:
step 1, an Internet of things receiving module collects satellite navigation parameters and observation parameters of a global navigation satellite system, and an Internet of things receiving device sends the satellite navigation parameters and the observation parameters to a clock data processing module through an Internet communication system;
step 2, a clock data processing module extracts satellite navigation parameters and observation parameters to perform parameter preprocessing, and the clock data processing module determines a system predicted orbit after eliminating a global navigation satellite system clock through the positioning parameters of a broadcast ephemeris;
step 3, the clock data processing module establishes a super-bandwidth nonlinear clock estimation observation model through a system prediction orbit, and gradually solves the ambiguity parameter of indoor positioning of the Internet of things according to an ambiguity integer constraint form;
Step 4, predicting clock parameters of the global navigation satellite system based on the ambiguity data in the step 3, and repeatedly executing the step 3 and the step 4 until the residual error value predicted by the clock is controlled within the precision threshold range;
step 5, constructing a convolutional neural network with a nonlinear structure, and training the convolutional neural network through the generated clock parameters of the global navigation satellite system;
and 6, predicting real-time clock parameters positioned in the Internet of things room through a convolutional neural network training result, and sending the predicted clock parameters to an authorized device in the Internet of things range through an Internet communication system.
The step 1 of collecting the satellite navigation parameters and the observation parameters comprises the following steps:
(1.1) Internet of things receiving module acquisition tlObservation parameter matrix G (t) of timel)；
(1.2) the receiving module of the Internet of things marks the time parameter;
(1.3) setting a clock state transition matrix phi by the receiving module of the Internet of things;
(1.4) the Internet communication system issues observation noise;
(1.5) Internet communication System estimation tlObservation parameter b (t) at timel)：
Figure BDA0002593782750000021
Wherein
Figure BDA0002593782750000031
l is a time parameter index, a0Representing the quantity of state to be estimated, t0At an initial time, phi (t)l,t0) Represents t0Time tlA phase difference function of the time of day.
The clock prediction method based on the indoor positioning of the Internet of things is characterized in that the satellite navigation parameters are calculated as follows:
(1.6) acquiring the three-dimensional position r of the global navigation satellite system in an inertial coordinate system0Speed s1Acceleration s2；
(1.7) collecting a dynamic parameter q of the global navigation satellite system by an Internet of things receiving module;
(1.8) calculating an Earth gravity constant HN by an Internet communication systeme；
(1.9) the receiving module of the Internet of things calculates the sum a of various perturbation forces acting on the global navigation satellite system1；
The satellite navigation parameters of the global navigation satellite system are as follows:
Figure BDA0002593782750000032
the establishing of the super-bandwidth nonlinear clock estimation observation model in the step 3 comprises the following steps:
(3.1) a clock data processing module acquires pseudo-range observed values and phases Q and K of a global navigation satellite system;
(3.2) the clock data processing module acquires the clock parameters of the global navigation satellite system
Figure BDA0002593782750000033
(3.3) estimating the clock error of the global navigation satellite system receiver by the clock data processing module
Figure BDA0002593782750000034
(3.4) confirming the communication wavelength lambda of the Internet communication systemIF；
(3.5) calculating ambiguity parameter M of global navigation satellite systemIF；
(3.6) acquisition of tropospheric projection function ntropAnd zenith troposphere area S;
(3.7) Carrier phase of Global navigation satellite System [ phi ]L,IFAnd pseudorange observation noise phiP,IF；
The super-bandwidth nonlinear clock estimation observation model comprises the following steps:
Figure BDA0002593782750000041
Figure BDA0002593782750000042
in order to effectively fix the ambiguity parameters of the global navigation satellite system, the ambiguity parameters are decomposed and fixed step by step.
Figure BDA0002593782750000043
Wherein w1And w2Respectively representing the frequencies, M, of carrier 1 and carrier 2 of the Internet of things1And M2Respectively representing ambiguity parameters of carrier 1 and carrier 2 of the Internet of things;
and fixing ambiguity parameters of the carrier 1 and the carrier 2 of the Internet of things:
Figure BDA0002593782750000044
wherein L is1And L2Respectively representing phase observations, P, of carrier 1 and carrier 2 of the Internet of things1And P2Respectively representing pseudo-range observed values of carrier 1 and carrier 2 of the Internet of things;
will M1-M2Solving ambiguity M in the retrospective clock estimation observation model1The floating point solution of (2).
The training of the convolutional neural network comprises:
(5.1) inputting a training sample set and a testing sample set; scaling any input image in a bilinear interpolation mode, and performing random horizontal turning, random translation, random image rotation and random image scaling on each image;
(5.2) adopting a plurality of loss functions as the training target function;
(5.3) ordering the loss functions;
(5.4) repeatedly carrying out dynamic target training on the deep neural network on the training sample set in sequence according to the sorted loss functions to obtain a recognition model;
and 5.5, classifying the input test sample according to the recognition model.
Wherein the net input S of the jth neuronjComprises the following steps:
Figure BDA0002593782750000051
through the excitation function f (·), the net output of the j-th neuron is obtained as:
yi＝f(Sj)
Wherein f (-) is Sigmoid function, yjRepresenting the j-th neuron net output.
Wherein the neural network sets the connection weight coefficient of the input layer and the hidden layer as vjThe connection weight coefficient of the hidden layer and the output layer is wjWhen the input is (v)i,ωi) When i is 0,1,2, …, n, the hidden layer neuron node output is:
Figure BDA0002593782750000052
Figure BDA0002593782750000053
and determining a model of the neural network according to the two formulas.
The neural network learning includes:
inputting p test samples to the neural network to obtain output values
Figure BDA0002593782750000054
Obtaining an error function E of the p samplepComprises the following steps:
Figure BDA0002593782750000055
wherein
Figure BDA0002593782750000056
For the desired output, the global error function E is, for all p samples:
Figure BDA0002593782750000057
where η ∈ (0,1) represents the learning rate of the neural network.
Preferably, the clock prediction method based on the internet of things indoor positioning is characterized in that: through the convolutional neural network training result, predict the real-time clock parameter of the indoor location of thing networking to the in-process of the device of authorizing in internet of things scope is sent the clock parameter of prediction through internet communication system, still include:
receiving a previous data packet related to the indoor positioning of the Internet of things, acquiring the arrival time of the previous data packet based on the Internet of things, and predicting the arrival time of the next data packet based on the Internet of things based on the convolutional neural network;
When the next data packet does not arrive at the predicted arrival time, automatically generating a new data packet, and performing smooth processing on clock information carried in the new data packet to enable the output clock signals to be uniformly distributed in a time domain;
when the next data packet arrives at the predicted arrival time, performing smoothing processing on clock information carried in the next data packet to enable output clock signals to be uniformly distributed in a time domain;
converting all output clock signals from digital signals into analog signals, so that all the output clock signals meet preset jitter and drift conditions;
time division multiplexing the clock signals meeting the preset jitter and drift conditions in a packet switching network, and outputting and displaying the time division multiplexing result;
and the previous data packet, the next data packet and the new data packet all carry clock information.
The invention has the beneficial effects that:
according to the clock prediction method based on the Internet of things indoor positioning, the convolutional neural network is fully utilized to carry out more accurate prediction of nonlinear time sequence characteristics through learning of a large number of samples of clock prediction data of the Internet of things positioning system, clock parameters of a global navigation satellite system are calculated through data of the Internet of things receiving module, the neural network is utilized to realize prediction of the clock parameters, and the clock parameters are broadcasted to an authorized device in the range of the Internet of things to correct clock errors of the global navigation satellite system, so that the positioning accuracy of the Internet of things system is really improved.
The clock prediction method based on the indoor positioning of the Internet of things can predict the clock more accurately based on the clock parameters of the global navigation satellite system through the artificial intelligent learning of the convolutional neural network, thereby improving the positioning accuracy of the Internet of things system.
Drawings
Fig. 1 is a schematic diagram of an embodiment of a clock prediction method based on internet of things indoor positioning, which applies the invention.
Detailed Description
The invention is further described below with reference to the accompanying drawings.
The invention provides a clock prediction method based on internet of things indoor positioning, which can more accurately predict the nonlinear time sequence characteristic by fully utilizing the learning of a convolutional neural network on a large number of samples of data running in an authorization device in the internet of things, realizes the prediction of a precise clock of the internet of things device through a posterior clock of the internet of things device, and broadcasts the precise clock to an authorization client side in the internet of things for correcting a clock error so as to fulfill the aim of performing high-level positioning on the device for managing the internet of things. The method simultaneously adopts the convolutional neural network and the post precision clock parameter estimation to ensure the clock prediction precision of the satellite navigation system. The invention discloses a clock prediction method for indoor positioning of an internet of things, which comprises the following steps:
Step 1, an Internet of things receiving module collects satellite navigation parameters and observation parameters of a global navigation satellite system, and an Internet of things receiving device sends the satellite navigation parameters and the observation parameters to a clock data processing module through an Internet communication system;
step 2, a clock data processing module extracts satellite navigation parameters and observation parameters to perform parameter preprocessing, and the clock data processing module determines a system predicted orbit after eliminating a global navigation satellite system clock through the positioning parameters of a broadcast ephemeris;
step 3, the clock data processing module establishes a super-bandwidth nonlinear clock estimation observation model through a system prediction orbit, and gradually solves the ambiguity parameter of indoor positioning of the Internet of things according to an ambiguity integer constraint form;
step 4, predicting clock parameters of the global navigation satellite system based on the ambiguity data in the step 3, and repeatedly executing the step 3 and the step 4 until the residual error value predicted by the clock is controlled within the precision threshold range;
step 5, constructing a convolutional neural network with a nonlinear structure, and training the convolutional neural network through the generated clock parameters of the global navigation satellite system;
and 6, predicting real-time clock parameters positioned in the Internet of things room through a convolutional neural network training result, and sending the predicted clock parameters to an authorized device in the Internet of things range through an Internet communication system.
According to the method, different weighted values are given to the clock data, the weighted values are dynamically adjusted according to the clock history, the clock deviation value is accurately predicted while the change of the clock offset rate is rapidly measured, the stability of the site clock is greatly improved, the time synchronization precision is high, finally, the clock prediction and compensation method is executed locally, only satellite data needs to be maintained without extra network communication resources, in addition, the interactive time deviation value is piggybacked through data, and the expense of control messages is reduced. The method integrates the global navigation satellite system receiver technology, computer data processing, neural network, satellite motion, wireless communication and other technologies, can accurately predict the nonlinear time sequence characteristic by learning a large number of samples through the convolutional neural network, calculates and predicts satellite clock parameters based on the observation data of the global satellite observation station, and effectively assists users of the positioning device used by the Internet of things to obtain high-precision positioning information.
Example (b):
the invention relates to a clock prediction method based on internet of things indoor positioning, which comprises the following specific steps:
step 1, collecting observation parameters by an Internet of things receiving module
The method comprises the steps that firstly, satellite navigation parameters and observation parameters of a global navigation satellite system are collected by using the internet of things receiving modules which are uniformly distributed in the range of the internet of things, and the internet of things receiving device is sent to the clock data processing module through the internet communication system. On the other hand, the requirements of real time and high speed of the network are met.
Step 2, generating system prediction orbit
The clock data processing module extracts satellite navigation parameters and observation parameters to perform parameter preprocessing, and determines a system predicted orbit after eliminating a global navigation satellite system clock through the positioning parameters of the broadcast ephemeris;
satellite navigation parameter and observation parameter b (t) in step 2l)，
Figure BDA0002593782750000081
Wherein
Figure BDA0002593782750000091
b(tl) Represents tlThe observed parameter at the moment, G (t)l) Represents tlAn observation parameter matrix of time, l is a time parameter index, phi denotes a clock state transition matrix, a0Representing the state quantity to be estimated, representing the observation noise, t0Is the initial time.
The satellite navigation parameters s' of the global navigation satellite system are:
Figure BDA0002593782750000092
in the formula, r0,s1,s2Three-dimensional position, speed and acceleration of the global navigation satellite system in an inertial coordinate system, q is a kinetic parameter of the global navigation satellite system, HNeIs the constant of the earth's gravity, a1Is the sum of various perturbation forces acting on the global navigation satellite system, a1Representing the total acceleration of the global navigation satellite system in motion.
Step 3, solving ambiguity parameter of indoor positioning of Internet of things
The clock data processing module establishes a super-bandwidth nonlinear clock estimation observation model through a system prediction orbit, and gradually solves the ambiguity parameter of indoor positioning of the Internet of things according to an ambiguity integer constraint form;
The super-bandwidth nonlinear clock estimation observation model comprises the following steps:
Figure BDA0002593782750000093
Figure BDA0002593782750000094
wherein Q and K representPseudorange observations and phase of the ball navigation satellite system,
Figure BDA0002593782750000095
representing a global navigation satellite system clock parameter,
Figure BDA0002593782750000096
indicating global navigation satellite system receiver clock error, λIFRepresenting the communication wavelength, MIFRepresenting an ambiguity parameter of a global navigation satellite system, ntropAnd S represents the tropospheric projection function and zenith tropospheric, phi, respectivelyL,IFAnd phiP,IFRespectively, carrier phase and pseudorange observation noise of a global navigation satellite system.
The technical scheme utilizes the advantages of model predictive control to accelerate the convergence speed of the sensor clock synchronization.
In order to effectively fix the ambiguity parameters of the global navigation satellite system, the ambiguity parameters are decomposed and fixed step by step.
Figure BDA0002593782750000101
Wherein w1And w2Respectively representing the frequencies, M, of carrier 1 and carrier 2 of the Internet of things1And M2And respectively representing the ambiguity parameters of the carrier 1 and the carrier 2 of the internet of things.
And fixing ambiguity parameters of the carrier 1 and the carrier 2 of the Internet of things:
Figure BDA0002593782750000102
wherein L is1And L2Respectively representing phase observations, P, of carrier 1 and carrier 2 of the Internet of things1And P2And respectively representing pseudo-range observed values of carrier 1 and carrier 2 of the internet of things.
Will M1-M2Solving ambiguity M in the retrospective clock estimation observation model 1The floating point solution of (2).
Step 4, estimating clock parameters
Predicting clock parameters of the global navigation satellite system based on the ambiguity data in the step 3, and repeatedly executing the step 3 and the step 4 until the residual error value of clock prediction is controlled within the precision threshold range;
step 5, constructing and training a neural network
Constructing a convolutional neural network with a nonlinear structure, and training the convolutional neural network through the generated clock parameters of the global navigation satellite system; the convolutional neural network mainly comprises an input layer, a convolutional layer, a pooling layer, a full-link layer and an output layer. The input layer mainly comprises clock parameters; the convolutional layer mainly extracts a plurality of characteristic graphs of the data; performing aggregation statistics on certain features by adopting a maximum pooling technology in a pooling layer, so that the dimensionality of a feature map is reduced, and the number of neural network nodes is reduced; integrating local information with obvious distinction in the convolution layer and the pooling layer in the full-connection layer, extracting characteristics and classifying; and finally, calculating the neural network parameters by taking a mean square error function between the observed value and the predicted value as a cost loss function and taking the minimum loss function of the network as a criterion. In order to obtain the parameters of the neural network, the neural network is trained by two modes, namely forward propagation and backward propagation, based on the clock parameters generated in the step 4. The forward propagation is mainly a characteristic diagram conversion process, an input sample is classified in a full connection layer after being subjected to layer-by-layer characteristic conversion, then a neural network prediction result is obtained, and a deviation term between the neural network prediction result and a real observation is calculated. And the reverse propagation is to transmit the error in the forward propagation from the last output layer to the input layer of the first layer, calculate the gradient of the error relative to the weight parameter, and correct the weight vector and the offset according to the gradient direction and magnitude of the error. The training process is based on the two propagation direction iterations, and when the output error is reduced to a certain degree, the model converges and the iterations stop.
And 6, predicting real-time clock parameters positioned in the Internet of things room through a convolutional neural network training result, and sending the predicted clock parameters to an authorized device in the Internet of things range through an Internet communication system.
The training of the convolutional neural network comprises:
net input S to jth neuronjComprises the following steps:
Figure BDA0002593782750000111
through the excitation function f (·), the net output of the j-th neuron is obtained as:
yi＝f(Sj)
wherein f (-) is Sigmoid function, yjRepresenting the j-th neuron net output.
Wherein the neural network sets the connection weight coefficient of the input layer and the hidden layer as vjThe connection weight coefficient of the hidden layer and the output layer is wjWhen the input is (v)i,ωi) When i is 0,1,2, …, n, the hidden layer neuron node output is:
Figure BDA0002593782750000112
Figure BDA0002593782750000113
and determining a model of the neural network according to the two formulas.
The neural network learning includes:
inputting p test samples to the neural network to obtain output values
Figure BDA0002593782750000121
Obtaining an error function E of the p samplepComprises the following steps:
Figure BDA0002593782750000122
wherein
Figure BDA0002593782750000123
For the desired output, the global error function E is, for all p samples:
Figure BDA0002593782750000124
where η ∈ (0,1) represents the learning rate of the neural network.
In summary, the invention discloses a clock prediction method based on internet of things indoor positioning. Based on the characteristic that the convolutional neural network can accurately predict the nonlinear time sequence through learning samples, the intelligent prediction of the clock in the range of the Internet of things is realized by constructing the convolutional neural network with a multilayer structure, and a user is assisted to obtain a high-precision positioning result. The invention uses the observation information of the satellite and the characteristic of the multilayer convolution neural network for predicting the nonlinear time sequence, thereby really predicting the real-time precise clock and assisting the user to obtain the real-time high-precision positioning result. According to the clock prediction method based on the Internet of things indoor positioning, the convolutional neural network is fully utilized to carry out more accurate prediction of nonlinear time sequence characteristics through learning of a large number of samples of clock prediction data of the Internet of things positioning system, clock parameters of a global navigation satellite system are calculated through data of the Internet of things receiving module, the neural network is utilized to realize prediction of the clock parameters, and the clock parameters are broadcasted to an authorized device in the range of the Internet of things to correct clock errors of the global navigation satellite system, so that the positioning accuracy of the Internet of things system is really improved.
The invention provides a clock prediction method based on internet of things indoor positioning, which predicts real-time clock parameters of the internet of things indoor positioning through a convolutional neural network training result and sends the predicted clock parameters to an authorized device in the internet of things range through an internet communication system, and further comprises the following steps:
receiving a previous data packet related to the indoor positioning of the Internet of things, acquiring the arrival time of the previous data packet based on the Internet of things, and predicting the arrival time of the next data packet based on the Internet of things based on the convolutional neural network;
when the next data packet does not arrive at the predicted arrival time, automatically generating a new data packet, and performing smooth processing on clock information carried in the new data packet to enable the output clock signals to be uniformly distributed in a time domain;
when the next data packet arrives at the predicted arrival time, performing smoothing processing on clock information carried in the next data packet to enable output clock signals to be uniformly distributed in a time domain;
converting all output clock signals from digital signals into analog signals, so that all the output clock signals meet preset jitter and drift conditions;
Time division multiplexing the clock signals meeting the preset jitter and drift conditions in a packet switching network, and outputting and displaying the time division multiplexing result;
and the previous data packet, the next data packet and the new data packet all carry clock information.
The working principle and the beneficial effects of the technical scheme are as follows: according to the clock prediction method based on the indoor positioning of the Internet of things, the time division multiplexing result is displayed on the electronic display screen through the artificial intelligence learning of the convolutional neural network and based on the clock parameters of the global navigation satellite system, so that the clock is predicted more accurately, and the positioning accuracy of the Internet of things system is improved.
The present invention may be embodied in other specific forms without departing from the spirit or essential attributes thereof, and it should be understood that various changes and modifications can be made therein by those skilled in the art without departing from the spirit and scope of the invention as defined in the appended claims.

Claims (9)

1. A clock prediction method based on Internet of things indoor positioning is characterized by comprising the following steps:
step 1, an Internet of things receiving module collects satellite navigation parameters and observation parameters of a global navigation satellite system, and an Internet of things receiving device sends the satellite navigation parameters and the observation parameters to a clock data processing module through an Internet communication system;
Step 2, a clock data processing module extracts satellite navigation parameters and observation parameters to perform parameter preprocessing, and the clock data processing module determines a system predicted orbit after eliminating a global navigation satellite system clock through the positioning parameters of a broadcast ephemeris;
step 3, the clock data processing module establishes a super-bandwidth nonlinear clock estimation observation model through a system prediction orbit, and gradually solves the ambiguity parameter of indoor positioning of the Internet of things according to an ambiguity integer constraint form;
step 4, predicting clock parameters of the global navigation satellite system based on the ambiguity data in the step 3, and repeatedly executing the step 3 and the step 4 until the residual error value predicted by the clock is controlled within the precision threshold range;
step 5, constructing a convolutional neural network with a nonlinear structure, and training the convolutional neural network through the generated clock parameters of the global navigation satellite system;
and 6, predicting real-time clock parameters positioned in the Internet of things room through a convolutional neural network training result, and sending the predicted clock parameters to an authorized device in the Internet of things range through an Internet communication system.
2. The method for predicting the clock based on the indoor positioning of the internet of things according to claim 1, wherein the step 1 of collecting the satellite navigation parameters and the observation parameters comprises the following steps:
(1.1) Internet of things receiving module acquisition tlObservation parameter matrix G (t) of timel)；
(1.2) the receiving module of the Internet of things marks the time parameter;
(1.3) setting a clock state transition matrix phi by the receiving module of the Internet of things;
(1.4) the Internet communication system issues observation noise;
(1.5) Internet communication System estimation tlObservation parameter b (t) at timel)：
Figure FDA0002593782740000011
Wherein
Figure FDA0002593782740000021
Wherein, l is a time parameter index, a0Representing the quantity of state to be estimated, t0At an initial time, phi (t)l,t0) Represents t0Time tlA phase difference function of the time of day.
3. The method of claim 1, wherein the satellite navigation parameters are calculated as:
(1.6) acquiring the three-dimensional position r of the global navigation satellite system in an inertial coordinate system0Speed s1Acceleration s2；
(1.7) collecting a dynamic parameter q of the global navigation satellite system by an Internet of things receiving module;
(1.8) calculating an Earth gravity constant HN by an Internet communication systeme；
(1.9) the receiving module of the Internet of things calculates the sum a of various perturbation forces acting on the global navigation satellite system1；
The satellite navigation parameters s' of the global navigation satellite system are:
Figure FDA0002593782740000022
4. the method for predicting the clock based on the indoor positioning of the internet of things according to claim 1, wherein the establishing of the clock estimation observation model of the ultra-bandwidth nonlinearity in the step 3 comprises:
(3.1) a clock data processing module acquires pseudo-range observed values and phases Q and K of a global navigation satellite system;
(3.2) the clock data processing module acquires the clock parameters of the global navigation satellite system
Figure FDA0002593782740000023
(3.3) clock data processing Module evaluation TotalReceiving clock error of ball navigation satellite system
Figure FDA0002593782740000024
(3.4) confirming the communication wavelength lambda of the Internet communication systemIF；
(3.5) calculating ambiguity parameter M of global navigation satellite systemIF；
(3.6) acquisition of tropospheric projection function ntropAnd zenith troposphere area S;
(3.7) Carrier phase of Global navigation satellite System [ phi ]L,IFAnd pseudorange observation noise phiP,IF；
The super-bandwidth nonlinear clock estimation observation model comprises the following steps:
Figure FDA0002593782740000031
Figure FDA0002593782740000032
5. the method of claim 4, wherein the ambiguity parameters are decomposed and fixed step by step for effectively fixing the ambiguity parameters of the global navigation satellite system;
Figure FDA0002593782740000033
wherein w1And w2Respectively representing the frequencies, M, of carrier 1 and carrier 2 of the Internet of things1And M2Respectively representing ambiguity parameters of carrier 1 and carrier 2 of the Internet of things;
and fixing ambiguity parameters of the carrier 1 and the carrier 2 of the Internet of things:
Figure FDA0002593782740000034
wherein L is1And L2Respectively represent phase observed values, P, of carrier 1 and carrier 2 of the Internet of things1And P2Respectively representing pseudo-range observed values of carrier 1 and carrier 2 of the Internet of things;
Will M1-M2Solving ambiguity M in the retrospective clock estimation observation model1The floating point solution of (2).
6. The method of claim 5, wherein the training of the convolutional neural network comprises:
(5.1) inputting a training sample set and a testing sample set; scaling any input image in a bilinear interpolation mode, and performing random horizontal turning, random translation, random image rotation and random image scaling on each image;
(5.2) adopting a plurality of loss functions as the training target function;
(5.3) ordering the loss functions;
(5.4) repeatedly carrying out dynamic target training on the deep neural network on the training sample set in sequence according to the sorted loss functions to obtain a recognition model;
(5.5) classifying the input test samples according to the recognition models;
wherein the net input S of the jth neuronjComprises the following steps:
Figure FDA0002593782740000041
through the excitation function f (·), the net output of the j-th neuron is obtained as:
yj＝f(Sj)
wherein f (-) is Sigmoid function, yjRepresenting the j-th neuron net output.
7. The method of claim 6, wherein the neural network sets the connection weight coefficient between the input layer and the hidden layer as vjThe connection weight coefficient of the hidden layer and the output layer is wjWhen inputting (v)j,wj) When j is 0,1,2, …, n, the hidden layer neuron node output is:
Figure FDA0002593782740000042
Figure FDA0002593782740000043
and determining a model of the neural network according to the two formulas.
8. The method of claim 7, wherein the neural network learning comprises:
inputting p test samples to the neural network to obtain output values
Figure FDA0002593782740000044
Obtaining an error function E of the p samplepComprises the following steps:
Figure FDA0002593782740000045
wherein
Figure FDA0002593782740000046
For the desired output, the global error function E is, for all p samples:
Figure FDA0002593782740000051
where η ∈ (0,1) represents the learning rate of the neural network.
9. The clock prediction method based on the indoor positioning of the internet of things according to claim 1, wherein: through the convolutional neural network training result, predict the real-time clock parameter of the indoor location of thing networking to the in-process of the device of authorizing in internet of things scope is sent the clock parameter of prediction through internet communication system, still include:
receiving a previous data packet related to the indoor positioning of the Internet of things, acquiring the arrival time of the previous data packet based on the Internet of things, and predicting the arrival time of the next data packet based on the Internet of things based on the convolutional neural network;
When the next data packet does not arrive at the predicted arrival time, automatically generating a new data packet, and performing smooth processing on clock information carried in the new data packet to enable the output clock signals to be uniformly distributed in a time domain;
when the next data packet arrives at the predicted arrival time, performing smoothing processing on clock information carried in the next data packet to enable output clock signals to be uniformly distributed in a time domain;
converting all output clock signals from digital signals into analog signals, so that all the output clock signals meet preset jitter and drift conditions;
time division multiplexing the clock signals meeting the preset jitter and drift conditions in a packet switching network, and outputting and displaying the time division multiplexing result;
and the previous data packet, the next data packet and the new data packet all carry clock information.
CN202010703571.7A 2020-07-21 2020-07-21 Clock prediction method based on Internet of things indoor positioning Withdrawn CN111854745A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010703571.7A CN111854745A (en) 2020-07-21 2020-07-21 Clock prediction method based on Internet of things indoor positioning

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010703571.7A CN111854745A (en) 2020-07-21 2020-07-21 Clock prediction method based on Internet of things indoor positioning

Publications (1)

Publication Number Publication Date
CN111854745A true CN111854745A (en) 2020-10-30

Family

ID=73002147

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010703571.7A Withdrawn CN111854745A (en) 2020-07-21 2020-07-21 Clock prediction method based on Internet of things indoor positioning

Country Status (1)

Country Link
CN (1) CN111854745A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113608248A (en) * 2021-06-25 2021-11-05 北京建筑大学 Beidou 5G fused high-precision routing inspection personnel positioning method and related equipment
CN113723591A (en) * 2021-07-29 2021-11-30 江苏师范大学 GNSS time service filtering method, system and device based on RBPNN and storage medium

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113608248A (en) * 2021-06-25 2021-11-05 北京建筑大学 Beidou 5G fused high-precision routing inspection personnel positioning method and related equipment
CN113723591A (en) * 2021-07-29 2021-11-30 江苏师范大学 GNSS time service filtering method, system and device based on RBPNN and storage medium

Similar Documents

Publication Publication Date Title
Zhu et al. Adapted error map based mobile robot UWB indoor positioning
CN113945206B (en) Positioning method and device based on multi-sensor fusion
Sermet et al. Crowdsourced approaches for stage measurements at ungauged locations using smartphones
CN110118549A (en) A kind of Multi-source Information Fusion localization method and device
CN112433236B (en) Error model calibration method, device, equipment and computer readable storage medium
Elbes et al. An indoor localization approach based on deep learning for indoor location-based services
CN111854745A (en) Clock prediction method based on Internet of things indoor positioning
CN114216459B (en) ELM-assisted GNSS/INS combined navigation unmanned target vehicle positioning method
CN106353722A (en) RSSI (received signal strength indicator) distance measuring method based on cost-reference particle filter
Liu et al. Geomagnetism-based indoor navigation by offloading strategy in NB-IoT
CN116391138A (en) Positioning method, device, equipment, system and storage medium
Yu et al. Floor positioning method indoors with smartphone’s barometer
Li et al. Loosely coupled GNSS/INS integration based on factor graph and aided by ARIMA model
CN114706110A (en) Vehicle satellite dynamic positioning method and system based on vehicle-road cooperation
Zhao et al. Fusing vehicle trajectories and GNSS measurements to improve GNSS positioning correction based on actor-critic learning
Huan et al. Indoor location fingerprinting algorithm based on path loss parameter estimation and bayesian inference
Wei et al. MM-Loc: Cross-sensor indoor smartphone location tracking using multimodal deep neural networks
CN102830391B (en) Accuracy index calculating method of infrared search and track system
CN112651991B (en) Visual positioning method, device and computer system
CN111025364A (en) Machine vision positioning system and method based on satellite assistance
Jia et al. An indoor and outdoor seamless positioning system based on android platform
CN104297761B (en) Based on the localization method that non-concurrent is received
CN111194000A (en) Distance measurement method and system based on Bluetooth fusion hybrid filtering and neural network
CN109581280A (en) The adaptive tuning on-line method, system and device of terminal
CN112907644B (en) Machine map-oriented visual positioning method

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
WW01 Invention patent application withdrawn after publication
WW01 Invention patent application withdrawn after publication

Application publication date: 20201030