CN111949135B - Touch communication fault tolerance method and system based on hybrid prediction - Google Patents

Touch communication fault tolerance method and system based on hybrid prediction Download PDF

Info

Publication number
CN111949135B
CN111949135B CN202010895087.9A CN202010895087A CN111949135B CN 111949135 B CN111949135 B CN 111949135B CN 202010895087 A CN202010895087 A CN 202010895087A CN 111949135 B CN111949135 B CN 111949135B
Authority
CN
China
Prior art keywords
data
prediction
value
theta
predicted
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.)
Active
Application number
CN202010895087.9A
Other languages
Chinese (zh)
Other versions
CN111949135A (en
Inventor
房颖
林擎旭
郑权斐
徐艺文
赵铁松
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.)
Fuzhou University
Original Assignee
Fuzhou University
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 Fuzhou University filed Critical Fuzhou University
Priority to CN202010895087.9A priority Critical patent/CN111949135B/en
Publication of CN111949135A publication Critical patent/CN111949135A/en
Application granted granted Critical
Publication of CN111949135B publication Critical patent/CN111949135B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
    • G06F3/01—Input arrangements or combined input and output arrangements for interaction between user and computer
    • G06F3/016—Input arrangements with force or tactile feedback as computer generated output to the user
    • 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
    • G06N3/044—Recurrent networks, e.g. Hopfield 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
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2203/00—Indexing scheme relating to G06F3/00 - G06F3/048
    • G06F2203/01—Indexing scheme relating to G06F3/01
    • G06F2203/012—Walk-in-place systems for allowing a user to walk in a virtual environment while constraining him to a given position in the physical environment

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • General Health & Medical Sciences (AREA)
  • Computing Systems (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Biomedical Technology (AREA)
  • Molecular Biology (AREA)
  • Biophysics (AREA)
  • Artificial Intelligence (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Human Computer Interaction (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)
  • Compression, Expansion, Code Conversion, And Decoders (AREA)

Abstract

The invention relates to a touch communication fault-tolerant method and a touch communication fault-tolerant system based on hybrid prediction, wherein when data are sent, data to be sent are subjected to prediction compression, and only data frames with the relative deviation between a predicted value and a true value larger than a dead zone parameter are sent to reduce the data volume to be sent; when receiving data, if missing data is met, counting and adding 1, comparing a current count value n with a previously received prediction threshold value L, if n is larger than L, predicting the missing data by adopting a zero-order retention prediction algorithm, and otherwise predicting the missing data by adopting an LSTM model; and if the data is received, resetting the count value, wherein the predicted value is a true value, and simultaneously acquiring a newly received prediction threshold value L. According to the invention, through setting the continuous prediction threshold, the receiving end can autonomously select a relatively conservative prediction algorithm, so that the influence caused by accumulated errors caused by continuous prediction is reduced, and the stability of the touch communication system is further improved.

Description

Touch communication fault tolerance method and system based on hybrid prediction
Technical Field
The invention relates to the technical field of touch communication, in particular to a touch communication fault tolerance method and system based on hybrid prediction.
Background
In recent years, with the development of the field of Virtual Reality (VR) and artificial intelligence, people have been unable to satisfy the sense of reality and immersion obtained visually and aurally, and thus the sense of touch has received increasing attention in the industry. In the transmission process of the touch data, transmission errors may be caused by problems such as network jitter, and further, the touch communication system may be unstable. Traditionally, zero order hold and first order linearity have been used to solve this problem, but research has shown that haptic data is not simply a linear relationship. In recent years, long-short term memory (LSTM) neural networks have been utilized by researchers to implement nonlinear prediction of haptic data due to their advantages in time series prediction, and have greatly improved fault tolerance. However, the method has the problem of error accumulation under the long-term prediction condition, and the performance still has a promotion space.
Disclosure of Invention
In view of the above, the present invention provides a haptic communication fault tolerance method and system based on hybrid prediction, which enable a receiving end to autonomously select a relatively conservative prediction algorithm by setting a continuous prediction threshold, reduce the influence caused by accumulated errors due to continuous prediction, and further improve the stability of a haptic communication system.
The invention is realized by adopting the following scheme: a haptic communication fault tolerance method based on hybrid prediction specifically comprises the following steps:
when data is sent, the data to be sent is subjected to prediction compression, and only data frames with the relative deviation between the predicted value and the true value larger than the dead zone parameter are sent, so that the data volume needing to be sent is reduced; when data is transmitted, a prediction threshold value L is calculated and transmitted together with the data;
when receiving data, if missing data is encountered, adding 1 to the count, comparing the current count value n with the previously received prediction threshold value L, if n is greater than L, predicting the missing data by adopting a zero-order retention prediction algorithm, otherwise predicting the missing data by adopting an LSTM model; if data is received, resetting the count value, wherein the predicted value is a true value, and simultaneously acquiring a newly received prediction threshold value L;
the missing data is packet loss data or data discarded due to compression.
Further, the performing the predictive compression on the data to be transmitted specifically includes: the data is predictively compressed using the following equation:
Figure BDA0002658190750000021
in the formula,
Figure BDA0002658190750000022
for the actual value to be currently transmitted,
Figure BDA0002658190750000023
k is a dead zone parameter, which is a predicted value obtained by a prediction algorithm (the invention adopts an LSTM model for prediction).
Further, the value of the dead zone parameter k is 0.15.
Further, the prediction threshold L is calculated by the following formula:
L＝Lmax*Qθ；
in the formula, LmaxContinuous prediction of extremum, Q, for LSTMθAre gradient influencing parameters.
Further, LmaxThe value of (d) is 8.
Further, the gradient influence parameter QθThe calculation of (a) is specifically:
Figure BDA0002658190750000031
in the formula, thetamaxAnd theta x, theta y and theta z are predicted gradients in x, y and z directions, wherein theta x, theta y and theta z are maximum values of the predicted gradients in the communication process.
Further, when the method is tested, whether packet loss is sent or not in the communication process is represented by the state of a Boolean type pointer TransmitFlag, the state of each time TransmitFlag is determined by randomly generating a normal distribution value, eight different packet loss rates of 0% -50% are set, and the state of the TransmitFlag is stored in a flag bit loss _ v to represent whether packet loss is sent or not at each time; when data is received, if TransmitFlag is True, the data is sent, and the predicted value is a True value, namely prediction is not needed; and if TransmitFlag is equal to False, the fact that packet loss occurs in the communication process is indicated, and the predicted value is obtained through prediction of an LSTM model or a zero-order keeping algorithm.
The invention also provides a system based on the mixed prediction-based touch communication fault-tolerant method, which comprises a sending end and a receiving end;
when the sending end sends data, the data to be sent is subjected to predictive compression, and only the data frame with the relative deviation between the predicted value and the true value larger than the dead zone parameter is sent, so that the data volume needing to be sent is reduced; when data is transmitted, a prediction threshold value L is calculated and transmitted together with the data;
when receiving data, if missing data is met, counting and adding 1, comparing a current count value n with a previously received prediction threshold value L, if n is larger than L, predicting the missing data by adopting a zero-order retention prediction algorithm, and otherwise predicting the missing data by adopting an LSTM model; if data are received, resetting the count value, wherein the predicted value is a true value, and simultaneously acquiring a newly received prediction threshold value L;
the missing data is packet loss data or data discarded due to compression.
The invention also provides a touch communication fault-tolerant system based on hybrid prediction, which comprises a sending end, wherein when the sending end sends data, the sending end only sends a data frame of which the relative deviation between a predicted value and a true value is greater than a dead zone parameter by predicting and compressing the data to be sent so as to reduce the data volume to be sent; when data is transmitted, a prediction threshold value L is calculated and transmitted together with the data;
the performing the predictive compression on the data to be transmitted specifically includes: the data is predictively compressed using the following equation:
Figure BDA0002658190750000041
in the formula,
Figure BDA0002658190750000042
for the current transmissionThe actual value of the value is,
Figure BDA0002658190750000043
k is a predicted value obtained by a prediction algorithm (LSTM prediction is adopted in the invention), and is a dead zone parameter;
the prediction threshold L is calculated using the following equation:
L＝Lmax*Qθ；
in the formula, LmaxContinuous prediction of extremum, Q, for LSTMθIs a gradient influence parameter;
wherein the gradient influences the parameter QθThe calculation of (a) is specifically:
Figure BDA0002658190750000044
in the formula, thetamaxAnd theta x, theta y and theta z are predicted gradients in x, y and z directions, wherein theta x, theta y and theta z are maximum values of the predicted gradients in the communication process.
The invention also provides a touch communication fault-tolerant system based on hybrid prediction, which comprises a receiving end, wherein when the receiving end receives data, if the receiving end encounters missing data, the counting is added with 1, the current counting value n is compared with the previously received prediction threshold value L, if n is larger than L, the missing data is predicted by adopting a zero-order retention prediction algorithm, otherwise, the missing data is predicted by adopting an LSTM model; if data are received, resetting the count value, wherein the predicted value is a true value, and simultaneously acquiring a newly received prediction threshold value L; the missing data is packet loss data or data discarded due to compression.
Compared with the prior art, the invention has the following beneficial effects: according to the invention, through setting the continuous prediction threshold, the receiving end can autonomously select a relatively conservative prediction algorithm, so that the influence caused by accumulated errors caused by continuous prediction is reduced, and the stability of the touch communication system is further improved.
Drawings
FIG. 1 is a schematic flow chart of a method according to an embodiment of the present invention.
FIG. 2 is a diagram illustrating tactile data in three directions according to an embodiment of the present invention. Wherein (a) is the x direction, (b) is the y direction, and (c) is the z direction.
Detailed Description
The invention is further explained below with reference to the drawings and the embodiments.
It should be noted that the following detailed description is exemplary and is intended to provide further explanation of the disclosure. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.
It is noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of example embodiments according to the present application. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, and it should be understood that when the terms "comprises" and/or "comprising" are used in this specification, they specify the presence of stated features, steps, operations, devices, components, and/or combinations thereof, unless the context clearly indicates otherwise.
As shown in fig. 1, the present embodiment provides a haptic communication fault tolerance method based on hybrid prediction, which specifically includes the following steps:
when data is sent, the data to be sent is subjected to prediction compression, and only data frames with the relative deviation between the predicted value and the true value larger than the dead zone parameter are sent, so that the data volume needing to be sent is reduced; when data is transmitted, a prediction threshold value L is calculated and transmitted together with the data;
when receiving data, if missing data is met, counting and adding 1, comparing a current count value n with a previously received prediction threshold value L, if n is larger than L, predicting the missing data by adopting a zero-order retention prediction algorithm, and otherwise predicting the missing data by adopting an LSTM model; if data are received, resetting the count value, wherein the predicted value is a true value, and simultaneously acquiring a newly received prediction threshold value L;
the missing data is packet loss data or data discarded due to compression.
In this embodiment, the performing the predictive compression on the data to be transmitted specifically includes: the data is predictively compressed using the following equation:
Figure BDA0002658190750000071
in the formula,
Figure BDA0002658190750000072
for the actual value to be currently transmitted,
Figure BDA0002658190750000073
k is a dead zone parameter for a predicted value obtained by a prediction algorithm (the invention uses an LSTM model for prediction).
In this embodiment, the value of the dead zone parameter k is 0.15.
In this embodiment, the prediction threshold L is calculated by the following formula:
L＝Lmax*Qθ；
in the formula, LmaxContinuous prediction of extremum, Q, for LSTMθAre gradient influencing parameters.
In this embodiment, LmaxThe value of (d) is 8.
In the present embodiment, the gradient influence parameter QθThe calculation of (a) is specifically:
Figure BDA0002658190750000074
in the formula, thetamaxAnd theta x, theta y and theta z are predicted gradients in x, y and z directions, wherein theta x, theta y and theta z are maximum values of the predicted gradients in the communication process.
Specifically, as shown in fig. 2, since the tactile sensation data is three-dimensional velocity data, the present embodiment divides the velocity data in three directions into
Figure BDA0002658190750000075
Corresponding predictive value division
Figure BDA0002658190750000076
Predicting a gradient θ into [ θ ]x,θy,θz]，[rx,ry,rz]Is the true value of the previous time in three directions, [ m ]x,my,mz]The predicted values of the next time in the three directions and the actual value variation of the previous time, i.e. the predicted difference value, [ m ]ref x,mref y,mref z]The actual difference, which is the variation between the actual value at the next time and the actual value at the previous time in the three directions, is defined as follows from the formula of fig. 2, where the predicted gradient θ in a single dimension is obtained as follows:
Figure BDA0002658190750000081
in the formula, ghAnd rhValue representing the horizontal direction, gvAnd rvRepresenting the values in the vertical direction, the following formula can be derived from fig. 2:
Figure BDA0002658190750000082
in the formula, mjRepresenting the predicted difference in a certain direction.
Substituting the formula into the above formula
Figure BDA0002658190750000083
Where θ is the predicted gradient value in a dimension.
By predicting gradients [ theta x, theta y, theta z ] in three directions]The gradient influence parameter Q can be derivedθThe following formula:
Figure BDA0002658190750000084
wherein theta ismaxThe maximum value of the prediction gradient appeared in the communication process is the initial value pi/10000, and is updated when a larger value appears.
In this embodiment, when performing the method test, the boolean pointer × TransmitFlag is used to represent whether packet loss is sent in the communication process, the state of each time × TransmitFlag is determined by randomly generating a normal distribution value, eight different packet loss rates (0% to 50% (in this embodiment, 0%, 5%, 10%, 15%, 20%, 30%, 40%, 50%) are set, and the state of the TransmitFlag is stored in a flag loss _ v, which represents whether packet loss is sent at each time; when data is received, if TransmitFlag is True, the data is sent, and the predicted value is a True value, namely prediction is not needed; if transmit flag is equal to False, it indicates that packet loss occurs in the communication process, and the predicted value is predicted by an LSTM model or a zero-order hold algorithm.
The embodiment also provides a system based on the haptic communication fault tolerance method based on the hybrid prediction, which comprises a sending end and a receiving end;
when the sending end sends data, the data to be sent is subjected to predictive compression, and only the data frame with the relative deviation between the predicted value and the true value larger than the dead zone parameter is sent, so that the data volume needing to be sent is reduced; when data is transmitted, a prediction threshold value L is calculated and transmitted together with the data;
when receiving data, if missing data is met, counting and adding 1, comparing a current count value n with a previously received prediction threshold value L, if n is larger than L, predicting the missing data by adopting a zero-order retention prediction algorithm, and otherwise predicting the missing data by adopting an LSTM model; if data is received, resetting the count value, wherein the predicted value is a true value, and simultaneously acquiring a newly received prediction threshold value L;
the missing data is packet loss data or data discarded due to compression.
The embodiment also provides a touch communication fault-tolerant system based on hybrid prediction, which comprises a sending end, wherein when the sending end sends data, the sending end only sends a data frame of which the relative deviation between a predicted value and a true value is greater than a dead zone parameter by performing prediction compression on the data to be sent so as to reduce the data volume needing to be sent; when data is transmitted, a prediction threshold value L is calculated and transmitted together with the data;
the performing the predictive compression on the data to be transmitted specifically includes: the data is predictively compressed using the following equation:
Figure BDA0002658190750000101
in the formula,
Figure BDA0002658190750000102
for the actual value to be currently transmitted,
Figure BDA0002658190750000103
k is a dead zone parameter, which is a predicted value obtained by a prediction algorithm (LSTM prediction is used in the present invention).
The prediction threshold L is calculated using the following equation:
L＝Lmax*Qθ；
in the formula, LmaxContinuous prediction of extremum, Q, for LSTMθIs a gradient-influencing parameter;
wherein the gradient influences the parameter QθThe calculation of (a) is specifically:
Figure BDA0002658190750000104
in the formula, thetamaxAnd theta x, theta y and theta z are the predicted gradients in the x direction, the y direction and the z direction which are the maximum values of the predicted gradients in the communication process. The remaining specific methods were consistent with the above.
The embodiment also provides a touch communication fault-tolerant system based on hybrid prediction, which comprises a receiving end, wherein when the receiving end receives data, if the receiving end encounters missing data, the counting is added with 1, the current counting value n is compared with the previously received prediction threshold value L, if n is larger than L, the missing data is predicted by adopting a zero-order retention prediction algorithm, otherwise, the missing data is predicted by adopting an LSTM model; if data are received, resetting the count value, wherein the predicted value is a true value, and simultaneously acquiring a newly received prediction threshold value L; the missing data is packet loss data or data discarded due to compression. The remaining specific methods were consistent with the above.
Specifically, the following steps are adopted in the construction of the method of the embodiment:
and step S1, acquiring data for model training and experimental testing. The model used is a long-short term memory network model, i.e. the LSTM model. Determining model parameters and a loss function of a neural network, and optimizing the model to be optimal;
step S2, at the transmitting end, performing data compression by combining dead zone compression with a trained LSTM network, calculating a continuous prediction threshold L for data to be transmitted outside the dead zone, and transmitting the continuous prediction threshold L together with the data;
step S3, packet loss occurs randomly through normal distribution simulation in the network, and eight kinds of packet loss rates are set, wherein the packet loss rates are respectively 0%, 5%, 10%, 15%, 20%, 30%, 40% and 50%;
step S4, the receiving end carries out prediction reconstruction on lost data packets and compressed data packets in the network, and original data are recovered;
wherein, the step S1 is specifically implemented as follows:
step S11, manually operating equipment to perform a complex maze walking experiment through touch experimental equipment, recording the speed value, the position value and the force value of the small ball at each moment by taking 1ms as a unit, and repeatedly obtaining a large enough data sample;
step S12, constructing an LSTM network model, wherein the model is based on an LSTM original model integrated by Keras, the input of the model is speed values in x, y and z directions, normalization is needed before the model is input to reduce the influence caused by a large data change range, the normalization mode is that data is (normal _ data +10)/20, the output of the LSTM layer needs to be connected with a dense full connection layer, and finally reverse normalization is carried out to obtain a final predicted value;
step S13, setting a loss function of the model, which is a reliability index of the tactile communication system, and setting a quality index, which is commonly used for the stability of the tactile communication system, as a Mean Square Error (MSE), so that the loss function is set as MSE in the links of model training and quality evaluation, and the formula is as follows:
Figure BDA0002658190750000121
in the formula,
Figure BDA0002658190750000122
in order to be the true value of the value,
Figure BDA0002658190750000123
the predicted value is obtained through a prediction algorithm, and N is the total number of samples;
and step S14, optimizing the model performance to the best, determining other parameters of the LSTM network, and finally determining the model parameters of look _ back being 1, units being 10, and lr being 0.001 through model optimization.
Wherein, the step S2 is specifically implemented as follows:
step S21, compressing the data by a method of combining a prediction algorithm and dead zone compression according to the human body perception characteristic and the dead zone proposed by Weber, wherein the compression principle is as follows:
Figure BDA0002658190750000124
wherein,
Figure BDA0002658190750000125
for the actual value to be currently transmitted,
Figure BDA0002658190750000126
k is a dead zone parameter, which is a predicted value obtained by a prediction algorithm (LSTM prediction is adopted in the present invention), and a reference value is taken in an experiment to be 0.15.
In step S22, based on the data compression method determined in step S21, the conventional prediction algorithm is not suitable for prediction of the haptic data because the haptic data is not simple linear data and has a poor change regularity, so that the data compression is performed by combining the LSTM network with dead zone compression.
In step S23, based on step S21, if the data at the current time is not compressed, that is, the current data is to be transmitted, the continuous prediction threshold L at the current time is calculated. Since the influence parameter for determining the continuous prediction threshold L is not able to use only the LSTM continuous prediction extremum as the continuous prediction threshold for the hybrid prediction due to the poor regularity of the change of the haptic data, the present embodiment proposes to determine the continuous prediction threshold L according to the change of the haptic data at each moment, i.e. the prediction gradient, in combination with the LSTM continuous prediction extremum. That is, the hybrid prediction continuous prediction threshold L can be expressed by the following equation:
L＝Lmax*Qθ
in the formula, LmaxThe extremum is continuously predicted for LSTM and is experimentally determined to be 8. QθAre gradient influencing parameters.
Step S24, determining gradient influence parameter Q based on step S23θAnd (4) determining the formula.
Wherein, the step S3 is specifically implemented as follows:
step S31, the data sent by the sending end generates a random number in the network through normal distribution, the packet loss problem in the actual communication process is simulated according to the size of the random number, and 8 different packet loss rates are set in the experiment, which are 0%, 5%, 10%, 15%, 20%, 30%, 40%, and 50%, respectively.
Step S32, based on step S31, representing whether packet loss occurs in the communication process through the state of the Boolean type pointer TransmitFlag, determining the state of each time TransmitFlag through randomly generating a normal distribution value, setting eight different packet loss rates of 0% -50%, and storing the state of the TransmitFlag in a flag bit loss _ v to represent whether packet loss occurs in the data at each time;
wherein, the step S4 is specifically implemented as follows:
step S41, setting a continuous prediction counter n at the receiving end for comparing with a continuous prediction threshold L;
step S42, at the receiving end, if the transmit flag is True, it indicates that the receiving end receives the data sent by the sending end, and the predicted value at this time is a True value, that is, prediction is not needed, and at the same time, the receiving end receives a new continuous prediction threshold L, clears the counter n to zero, and counts again;
step S43, if transmit flag is False, it indicates packet loss occurs in the communication process, at this time, 1 is added to the counter n, the size of n and the continuous prediction threshold L is compared, if n is less than or equal to L, LSTM is used for prediction, otherwise, zeroth order prediction is used;
step S44, the data compressed at the transmitting end and the data lost in the communication process are predicted and reconstructed using the same hybrid prediction concept.
As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
The foregoing is directed to preferred embodiments of the present invention, other and further embodiments of the invention may be devised without departing from the basic scope thereof, and the scope thereof is determined by the claims that follow. However, any simple modification, equivalent change and modification of the above embodiments according to the technical essence of the present invention are within the protection scope of the technical solution of the present invention.

Claims (7)

1. A haptic communication fault tolerance method based on hybrid prediction is characterized by comprising the following steps:
when data is sent, the data to be sent is subjected to prediction compression, and only data frames with the relative deviation between the predicted value and the true value larger than the dead zone parameter are sent, so that the data volume needing to be sent is reduced; when data is transmitted, a prediction threshold value L is calculated and transmitted together with the data;
when receiving data, if missing data is met, counting and adding 1, comparing a current count value n with a previously received prediction threshold value L, if n is larger than L, predicting the missing data by adopting a zero-order retention prediction algorithm, and otherwise predicting the missing data by adopting an LSTM model; if data are received, resetting the count value, wherein the predicted value is a true value, and simultaneously acquiring a newly received prediction threshold value L; the prediction threshold L is calculated using the following equation:
L＝Lmax*Qθ；
in the formula, LmaxContinuous prediction of extremum, L, for LSTMmaxHas a value of 8, QθIs a gradient influence parameter;
the missing data is packet loss data or data discarded due to compression;
gradient influence parameter QθThe calculation of (a) is specifically:
Figure FDA0003590371460000011
in the formula, thetamaxAnd theta x, theta y and theta z are predicted gradients in x, y and z directions, wherein theta x, theta y and theta z are maximum values of the predicted gradients in the communication process.
2. A haptic communication fault tolerant method based on hybrid prediction according to claim 1, wherein the predictive compression of the data to be transmitted is specifically: the data is predictively compressed using the following equation:
Figure FDA0003590371460000021
in the formula,
Figure FDA0003590371460000022
for the actual value to be currently transmitted,
Figure FDA0003590371460000023
k is a dead zone parameter for the predicted value obtained by the prediction of the LSTM model.
3. A haptic communication fault-tolerant method based on hybrid prediction according to claim 2, wherein the dead zone parameter takes a value of 0.15.
4. The fault-tolerant method for touch communication based on hybrid prediction as claimed in claim 1, wherein in the testing of the method, the status of boolean pointer transmit flag is used to represent whether packet loss occurs during communication, the status of transmit flag at each time is determined by randomly generating a normal distribution value, eight different packet loss rates of 0% to 50% are set, and the status of transmit flag is stored in a flag bit loss _ v to represent whether packet loss occurs at each time; when data is received, if TransmitFlag is True, the data is sent, and the predicted value is a True value, namely prediction is not needed; if transmit flag is equal to False, it indicates that packet loss occurs in the communication process, and the predicted value is predicted by an LSTM model or a zero-order hold algorithm.
5. A system based on the hybrid prediction-based haptic communication fault-tolerant method of any one of claims 1-4, comprising a transmitting end, a receiving end;
when the sending end sends data, the data to be sent is subjected to predictive compression, and only the data frame with the relative deviation between the predicted value and the true value larger than the dead zone parameter is sent, so that the data volume needing to be sent is reduced; when data is transmitted, a prediction threshold value L is calculated and transmitted together with the data;
when receiving data, if missing data is met, counting and adding 1, comparing a current count value n with a previously received prediction threshold value L, if n is larger than L, predicting the missing data by adopting a zero-order retention prediction algorithm, and otherwise predicting the missing data by adopting an LSTM model; if data are received, resetting the count value, wherein the predicted value is a true value, and simultaneously acquiring a newly received prediction threshold value L;
the missing data is packet loss data or data discarded due to compression;
the prediction threshold L is calculated using the following equation:
L＝Lmax*Qθ；
in the formula, LmaxExtreme values are continuously predicted for LSTM, and the magnitude is determined to be 8, QθIs a shadow of gradientA sound parameter;
wherein the gradient influences the parameter QθThe calculation of (a) is specifically:
Figure FDA0003590371460000031
in the formula, thetamaxAnd theta x, theta y and theta z are predicted gradients in x, y and z directions, wherein theta x, theta y and theta z are maximum values of the predicted gradients in the communication process.
6. A touch communication fault-tolerant system based on hybrid prediction is characterized by comprising a sending end, wherein when the sending end sends data, the sending end only sends a data frame of which the relative deviation between a predicted value and a real value is larger than a dead zone parameter by performing prediction compression on the data to be sent so as to reduce the data volume needing to be sent; when data is transmitted, a prediction threshold value L is calculated and transmitted together with the data;
the performing the predictive compression on the data to be transmitted specifically includes: the data is predictively compressed using the following equation:
Figure FDA0003590371460000041
in the formula
Figure FDA0003590371460000042
For the actual value to be currently transmitted,
Figure FDA0003590371460000043
the predicted value is obtained through prediction of an LSTM model, and k is a dead zone parameter;
the prediction threshold L is calculated using the following equation:
L＝Lmax*Qθ；
in the formula, LmaxExtreme values are continuously predicted for LSTM, and the magnitude is determined to be 8, QθIs a gradient influence parameter;
wherein the gradient influences the parameterQθThe calculation of (a) is specifically:
Figure FDA0003590371460000044
in the formula, thetamaxAnd theta x, theta y and theta z are predicted gradients in x, y and z directions, wherein theta x, theta y and theta z are maximum values of the predicted gradients in the communication process.
7. A touch communication fault-tolerant system based on hybrid prediction is characterized by comprising a receiving end, wherein when the receiving end receives data, if the receiving end encounters missing data, the counting is added with 1, the current counting value n is compared with a previously received prediction threshold value L, if n is larger than L, a zero-order retention prediction algorithm is adopted to predict the missing data, otherwise, an LSTM model is adopted to predict the missing data; if data are received, resetting the count value, wherein the predicted value is a true value, and simultaneously acquiring a newly received prediction threshold value L; the missing data is packet loss data or data discarded due to compression;
the prediction threshold L is calculated using the following equation:
L＝Lmax*Qθ；
in the formula, LmaxExtreme values are continuously predicted for LSTM, and the magnitude is determined to be 8, QθIs a gradient influence parameter;
wherein the gradient influences the parameter QθThe calculation of (a) is specifically:
Figure FDA0003590371460000051
in the formula, thetamaxAnd theta x, theta y and theta z are predicted gradients in x, y and z directions, wherein theta x, theta y and theta z are maximum values of the predicted gradients in the communication process.
CN202010895087.9A 2020-08-31 2020-08-31 Touch communication fault tolerance method and system based on hybrid prediction Active CN111949135B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010895087.9A CN111949135B (en) 2020-08-31 2020-08-31 Touch communication fault tolerance method and system based on hybrid prediction

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010895087.9A CN111949135B (en) 2020-08-31 2020-08-31 Touch communication fault tolerance method and system based on hybrid prediction

Publications (2)

Publication Number Publication Date
CN111949135A CN111949135A (en) 2020-11-17
CN111949135B true CN111949135B (en) 2022-06-17

Family

ID=73367667

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010895087.9A Active CN111949135B (en) 2020-08-31 2020-08-31 Touch communication fault tolerance method and system based on hybrid prediction

Country Status (1)

Country Link
CN (1) CN111949135B (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115484002A (en) * 2022-09-09 2022-12-16 中国电信股份有限公司 Information processing method and device, electronic equipment and computer readable medium

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20090001411A (en) * 2007-03-14 2009-01-09 인천대학교 산학협력단 Haptic data transferring system for networked collaboration and method thereof
CN103607395A (en) * 2013-11-20 2014-02-26 深圳先进技术研究院 Touch sense data transmission method and apparatus
CN110502806A (en) * 2019-07-31 2019-11-26 电子科技大学 A kind of wireless frequency spectrum degree prediction technique based on LSTM network
CN110912737A (en) * 2019-11-14 2020-03-24 国网浙江省电力有限公司信息通信分公司 Dynamic perception performance early warning method based on hybrid model

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
GB0322875D0 (en) * 2003-09-30 2003-10-29 British Telecomm Haptics transmission systems
US10560530B2 (en) * 2018-02-16 2020-02-11 Tata Consultancy Services Limited Data analyses using compressive sensing for internet of things (IoT) networks
CN108764517B (en) * 2018-04-08 2020-12-04 中南大学 Method, equipment and storage medium for predicting change trend of silicon content in molten iron of blast furnace
US10579146B2 (en) * 2018-06-15 2020-03-03 Immersion Corporation Systems and methods for multi-level closed loop control of haptic effects
CN108694475B (en) * 2018-06-26 2022-04-01 南京航空航天大学 Short-time-scale photovoltaic cell power generation capacity prediction method based on hybrid model

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20090001411A (en) * 2007-03-14 2009-01-09 인천대학교 산학협력단 Haptic data transferring system for networked collaboration and method thereof
CN103607395A (en) * 2013-11-20 2014-02-26 深圳先进技术研究院 Touch sense data transmission method and apparatus
CN110502806A (en) * 2019-07-31 2019-11-26 电子科技大学 A kind of wireless frequency spectrum degree prediction technique based on LSTM network
CN110912737A (en) * 2019-11-14 2020-03-24 国网浙江省电力有限公司信息通信分公司 Dynamic perception performance early warning method based on hybrid model

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
Error Resilience of Haptic Data in Interactive Systems;Yiwen Xu等;《2019 11th International Conference on Wireless Communications and Signal Processing (WCSP)》;20191209;全文 *
基于TCP/IP的远程触觉交互系统力反馈数据传输方法;蒋峥峥等;《计算机科学》;20131115(第11期);第58-60、69页 *

Also Published As

Publication number Publication date
CN111949135A (en) 2020-11-17

Similar Documents

Publication Publication Date Title
CN109819233B (en) Digital twin system based on virtual imaging technology
CN106027300B (en) A kind of intelligent robot Parameter Optimization System and method using neural network
CN113033712B (en) Multi-user cooperative training people flow statistical method and system based on federal learning
CN108900542A (en) Ddos attack detection method and device based on LSTM prediction model
CN111898769A (en) Method and system for establishing user behavior period model based on horizontal federal learning
CN111191897A (en) Service flow online compliance prediction method and system based on bidirectional GRU neural network
CN107291879A (en) The method for visualizing of three-dimensional environment map in a kind of virtual reality system
CN111949135B (en) Touch communication fault tolerance method and system based on hybrid prediction
CN111966226B (en) Touch communication fault-tolerant method and system based on compensation type long-term and short-term memory network
CN114626293A (en) Method, device, equipment and storage medium for predicting collision simulation result
De Freitas et al. Perception policies for intelligent virtual agents
CN113657644B (en) False data detection method, device, equipment and storage medium in smart power grid
CN113887501A (en) Behavior recognition method and device, storage medium and electronic equipment
CN114363212B (en) Equipment detection method, device, equipment and storage medium
JP7348243B2 (en) Test methods, devices and equipment for traffic flow monitoring measurement systems
CN111611703A (en) Sand table deduction method, device, equipment and storage medium based on digital twins
CN103227644A (en) Compression method of automobile body small-format data
CN116811895B (en) Vehicle running speed determination model processing method and vehicle running speed determination method
CN107743077B (en) Method and device for evaluating network performance of information-physical fusion system
CN116915720A (en) Internet of things equipment flow identification method and system, electronic equipment and storage medium
CN110288666A (en) A kind of data compression method and apparatus
CN116844340A (en) Road traffic risk prediction method based on artificial intelligence
CN113934862B (en) Community security risk prediction method, device, electronic equipment and medium
CN113961962A (en) Model training method and system based on privacy protection and computer equipment
CN115499657A (en) Training method, application method, device and equipment of video code rate adaptive network

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant