CN111163530A - Wireless local area network performance enhancing method based on neural network algorithm - Google Patents

Wireless local area network performance enhancing method based on neural network algorithm Download PDF

Info

Publication number
CN111163530A
CN111163530A CN201911126396.3A CN201911126396A CN111163530A CN 111163530 A CN111163530 A CN 111163530A CN 201911126396 A CN201911126396 A CN 201911126396A CN 111163530 A CN111163530 A CN 111163530A
Authority
CN
China
Prior art keywords
model
data
neural network
output
system performance
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.)
Pending
Application number
CN201911126396.3A
Other languages
Chinese (zh)
Inventor
李俊超
邬永强
李兆刚
章为昆
陈晨
何日阳
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.)
Zhejiang Wellsun Intelligent Technology Co Ltd
Original Assignee
Zhejiang Wellsun Intelligent 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 Zhejiang Wellsun Intelligent Technology Co Ltd filed Critical Zhejiang Wellsun Intelligent Technology Co Ltd
Priority to CN201911126396.3A priority Critical patent/CN111163530A/en
Publication of CN111163530A publication Critical patent/CN111163530A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W74/00—Wireless channel access
    • H04W74/08—Non-scheduled access, e.g. ALOHA
    • H04W74/0808—Non-scheduled access, e.g. ALOHA using carrier sensing, e.g. carrier sense multiple access [CSMA]
    • H04W74/0816—Non-scheduled access, e.g. ALOHA using carrier sensing, e.g. carrier sense multiple access [CSMA] with collision avoidance
    • 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
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/084—Backpropagation, e.g. using gradient descent

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Health & Medical Sciences (AREA)
  • Computing Systems (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Molecular Biology (AREA)
  • Artificial Intelligence (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Mobile Radio Communication Systems (AREA)

Abstract

The invention discloses a wireless local area network performance enhancing method based on a neural network algorithm, which comprises the steps of carrying out normalized processing on original data, establishing a nonlinear regression model by using a neural network, and repeatedly carrying out training modeling by using the original data until the model can reflect the relationship between the backoff window increasing multiple and the system throughput, the system delay and the data packet retransmission rate, predicting the backoff window increasing multiple of different service requirements by using the model to obtain the backoff window increasing multiple under the optimal system performance requirement, so as to solve the problem that the backoff window increasing multiple cannot be reasonably changed when a channel conflicts.

Description

Wireless local area network performance enhancing method based on neural network algorithm
Technical Field
The invention relates to the field of wireless communication and the technical field of machine learning, in particular to a method for enhancing the performance of a wireless local area network based on a neural network algorithm.
Background
In the field of wireless communication, the bandwidth of a wireless channel given by a radio management committee is limited, which inevitably causes channel collision, thereby affecting the communication quality. At present, enterprises mostly adopt the carrier sense multiple access collision avoidance technology of IEEE802.11 series standards and a request to send/clear to send protocol (RTS/CTS) mechanism to reduce wireless channel collision and realize access to a shared channel. The basic idea of the carrier sense multiple access collision avoidance mechanism is that when collision occurs during access, a random backoff process is generated, the value of a backoff counter is generated through a backoff window, the backoff window becomes larger according to a binary exponential backoff strategy, and the value of the backoff counter is reduced to zero along with the idle state of a channel, and the device accesses the channel again. However, the backoff window is increased twice by a fixed multiple, and the node priority, the packet retransmission rate, and other factors are not fully considered, which may have a large negative impact on the system throughput and the system delay. The method improves the performance of the system to a certain extent, but the back-off window increase multiple cannot be reasonably changed, cannot meet the compromise of various performance parameters of the system, can cause the increase of the system delay performance to be smaller than the decrease of the system throughput performance, and the like, and cannot be effectively predicted, so that the back-off window increase multiple cannot be reasonably changed to the problem which needs to be solved urgently when the channel conflicts.
Disclosure of Invention
In order to solve the problem that the back-off window increase times can not be reasonably changed when the channels conflict, the invention provides a wireless local area network performance enhancing method based on a neural network algorithm, which comprises the following steps:
acquiring and recording system performance parameters under different backoff window increasing multiples when the number of the sending nodes is fixed through an existing protocol system, repeating the acquisition and recording process for multiple times, and taking the average value of the system performance parameters as final original data;
normalizing the obtained original data to keep the magnitude of the normalized data within a preset range;
dividing the normalized data into three types of data according to a proportion, wherein the three types of data are respectively training data, verification data and test data;
establishing a nonlinear regression model by using a neural network, wherein the neural network architecture is divided into an input layer, a hidden layer and an output layer;
training and modeling input parameters and output parameters for multiple times by adopting a neural network algorithm, and saving the training model when the mean square error loss function value output by the model meets the requirement;
and (3) performing data prediction on the system performance parameters by using the stored training model, performing denormalization processing on model output data after the data prediction is performed, wherein the processed data is real prediction data.
Testing the original data by adopting a training model, and directly storing the model as a final network model if the regression fitting degree of the model to the original data is good; and if the regression fitting degree of the model is poor, abandoning the model, reconstructing the neural network architecture until obtaining the model meeting the requirements, storing the model and taking the model as a final model.
Selecting the back-off window increasing times in a collision avoidance mechanism, predicting and comparing system performance parameters of different back-off window increasing times by using a final model, obtaining the back-off window increasing times which enable the system performance to be superior through comparison, and applying the back-off window increasing times to a carrier sense multiple access collision avoidance technical mechanism.
The protocol system refers to an existing IEEE802.11 protocol system.
Wherein the system performance parameters include system throughput, system delay, and packet retransmission rate.
The normalization processing means that the output data after the normalization processing is equal to a value obtained by adding the product value of the difference between the upper limit of the normalization range and the lower limit of the normalization range and the first preset value to the lower limit of the normalization range.
Wherein the neural network comprises a feed-forward neural network, a recurrent neural network, and a feedback neural network.
The data of the back-off window increasing multiple after the normalization processing is used as the input parameters of an input layer, and the data of the system throughput, the system delay and the data packet retransmission rate are used as the output parameters of an output layer.
The hidden layer can be set to be a plurality of hidden layers according to requirements.
The normalization processing means that the output data after the normalization processing is equal to a value obtained by adding the product value of the difference between the upper limit of the normalization range and the lower limit of the normalization range and the first preset value to the lower limit of the normalization range.
Wherein the neural network comprises a feed-forward neural network, a recurrent neural network, and a feedback neural network.
The data of the back-off window increasing multiple after the normalization processing is used as the input parameters of an input layer, and the data of the system throughput, the system delay and the data packet retransmission rate are used as the output parameters of an output layer.
Wherein the mean square error function value is in the order of 10-3The following.
When the backoff window is selected to increase the multiple, the system performance parameters need to be comprehensively considered.
The invention has the beneficial effects that: and establishing a model between the back-off window increase multiple and the system throughput, the system delay and the data packet retransmission rate through a neural network algorithm, and predicting the back-off window increase multiple of different service requirements through the model so as to obtain the back-off window increase multiple under the optimal system performance requirement.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
Fig. 1 is a schematic flow structure diagram of a method for enhancing the performance of a wireless local area network based on a neural network algorithm according to the present invention.
Fig. 2 is a schematic flow chart and structure diagram of a wlan performance enhancing method based on a neural network algorithm according to an embodiment of the present invention.
Fig. 3 is a schematic structural diagram of a neural network model architecture of a wireless local area network performance enhancement method based on a neural network algorithm.
Detailed Description
Reference will now be made in detail to embodiments of the present invention, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to the same or similar elements or elements having the same or similar function throughout. The embodiments described below with reference to the drawings are illustrative and intended to be illustrative of the invention and are not to be construed as limiting the invention.
Referring to fig. 1 to 3, the present invention provides a technical solution:
a wireless local area network performance enhancing method based on a neural network algorithm comprises the following steps:
s101: acquiring and recording system performance parameters under different backoff window increasing multiples when the number of the sending nodes is fixed through an existing protocol system, repeating the acquisition and recording process for multiple times, and taking the average value of the system performance parameters as final original data;
the protocol system refers to the existing IEEE802.11 protocol system.
The system performance parameters include system throughput, system delay, and packet retransmission rate.
In the present embodiment, increasing the data size of the raw data allows the training model to more accurately reflect the relationship between the input parameters and the output parameters. It should be noted that the system performance parameters are not limited to the system throughput, the system delay and the data packet retransmission rate, and all parameters affecting the performance of the wlan can be collected and recorded as system performance parameters.
S102: normalizing the obtained original data to keep the magnitude of the normalized data within a preset range;
the normalization processing means that the output data after the normalization processing is equal to a value obtained by adding the product value of the difference between the upper limit of the normalization range and the lower limit of the normalization range and the first preset value to the lower limit of the normalization range.
In the present embodiment, the normalized output data satisfies the following equation:
Figure BDA0002276965100000061
ymaxis the upper limit of the normalized range, yminThe lower limit of the normalized range is defined, x is a size value in the normalized data, y is the normalized output data, the data magnitude is specifically set according to the required accuracy, the first preset value is a value obtained by obtaining a difference value between a normalized data value and a normalized data minimum value, the difference value is a value obtained by removing the difference value between the normalized data maximum value and the normalized data minimum value, and the preset range is artificially determined and preferably is [0, 1]]。
S103: dividing the normalized data into three types of data according to a proportion, wherein the three types of data are respectively training data, verification data and test data;
in the present embodiment, the ratio is determined artificially, and is preferably 7: 1.5: 1.5, wherein the training data accounts for 70% of the total normalized data, and the verification data and the test data respectively account for 15% of the total normalized data.
S104: establishing a nonlinear regression model by using a neural network, wherein the neural network architecture is divided into an input layer, a hidden layer and an output layer;
the neural network includes a feedforward neural network, a recurrent neural network, and a feedback neural network.
And taking the backoff window increase multiple data after the normalization processing as input parameters of an input layer, and taking system throughput, system delay and data packet retransmission rate data as output parameters of an output layer.
The hidden layer can be set as at least one hidden layer according to requirements.
In the embodiment, a feedforward neural network is used for nonlinear regression prediction, a hidden layer neuron should select a nonlinear activation function to improve the approximation capability of the network, and the network selects a fast-convergence training function to improve the convergence speed of the network.
S105: and training and modeling the input parameters and the output parameters for multiple times by adopting a neural network algorithm, and storing the training model when the mean square error loss function value output by the model meets the requirement.
The magnitude range of the mean square error function value is 10-3The following.
In the present embodiment, the mean square error function is typically on the order of 10-3It is preferable that if the mean square error value of the model output is always kept at a larger magnitude, the number of hidden layers should be increased, or other activating functions, training functions and neural network models should be selected to reconstruct the network architecture.
S106: and (3) performing data prediction on the system performance parameters by using the stored training model, performing denormalization processing on model output data after the data prediction is performed, wherein the processed data is real prediction data.
In the present embodiment, the denormalization is to obtain real prediction data by inverting the obtained result by using a normalization processing formula to obtain a numerical value.
S107: testing the original data by adopting a training model, and directly storing the model as a final network model if the regression fitting degree of the model to the original data is good; and if the regression fitting degree of the model is poor, abandoning the model, reconstructing the neural network architecture until obtaining the model meeting the requirements, storing the model and taking the model as a final model.
In this embodiment, all backoff window increasing multiples in the original data are sequentially input into the final model to obtain output data, and denormalization processing is performed on the output data, where denormalization processing coefficients correspond to normalization processing coefficients of the network model one to one, and the denormalized data is compared with system throughput, system delay, and data packet retransmission rate data in the original data in a drawing manner to check the regression fitting degree of the final model. If the regression fitting curve is basically overlapped with the original data curve, the reliability and the usability of the final model are high, and the final model can be output; and if the regression fitting degree is poor and the original data curve cannot be accurately corresponded, rebuilding a neural network framework, and carrying out model training until obtaining a model with good regression fitting degree.
S108: selecting the back-off window increasing times in a collision avoidance mechanism, predicting and comparing system performance parameters of different back-off window increasing times by using a final model, obtaining the back-off window increasing times which enable the system performance to be superior through comparison, and applying the back-off window increasing times to a carrier sense multiple access collision avoidance technical mechanism.
In the present embodiment, the predictable system performance includes system throughput, system delay, packet retransmission rate, etc., and the system performance parameters should be considered comprehensively when selecting the backoff window increase factor.
The specific embodiment is as follows:
s201: according to the existing IEEE802.11 protocol system, system performance data such as system throughput, system delay, data packet retransmission rate and the like under the condition that different backoff windows are increased by multiple when the number of fixed sending nodes is 20 are collected and recorded;
in the embodiment, the data acquisition is performed on an MATLAB simulation software platform, an MATLAB simulation carrier sense multiple access collision avoidance technology mechanism is used for acquiring data, and the increase multiple of a backoff window in the carrier sense multiple access collision avoidance technology mechanism is adjusted to be increased from 1.1 times to 20 times according to 0.1 step length, so that 200 groups of original data are generated. And storing corresponding system performance data under different backoff window increasing multiples, executing the simulation process of the same backoff window increasing multiple 100 times, averaging 100 results, and storing the value as original data.
S202: after the original data are obtained, carrying out normalization processing on the data, and normalizing the original data to be in a range of [0,1 ];
the normalization process satisfies the following formula:
Figure BDA0002276965100000091
wherein y ismaxIs that the upper limit of the normalized range is 1, yminThe lower limit of the normalization range is 0, x is the size of the input data, and y is the output data after normalization processing.
S203: classifying the data after the normalization processing, wherein the neural network is divided according to the following steps: 1.5: 1.5, wherein the training data accounts for 70% of the total number of the normalized data, and the verification data and the test data respectively account for 15% of the total number of the normalized data;
s204: determining the type of the neural network, and building a neural network architecture. The method is built on the basis of a BP neural network in a feedforward neural network algorithm;
in this embodiment, the network architecture is sequentially divided into an input layer, a hidden layer and an output layer. The input parameters of the input layer are the backoff window increasing times, and the number of the neurons is 1; the hidden layer can be set into multiple hidden layers according to requirements, in this embodiment, the hidden layer is set into a double hidden layer, wherein the number of neurons in the hidden layer 1 is 8, and the number of neurons in the hidden layer 2 is 5; the output parameters of the output layer are system throughput, system delay and data packet retransmission rate, and the number of neurons is 3. The activation functions from the network input layer to the hidden layer 1 and from the hidden layer 1 to the hidden layer 2 are both non-linear activation functions Relu, and the activation functions from the hidden layer 2 to the output layer are linear activation functions Purelin. The training function selects a fast converging Levenberg-Marquardt optimization algorithm. The loss function is a mean square error loss function.
S205: using a training network model after the network construction is finished;
in this embodiment, first, the network mean square error threshold is set to 10-4The maximum iteration number is 1000, the normalized ability check value is 15, and the network learning rate is 0.01. The input parameters and the output parameters are then input into a network training network model. And judging whether the training model meets the requirement of the network mean square error according to the output result of the MATLAB neural network tool, checking a Regression graph while meeting the requirement, checking the corresponding relation between the output value of the training, verifying and testing data and a target value, and storing the network model of which all data points are close to the 45-degree angle bisector.
S206: verifying the stored network model by using the original data;
in the present embodiment, all backoff windows in the original data are sequentially input to the model by increasing the multiple, and the output data of the network model is obtained. And performing denormalization processing on the output data, wherein denormalization processing coefficients correspond to normalization processing coefficients of the network model one to one. And (4) carrying out drawing comparison on the denormalized data and the data of the system throughput, the system delay and the data packet retransmission rate in the original data, and checking the regression fitting degree of the network model. If the regression fitting curve is basically overlapped with the original data curve, the reliability and the usability of the network model are high, and the network model can be output as a final network model; and if the regression fitting degree is poor and the original data curve cannot be accurately corresponded, rebuilding a neural network framework, and training until a network model with good regression fitting degree is obtained.
It should be noted that if the network regression fitting degree is still not ideal after the neural network architecture is rebuilt, the number of neurons in the hidden layer of the network and the number of layers of the hidden layer need to be changed, and the types of the training function and the nonlinear activation function can be changed simultaneously. If an effective network model cannot be obtained by adopting the method, the type of the neural network is reselected.
S207: and intelligently predicting the system performance parameters under different backoff window increasing multiples by using the trained neural network model.
To select the optimum back-off window increasing times; meanwhile, according to the system performance priority considered by the user, a proper backoff window increasing multiple is selected to meet the performance compromise among all parameters of the system performance index, so that the reasonable selection of the system performance parameters is realized.
For example: if the user gives priority to the performance of the system throughput, when the backoff window is selected to be increased by multiple, the performance indexes such as system delay, data packet retransmission rate and the like can be properly reduced for improving the system throughput; similarly, if the user gives priority to the performance of system delay or packet retransmission rate, other performance index parameters of the system can be properly reduced to satisfy the priority system performance index.
The invention relates to a wireless local area network performance enhancing method based on a neural network algorithm, which is characterized in that a model between a backoff window increasing multiple and system throughput, system delay and data packet retransmission rate is established through the neural network algorithm, and backoff window increasing multiples of different service requirements are predicted through the model so as to obtain the backoff window increasing multiple under the optimal system performance requirement.
While the invention has been described with reference to a preferred embodiment, it will be understood by those skilled in the art that various changes in form and detail may be made therein without departing from the spirit and scope of the invention as defined by the appended claims.

Claims (8)

1. A wireless local area network performance enhancing method based on a neural network algorithm is characterized by comprising the following steps:
according to the existing protocol system, collecting and recording system performance parameters under different backoff window increasing multiples when the number of the sending nodes is fixed, repeating the collecting and recording process for multiple times, and taking the average value of the system performance parameters as final original data;
normalizing the obtained original data to keep the magnitude of the normalized data within a preset range;
dividing the normalized data into three types of data according to a proportion, wherein the three types of data are respectively training data, verification data and test data;
establishing a nonlinear regression model by using a neural network, wherein the neural network architecture is divided into an input layer, a hidden layer and an output layer;
training and modeling input parameters and output parameters for multiple times by adopting a neural network algorithm, and saving the training model when the mean square error loss function value output by the model meets the requirement;
performing data prediction on system performance parameters by using a stored training model, performing denormalization processing on model output data after the data prediction is performed, wherein the data obtained after the denormalization processing is real prediction data;
testing the original data by adopting a training model, and directly storing the model as a final network model if the regression fitting degree of the model to the original data is good; if the regression fitting degree of the model is poor, abandoning the model, reconstructing a neural network architecture until obtaining the model meeting the requirements, storing the model and taking the model as a final model;
selecting the back-off window increasing times in a collision avoidance mechanism, predicting and comparing system performance parameters of different back-off window increasing times by using a final model, obtaining the back-off window increasing times which enable the system performance to be superior through comparison, and applying the back-off window increasing times to a carrier sense multiple access collision avoidance technical mechanism.
2. The method as claimed in claim 1, wherein the protocol system is an existing IEEE802.11 protocol system.
3. The method of claim 1, wherein the system performance parameters include system throughput, system delay, and packet retransmission rate.
4. The method as claimed in claim 1, wherein the normalization process is performed by adding the product of the difference between the upper limit of the normalization range and the lower limit of the normalization range and the first predetermined value to the lower limit of the normalization range.
5. The method of claim 1, wherein the neural network comprises a feedforward neural network, a recurrent neural network, and a feedback neural network.
6. The method of claim 3, wherein the backoff window increased by a multiple of the normalization processing is used as an input parameter of an input layer, and the system throughput, the system delay and the packet retransmission rate are used as output parameters of an output layer.
7. The method of claim 1, wherein the hidden layer is set as at least one hidden layer according to requirements.
8. The method of claim 1, wherein the mean square error function has a magnitude in the range of 10-3The following.
The method of claim 1, wherein system performance parameters are considered in combination when selecting the backoff window increase factor.
CN201911126396.3A 2019-11-18 2019-11-18 Wireless local area network performance enhancing method based on neural network algorithm Pending CN111163530A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201911126396.3A CN111163530A (en) 2019-11-18 2019-11-18 Wireless local area network performance enhancing method based on neural network algorithm

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201911126396.3A CN111163530A (en) 2019-11-18 2019-11-18 Wireless local area network performance enhancing method based on neural network algorithm

Publications (1)

Publication Number Publication Date
CN111163530A true CN111163530A (en) 2020-05-15

Family

ID=70555966

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201911126396.3A Pending CN111163530A (en) 2019-11-18 2019-11-18 Wireless local area network performance enhancing method based on neural network algorithm

Country Status (1)

Country Link
CN (1) CN111163530A (en)

Citations (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102143551A (en) * 2011-03-29 2011-08-03 华北电力大学 Wireless competition access control backoff method based on network load forecast
CN102256266A (en) * 2011-07-04 2011-11-23 重庆邮电大学 User application-oriented adaptive access network selection device and method
CN102421151A (en) * 2011-11-08 2012-04-18 哈尔滨工业大学 Minimum contention window adaptive adjustment method based on WLAN (Wireless Local Area Network) contention station number
CN102592171A (en) * 2011-12-30 2012-07-18 南京邮电大学 Method and device for predicting cognitive network performance based on BP (Back Propagation) neural network
CN103857055A (en) * 2014-03-20 2014-06-11 南京航空航天大学 Back-off parameter designing method for realizing fairness of wireless self-organizing network links
CN104185298A (en) * 2014-08-29 2014-12-03 华侨大学 Network load dynamic adaptive parameter adjusting method based on priorities
US20150098387A1 (en) * 2013-10-08 2015-04-09 Parallel Wireless, Inc. Parameter Optimization and Event Prediction Based on Cell Heuristics
CN105007596A (en) * 2015-08-13 2015-10-28 重庆大学 Method for designing WSANs (Wireless Sensor and Actuator Networks) packet loss decision device based on QoS (Quality of Service)
CN108428012A (en) * 2018-03-12 2018-08-21 株洲联诚集团控股股份有限公司 A kind of fan noise prediction technique of optimization neural network
CN109060716A (en) * 2018-08-31 2018-12-21 湖南农业大学 Near-infrared characteristic spectrum Variable Selection based on the competitive adaptive weight weight sampling strategy of window
CN109618302A (en) * 2019-01-25 2019-04-12 东华大学 Novel conflict supervision avoidance strategy and method towards car networking distributed MAC protocol
CN110336620A (en) * 2019-07-16 2019-10-15 沈阳理工大学 A kind of QL-UACW back-off method based on MAC layer fair exchange protocols
US20190339685A1 (en) * 2016-05-09 2019-11-07 Strong Force Iot Portfolio 2016, Llc Methods and systems for data collection, learning, and streaming of machine signals for analytics and maintenance using the industrial internet of things
CN110458195A (en) * 2019-07-10 2019-11-15 河海大学 A kind of abnormal data discriminating method based on the fusion of more criterions
CN110460953A (en) * 2019-08-12 2019-11-15 内蒙古大学 A kind of car networking back-off method and its device based on sender's packet loss differentiating mechanism

Patent Citations (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102143551A (en) * 2011-03-29 2011-08-03 华北电力大学 Wireless competition access control backoff method based on network load forecast
CN102256266A (en) * 2011-07-04 2011-11-23 重庆邮电大学 User application-oriented adaptive access network selection device and method
CN102421151A (en) * 2011-11-08 2012-04-18 哈尔滨工业大学 Minimum contention window adaptive adjustment method based on WLAN (Wireless Local Area Network) contention station number
CN102592171A (en) * 2011-12-30 2012-07-18 南京邮电大学 Method and device for predicting cognitive network performance based on BP (Back Propagation) neural network
US20150098387A1 (en) * 2013-10-08 2015-04-09 Parallel Wireless, Inc. Parameter Optimization and Event Prediction Based on Cell Heuristics
CN103857055A (en) * 2014-03-20 2014-06-11 南京航空航天大学 Back-off parameter designing method for realizing fairness of wireless self-organizing network links
CN104185298A (en) * 2014-08-29 2014-12-03 华侨大学 Network load dynamic adaptive parameter adjusting method based on priorities
CN105007596A (en) * 2015-08-13 2015-10-28 重庆大学 Method for designing WSANs (Wireless Sensor and Actuator Networks) packet loss decision device based on QoS (Quality of Service)
US20190339685A1 (en) * 2016-05-09 2019-11-07 Strong Force Iot Portfolio 2016, Llc Methods and systems for data collection, learning, and streaming of machine signals for analytics and maintenance using the industrial internet of things
CN108428012A (en) * 2018-03-12 2018-08-21 株洲联诚集团控股股份有限公司 A kind of fan noise prediction technique of optimization neural network
CN109060716A (en) * 2018-08-31 2018-12-21 湖南农业大学 Near-infrared characteristic spectrum Variable Selection based on the competitive adaptive weight weight sampling strategy of window
CN109618302A (en) * 2019-01-25 2019-04-12 东华大学 Novel conflict supervision avoidance strategy and method towards car networking distributed MAC protocol
CN110458195A (en) * 2019-07-10 2019-11-15 河海大学 A kind of abnormal data discriminating method based on the fusion of more criterions
CN110336620A (en) * 2019-07-16 2019-10-15 沈阳理工大学 A kind of QL-UACW back-off method based on MAC layer fair exchange protocols
CN110460953A (en) * 2019-08-12 2019-11-15 内蒙古大学 A kind of car networking back-off method and its device based on sender's packet loss differentiating mechanism

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
张延年: "基于ＲBF 神经网络的无线信道退避算法优化设计", 《计 算 机 与 现 代 化》 *

Similar Documents

Publication Publication Date Title
CN111867139B (en) Deep neural network self-adaptive back-off strategy implementation method and system based on Q learning
CN109245840B (en) Frequency spectrum prediction method based on convolutional neural network in cognitive radio system
CN113242568A (en) Task unloading and resource allocation method in uncertain network environment
CN110839184B (en) Method and device for adjusting bandwidth of mobile fronthaul optical network based on flow prediction
CN107949025B (en) Network selection method based on non-cooperative game
CN109976915B (en) Edge cloud collaborative demand optimization method and system based on edge computing
CN111328133A (en) V2X resource allocation method based on deep neural network
CN112383369A (en) Cognitive radio multi-channel spectrum sensing method based on CNN-LSTM network model
Lei et al. Learning-based resource allocation: Efficient content delivery enabled by convolutional neural network
CN110996365A (en) Heterogeneous network vertical switching algorithm and system based on multi-objective optimization model
CN105007596B (en) A kind of WSANs packet loss decision device design method based on QoS
CN111163530A (en) Wireless local area network performance enhancing method based on neural network algorithm
CN110751278A (en) Neural network bit quantization method and system
Dao et al. A new Markov model for non-saturated 802.11 networks
Mahmoudi et al. Cost-efficient distributed optimization in machine learning over wireless networks
US11489560B2 (en) Method of parameter estimation for a multi-input multi-output system
CN113038612A (en) Cognitive radio power control method based on deep learning
CN115379508A (en) Carrier management method, resource allocation method and related equipment
CN111124439B (en) Intelligent dynamic unloading algorithm with cloud edge cooperation
Zhang et al. Machine learning based protocol classification in unlicensed 5 GHz bands
CN117290090A (en) Node cooperative resource allocation method, device, equipment and storage medium
Zhao et al. C-LSTM: CNN and LSTM Based Offloading Prediction Model in Mobile Edge Computing (MEC)
Strelkovskaya et al. Multimedia traffic prediction based on wavelet-and spline-extrapolation
CN111010695A (en) Channel allocation method based on channel idle time prediction
CN114666917A (en) Method for optimizing time delay of LTE (Long term evolution) system in unauthorized frequency band

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
RJ01 Rejection of invention patent application after publication
RJ01 Rejection of invention patent application after publication

Application publication date: 20200515