CN102118803A - Video cross-layer scheduling method of mobile communication system on basis of QoE prediction - Google Patents
Video cross-layer scheduling method of mobile communication system on basis of QoE prediction Download PDFInfo
- Publication number
- CN102118803A CN102118803A CN2011100936179A CN201110093617A CN102118803A CN 102118803 A CN102118803 A CN 102118803A CN 2011100936179 A CN2011100936179 A CN 2011100936179A CN 201110093617 A CN201110093617 A CN 201110093617A CN 102118803 A CN102118803 A CN 102118803A
- Authority
- CN
- China
- Prior art keywords
- video
- user
- layer
- parameter
- mobile communication
- 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
Links
Images
Abstract
The invention discloses a video cross-layer scheduling method of a mobile communication system on the basis of QoE prediction. The method comprises the steps of: utilizing a clustering analysis method to divide videos into a slight motion, common scene switching and fast motion by content characteristics according to speed change parameter and brightness parameter of the videos; carrying out cross-layer analysis on a user application layer, network layer parameters and video content type information, and respectively establishing three user-received video quality prediction models by artificial nerve network studying method; and according to a model prediction result and a target function requirement for user-received video quality of the maximization system, adopting a PSO (Particle Swarm Optimization) algorithm to carry out the resource scheduling and distribution. Due to adoption of the technical scheme, the problem that the resource utilization ratio is low and the subjective perception effect of the user to the video quality is not fully considered in the existing mobile communication system video services can be solved, and the cross-layer resource scheduling of the video services based on a QoE evaluation method is realized.
Description
Technical field
The present invention relates to the mobile communication technology field, is to be used for next generation mobile communication system (IMT-Advanced) based on the QoE evaluation method video transmission business to be striden layer resource scheduling, and realization system user subjective quality is estimated optimized technical scheme.
Background technology
Along with the globalization of 3G network deployment and the maturation gradually of operation mode, the standard formulation of next generation mobile communication system IMT-Advanced and the research of key technology receive publicity day by day.ITU thinks that IMT-Advanced system (4G) has the new ability above IMT-2000, can be that the user provides that fast data inserts, unified information and various broadband multimedia services with novel interactive more form of service, business is more paid attention to property of participation, interactive, payes attention to the effect of service experience more.For different business service types, the QoS demand difference of customer requirements, the service quality feeling way is also different.
The video class business is a kind of High Data Rate and the strong business of content relevance, and the user is also higher for the quality of service requirement of video class business.In wireless environment, because the time variation of channel and the finiteness of Radio Resource cause video traffic efficiency of transmission on time and space low, serious waste of resources, thus the development of video traffic and application are restricted.Given this, thus video traffic service quality effectively detected and Internet resources are dispatched and optimized to improve service quality based on testing result become the critical problem that network provider, service provider need to be resolved hurrily.
At present, detect at video traffic service quality and the research of the network optimization is based on mostly that QoS carries out.ITU-T defines QoS at first for " resultant effect of the service performance of decision user satisfaction ", has comprised a plurality of aspects content comparatively widely.But present stage, usually with QoS narrow sense be interpreted as the performance index of bottom packet data transmission, as time delay, shake, bandwidth, the error rate etc.Main being responsible for of QoS mechanism carried out service management and guaranteed professional otherness from the angle of network, and network entity is handled different business according to different quality requirements.Adopting QoS is a kind of pure objective method as the detection and the optimization method of major parameter, main objective parameter on calculating according to the transmission network that obtains, do not consider to comprise other objective parameter such as video quality, video content, do not consider people's vision system and people's subjectivity experience factor yet.Therefore, adopting these class methods can not accurately obtain user's active service experiences, thereby can not carry out real-time network resource scheduling according to current service quality timely, cause Internet resources to be effectively utilized, the raising of Video service quality is restricted.In the case, notion and the correlative study of QoE (Quality of Experience) are arisen at the historic moment.
QoE can be understood as user experience or user's perception, i.e. the subjective feeling of terminal use's service feature that the mobile network is provided.It can be by representing experience and the impression of terminal use to professional and network and reflection current business and the quality of network and the gap of user expectation near the method that quantizes.For video traffic, the Video service quality is most important, weighing the most accurately should be from user's actual video service experience, promptly to the measurement of subjective parameters.Pure subjective measurement method all is just can be measured after business is finished mostly, the requirement of can not requirement of real time high video traffic.Therefore, be necessary to design a kind of effective ways that adopt retrievable objective parameter prediction and analog subscriber to experience for the subjectivity of video, set up the video tastes forecast model, thereby can realize the cross-layer scheduling and the optimization of video traffic based on the user QoE parameter that machine dopes.
Towards aspect the scheduling of resource and optimization of IMT-Advanced, cross-layer scheduling method makes network provide the better service quality support for the user under the situation of parameter dynamic change and resource-constrained by exchange message between each layer of protocol stack.Traditional cross-layer scheduling method mainly is towards the RRM of striding physical layer, link layer, towards many net coexistences of striding physical layer, link layer and network layer.These methods mainly adopt QoS as parameter, do not have to consider to comprise the parameter of videos such as video quality, video content itself, and people's vision system and people's subjectivity experience factor, therefore obtain the video client service experience in view of the above accurately.
This problem is based on the present Research of domestic and international association area, proposes a kind ofly based on the QoE appraisement system, is applicable to the cross-layer scheduling method of next generation mobile communication system video traffic, to reach the target of maximization system user video-aware quality.
Summary of the invention
The present invention is intended at the professional resource utilization that exists of video transmission in the present next generation mobile communication system low, and it is a kind of with the cross-layer scheduling scheme based on video content of QoE as the video quality evaluation standard for the situation proposition of quality of service perception not take into full account the user.
To achieve these goals, solve the corresponding techniques problem, the present invention realizes the corresponding techniques scheme by following process:
Step 1: utilize to extract video information and also adopt clustering method that institute's information extraction is analyzed, the video of server end is divided three classes according to the content character difference.
Employed video information process is for to analyze at original video in the step 1, extraction rate changes parameter (temporal signatures) and luminance video parameter (spatial feature), then utilize the classifying content estimation model that institute's extracting parameter is carried out cluster analysis, video is divided into light exercise, general scene switching, rapid movement three classes by content the most at last.
Step 2: take all factors into consideration application layer parameter, subscriber channel situation and user video type, utilize the artificial neural network learning method to set up user's receiver, video prediction of quality model respectively at three types videos to server requests.
Frame rate FR, Transmit Bit Rate SBR (application layer parameter) have been taken all factors into consideration in the step 2; Mistake packet rate PER, service bandwidth BW (network layer parameter); Video content classification situation (step 1 result) parameter utilizes machine learning method to set up three kinds of user's receiver, video prediction of quality models respectively.
Step 3: with maximization system user receiver, video quality is optimization aim, is reference with the video quality forecast model that obtains in the step 2, is constraints with the minimum threshold value of bearing of user's quality of reception, the objective definition function:
Max ∑ (ω
sQ
s+ ω
gQ
g+ ω
rQ
r) be confined to Q 〉=3.5
Step 4: be the required optimum parameters of each user determined value of initialization (carrying out value among the 30fps 10,15) in span as regulation FR among the present invention.Use the PSO particle cluster algorithm to find the solution the optimal solution of utility function.
Employed PSO algorithm is a kind of evolutionary computation technique based on animal population in the step 4, adopts the optimized scheduling of resource process of this algorithm solving system user video quality of reception in this example.Also can adopt other optimized Algorithm for trying to achieve the optimal resource allocation result.
Step 5:, under restrictive condition, try to achieve each scheduling of resource allocation result of each user successively according to the method for mentioning in the step 4.
As can be seen from the above technical solutions, technical scheme of the present invention is by the method predictive user receiver, video quality of machine self study, with the maximization user to the perceived quality of video traffic as target function, cross-layer scheduling video content and user application layer, network layer parameter, can reach the raising resource utilization ratio, promote the purpose of user the quality of service perception.
Below by accompanying drawing and specific embodiments technical scheme of the present invention is further set forth.
Description of drawings
In order to set forth embodiments of the invention and existing technical scheme more clearly, below the explanation accompanying drawing of using in technical scheme explanation accompanying drawing of the present invention and the description of the Prior Art is done simple introduction.Obviously, under the prerequisite of not paying creative work, those of ordinary skills can obtain other accompanying drawing by this accompanying drawing.
Fig. 1 is a video content categorizing system structural representation in the specific embodiment of the invention.
Fig. 2 is user's receiver, video prediction of quality flow chart in the specific embodiment of the invention.
Fig. 3 is user resources scheduling flow figure in the specific embodiment of the invention.
Embodiment
Clearer for what technical scheme advantage of the present invention was described, below in conjunction with accompanying drawing the specific embodiment of the present invention is described in further detail.
Fig. 1 is a video content categorizing system structural representation in the specific embodiment of the invention.As shown in Figure 1, this classifying content system comprises that original video input unit 101, velocity variations parameter extraction unit 102, brightness change parameter extraction unit 103, classifying content estimation model 104 and visual classification output unit 105 as a result.
The original video input unit is responsible for that video to be transmitted is sent into system and is carried out classifying content.
Velocity variations parameter extraction unit is used for extracting the velocity variations parameter of video time domain, and basic skills is to calculate the absolute value sum of the difference of two interframe pixel values.
The brightness that brightness variation parameter extraction unit user extracts in the video spatial domain changes parameter, and basic skills is the mould value of the difference of calculating two frame mean flow rates.
The cluster estimation model as input, will seem 102,103 result of calculation by the cluster multi-variate statistical analysis video with different characteristic and be divided three classes according to the feature of inherence:
Light exercise: refer to that basic scene does not change, the video that has only some details to change.As news hookup;
General scene is switched: refer to have the video that the continuous scene of low speed is switched.On the way walk as the personage;
Rapid movement: refer to the video that the whole video scene is switched fast.As the ball match video.
Visual classification output unit is as a result exported this system with video through the video information of classification.
Fig. 2 is user's receiver, video prediction of quality flow chart in the specific embodiment of the invention.As shown in Figure 2, this forecasting process may further comprise the steps:
Step 201: video is imported user's receiver, video prediction of quality system.
Step 202: extract the user application layer relevant parameter, comprise frame rate FR, Transmit Bit Rate SBR.
Step 203: extract with the visual classification information of exporting in Unit 105.
Step 204: extract the user network layer parameter, comprise service bandwidth BW, mistake packet rate PER.
Step 205: in the parameter extracted in 202,203,204 steps input artificial neural network learning model, the method by the machine self study draws the video quality predicted value.
To adopt artificial neural net machine self-learning method to set up forecast model in this process.Artificial neural net by simulating human for the non-linear relation between each factor of the perception of video quality and then analyzing influence video quality.This model is set up three kinds of different video quality forecast models according to visual classification result and application layer, network layer parameter for the influence of video quality.Learn to draw video quality predicted value Q for the customer parameter of importing in each scheduling process:
Q
s=f
s(FR, SBR, C
s, PER BW), is applicable to the light exercise video;
Q
g=f
g(FR, SBR, C
g, PER BW), is applicable to general scene Switch Video;
Q
r=f
r(FR, SBR, C
r, PER BW), is applicable to the rapid movement video.
Step 206: user's receiver, video prediction of quality value Q is exported.
Fig. 3 is user resources scheduling flow figure in the specific embodiment of the invention.As shown in Figure 3, this scheduling process may further comprise the steps:
Step 301: each user is sorted out video by video content after the video of server requests is by speed, the isoparametric extraction and analysis of brightness.Among the present invention video is divided into that light exercise, general scene switch, rapid movement three classes.
Step 302: video content classified information and user application layer, network layer relevant parameter that step 301 draws are sent into user's receiver, video prediction of quality model, draw the video quality predicted value according to three kinds of models.
Step 303: considering fairness condition between three kinds of quality forecast models and user, is target with maximization system user receiver, video quality, and the objective definition function is:
Max ∑ (ω
sQ
s+ ω
gQ
g+ ω
rQ
r) be confined to Q 〉=3.5
ω wherein
s, ω
g, ω
rBe the influence scale factor of dissimilar videos to oeverall quality.3.5 be the video quality minimum that the user can accept.
Step 304:, utilize the PSO algorithm to finish selection to the customer parameter value by the series of iterations process since No. 0 user circulation.
Step 305: calculate the historical optimized parameter value of historical optimized parameter value of user and colony, upgrade the customer parameter value according to target function.After reaching the utility function maximization, stop this scheduling process.
The above; only for the preferable embodiment of the present invention, but protection scope of the present invention is not limited thereto, and anyly is familiar with the people of this technology in the disclosed technical scope of the present invention; the variation that can expect easily or replacement all should be encompassed within protection scope of the present invention.Therefore, protection scope of the present invention should be as the criterion with the protection range of claim.
Claims (7)
1. mobile communication system video cross-layer scheduling method based on QoE prediction is characterized in that may further comprise the steps:
Step 1: utilize parameters such as velocity variations parameter that the clustering method analysis extracts and luminance video parameter from video, video is divided three classes by content character;
Step 2: utilize the artificial neural network learning method to set up user's receiver, video prediction of quality model respectively according to three types of videos;
Step 3: set up target function, determine constraints;
Step 4: each customer parameter value of initialization, use the PSO particle cluster algorithm to find the solution the utility function optimal solution.
Step 5: calculate the historical optimized parameter value of historical optimized parameter value of user and colony, upgrade the customer parameter value according to target function.After reaching the utility function maximization, stop this scheduling process.
2. a kind of mobile communication system video cross-layer scheduling method based on the QoE prediction according to claim 1 is characterized in that by clustering method video being classified in the step 1, thereby has considered the influence of video content for resource allocation scheduling.Video content extracts and comprises two parts parameter: velocity variations parameter, brightness change parameter.
3. according to a kind of mobile communication system video cross-layer scheduling method based on the QoE prediction described in the claim 1, it is characterized in that adopting in the step 1 clustering method that video is divided three classes by content: light exercise, general scene are switched, rapid movement.
4. a kind of mobile communication system video cross-layer scheduling method based on the QoE prediction according to claim 1 is characterized in that the video at three types adopts the artificial neural network learning method to set up user's receiver, video prediction of quality model respectively in the step 2.
5. a kind of mobile communication system video cross-layer scheduling method based on QoE prediction according to claim 1 is characterized in that prediction of quality model in the step 2 strides layer and considered user application layer parameter (frame rate, Transmit Bit Rate), network layer parameter (mistake packet rate, service bandwidth) and visual classification information.
6. a kind of mobile communication system video cross-layer scheduling method according to claim 1 based on the QoE prediction, it is characterized in that in the step 3 that with maximization system user receiver, video quality be optimization aim, and with the minimum threshold value of bearing of user's quality of reception is restrictive condition, sets up target function:
Max ∑ (ω
sQ
s+ ω
gQ
g+ ω
rQ
r) be confined to Q 〉=3.5.
7. a kind of mobile communication system video cross-layer scheduling method according to claim 1 based on the QoE prediction, it is characterized in that utilizing in the step 4 the PSO algorithm to find the solution target function, promptly constantly adjust the individual parameter value of user and colony's parameter value, when trying to achieve the target function optimal solution, stop this scheduling.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2011100936179A CN102118803A (en) | 2011-04-14 | 2011-04-14 | Video cross-layer scheduling method of mobile communication system on basis of QoE prediction |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2011100936179A CN102118803A (en) | 2011-04-14 | 2011-04-14 | Video cross-layer scheduling method of mobile communication system on basis of QoE prediction |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN102118803A true CN102118803A (en) | 2011-07-06 |
Family
ID=44217357
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2011100936179A Pending CN102118803A (en) | 2011-04-14 | 2011-04-14 | Video cross-layer scheduling method of mobile communication system on basis of QoE prediction |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102118803A (en) |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102291829A (en) * | 2011-09-05 | 2011-12-21 | 西安电子科技大学 | Method for developing data resources and optimally distributing frequency spectrums for user communication |
| CN102638730A (en) * | 2012-04-13 | 2012-08-15 | 北京邮电大学 | User perception based cross-layer optimization method for wireless video business |
| CN102665281A (en) * | 2012-04-13 | 2012-09-12 | 北京邮电大学 | Power distribution scheme on basis of MOS in wireless video transmission |
| CN102905380A (en) * | 2012-09-14 | 2013-01-30 | 西安交通大学 | Cross-layer scheduling method for real-time video in HSDPA (High Speed Downlink Packet Access) network |
| CN104023402A (en) * | 2014-05-28 | 2014-09-03 | 北京邮电大学 | Cross-layer resource allocation method facing user experience in open wireless network |
| CN104160659A (en) * | 2012-03-12 | 2014-11-19 | 诺基亚通信公司 | Prediction and root cause recommendations of service access quality of experience issues in communication networks |
| CN104380788A (en) * | 2012-03-30 | 2015-02-25 | 茨特里克斯系统公司 | Adaptive traffic management in cellular wireless networks |
| CN104969644A (en) * | 2013-06-20 | 2015-10-07 | 华为技术有限公司 | Resource allocation method and apparatus |
| CN105959685A (en) * | 2016-05-31 | 2016-09-21 | 上海交通大学 | Compression code rate prediction method based on video content and clustering analysis |
| CN106687995A (en) * | 2014-05-12 | 2017-05-17 | 高通股份有限公司 | Distributed model learning |
| CN103987125B (en) * | 2014-05-21 | 2017-10-20 | 西安交通大学 | The multi-user's real-time video cross-layer scheduling method optimized in HSDPA based on utility function |
| CN109165729A (en) * | 2018-08-22 | 2019-01-08 | 中科物栖（北京）科技有限责任公司 | The dispatching method and system of neural network |
| CN109302623A (en) * | 2018-11-01 | 2019-02-01 | 南京大学 | A kind of transmission method based on QoE model dynamic self-adapting video |
| WO2021026944A1 (en) * | 2019-08-09 | 2021-02-18 | 东北大学 | Adaptive transmission method for industrial wireless streaming media employing particle swarm and neural network |
-
2011
- 2011-04-14 CN CN2011100936179A patent/CN102118803A/en active Pending
Cited By (20)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102291829A (en) * | 2011-09-05 | 2011-12-21 | 西安电子科技大学 | Method for developing data resources and optimally distributing frequency spectrums for user communication |
| CN102291829B (en) * | 2011-09-05 | 2014-05-28 | 西安电子科技大学 | Method for developing data resources and optimally distributing frequency spectrums for user communication |
| CN104160659B (en) * | 2012-03-12 | 2018-06-26 | 诺基亚通信公司 | For the method and apparatus of management and the operation of communication network |
| CN104160659A (en) * | 2012-03-12 | 2014-11-19 | 诺基亚通信公司 | Prediction and root cause recommendations of service access quality of experience issues in communication networks |
| CN104380788B (en) * | 2012-03-30 | 2019-10-22 | 茨特里克斯系统公司 | Self-adapted service management in cellular radio |
| CN104380788A (en) * | 2012-03-30 | 2015-02-25 | 茨特里克斯系统公司 | Adaptive traffic management in cellular wireless networks |
| CN102638730A (en) * | 2012-04-13 | 2012-08-15 | 北京邮电大学 | User perception based cross-layer optimization method for wireless video business |
| CN102665281A (en) * | 2012-04-13 | 2012-09-12 | 北京邮电大学 | Power distribution scheme on basis of MOS in wireless video transmission |
| CN102905380A (en) * | 2012-09-14 | 2013-01-30 | 西安交通大学 | Cross-layer scheduling method for real-time video in HSDPA (High Speed Downlink Packet Access) network |
| CN102905380B (en) * | 2012-09-14 | 2015-03-04 | 西安交通大学 | Cross-layer scheduling method for real-time video in HSDPA (High Speed Downlink Packet Access) network |
| CN104969644A (en) * | 2013-06-20 | 2015-10-07 | 华为技术有限公司 | Resource allocation method and apparatus |
| CN104969644B (en) * | 2013-06-20 | 2019-04-26 | 华为技术有限公司 | Resource allocation methods and device |
| CN106687995A (en) * | 2014-05-12 | 2017-05-17 | 高通股份有限公司 | Distributed model learning |
| CN103987125B (en) * | 2014-05-21 | 2017-10-20 | 西安交通大学 | The multi-user's real-time video cross-layer scheduling method optimized in HSDPA based on utility function |
| CN104023402A (en) * | 2014-05-28 | 2014-09-03 | 北京邮电大学 | Cross-layer resource allocation method facing user experience in open wireless network |
| CN105959685A (en) * | 2016-05-31 | 2016-09-21 | 上海交通大学 | Compression code rate prediction method based on video content and clustering analysis |
| CN105959685B (en) * | 2016-05-31 | 2018-01-19 | 上海交通大学 | A kind of compression bit rate Forecasting Methodology based on video content and cluster analysis |
| CN109165729A (en) * | 2018-08-22 | 2019-01-08 | 中科物栖（北京）科技有限责任公司 | The dispatching method and system of neural network |
| CN109302623A (en) * | 2018-11-01 | 2019-02-01 | 南京大学 | A kind of transmission method based on QoE model dynamic self-adapting video |
| WO2021026944A1 (en) * | 2019-08-09 | 2021-02-18 | 东北大学 | Adaptive transmission method for industrial wireless streaming media employing particle swarm and neural network |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102118803A (en) | Video cross-layer scheduling method of mobile communication system on basis of QoE prediction | |
| CN107124375B (en) | Off-peak scheduling method, system and server for CDN (content delivery network) network bandwidth resources | |
| US10091675B2 (en) | System and method for estimating an effective bandwidth | |
| CN105794187A (en) | Predicting call quality | |
| Li et al. | Adaptive resource allocation based on the billing granularity in edge-cloud architecture | |
| Lekharu et al. | Deep learning based prediction model for adaptive video streaming | |
| CN102025732B (en) | Dynamic adaptive cognitive network quality of service (QoS) mapping method | |
| CN110519783A (en) | 5G network based on enhancing study is sliced resource allocation methods | |
| CN107087160A (en) | A kind of Forecasting Methodology of the user experience quality based on BP Adaboost neutral nets | |
| CN104426979A (en) | Distributed buffer scheduling system and method based on social relations | |
| CN110324260A (en) | A kind of network function virtualization intelligent dispatching method based on flow identification | |
| Wang | Application of DASH client optimization and artificial intelligence in the management and operation of big data tourism hotels | |
| CN103281355A (en) | Fuzzy decision-based server push method | |
| Huang et al. | Digital twin based user-centric resource management for multicast short video streaming | |
| CN117240887B (en) | Wisdom thing networking energy management platform system | |
| CN106789349B (en) | Quality of experience modeling analysis and conversation flow classification based method | |
| Huang et al. | Personalized QoE enhancement for adaptive video streaming: A digital twin-assisted scheme | |
| CN104683318B (en) | A kind of edge streaming server caching system of selection and system | |
| Shi et al. | CoLEAP: Cooperative learning-based edge scheme with caching and prefetching for DASH video delivery | |
| Beshley et al. | Quality of service management method in а heterogeneous wireless network using Big Data technology аnd mobile QoE application | |
| CN113055423B (en) | Policy pushing method, policy execution method, device, equipment and medium | |
| Lu et al. | Machine learning empowered content delivery: status, challenges, and opportunities | |
| CN113055212B (en) | Strategy pushing method, strategy execution method, device, equipment and medium | |
| Tu et al. | PSTile: Perception-Sensitivity Based 360$^\circ $ Tiled Video Streaming for Industrial Surveillance | |
| AT&T | CMSY8 |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20110706 |