AU2020103373A4 - Machine learning based network intelligentization for automatically- configurable cellular communication systems - Google Patents
Machine learning based network intelligentization for automatically- configurable cellular communication systems Download PDFInfo
- Publication number
- AU2020103373A4 AU2020103373A4 AU2020103373A AU2020103373A AU2020103373A4 AU 2020103373 A4 AU2020103373 A4 AU 2020103373A4 AU 2020103373 A AU2020103373 A AU 2020103373A AU 2020103373 A AU2020103373 A AU 2020103373A AU 2020103373 A4 AU2020103373 A4 AU 2020103373A4
- Authority
- AU
- Australia
- Prior art keywords
- networks
- machine learning
- network
- intelligentization
- smart
- 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.)
- Ceased
Links
Classifications
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
- G06N20/10—Machine learning using kernel methods, e.g. support vector machines [SVM]
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/16—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks using machine learning or artificial intelligence
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W24/00—Supervisory, monitoring or testing arrangements
- H04W24/02—Arrangements for optimising operational condition
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W28/00—Network traffic management; Network resource management
- H04W28/02—Traffic management, e.g. flow control or congestion control
- H04W28/08—Load balancing or load distribution
- H04W28/09—Management thereof
- H04W28/0958—Management thereof based on metrics or performance parameters
- H04W28/0967—Quality of Service [QoS] parameters
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
- H04W4/30—Services specially adapted for particular environments, situations or purposes
- H04W4/38—Services specially adapted for particular environments, situations or purposes for collecting sensor information
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W84/00—Network topologies
- H04W84/18—Self-organising networks, e.g. ad-hoc networks or sensor networks
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W28/00—Network traffic management; Network resource management
- H04W28/02—Traffic management, e.g. flow control or congestion control
- H04W28/08—Load balancing or load distribution
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W72/00—Local resource management
- H04W72/04—Wireless resource allocation
Landscapes
- Engineering & Computer Science (AREA)
- Signal Processing (AREA)
- Computer Networks & Wireless Communication (AREA)
- Theoretical Computer Science (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Medical Informatics (AREA)
- Computing Systems (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Databases & Information Systems (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Quality & Reliability (AREA)
- Mobile Radio Communication Systems (AREA)
- Telephonic Communication Services (AREA)
Abstract
MACHINE LEARNING BASED NETWORK INTELLIGENTIZATION
FOR AUTOMATICALLY- CONFIGURABLE CELLULAR
COMMUNICATION SYSTEMS
ABSTRACT
With the growing growth of intelligent interfaces and communications infrastructure, as well as
diversification technologies with vibrant criteria, existing networks will not be sufficient to entirely
satisfy progressively increasing traffic criteria. As a result, analysis on 6 G networks has also been
conducted by all industries and education. Artificial Intelligence has recently been used as a
modem model for the architecture and automation of 6 G networks with a high degree of
intelligence. A-enabled smart infrastructure for 6 G networks for information discovery,
intelligent resource management, automated network adaptation and automated service delivery,
where infrastructure is split into four levels: wearable devices layer, data mining and analytics
layer, remote communication module and smart device layer. The proposed methodology has been
implemented as the ultimate strategy in a variety of highly complicated situations. Network
Intelligentization would be a modem development to solve the complexities of an increasingly
growing number of distributed devices linked. Even so, relative to the use of machine learning in
other areas, such as video gaming, recent development on smart networking also has a significant
way to go to achieve to realise automated cellular communication networks. Numerous challenges
in spite of communication structures, machine learning technologies and computing performance
should be tackled for the best utilization of this approach in 6G. Al methods for 6 G improve
network efficiency, reliably and efficiently, namely Al-powered cellular device networking, smart
convergence and upload control, and intelligent spectral efficiency. The significant upcoming
exploration goals are spotlighted and possible strategies for Al-enabled smart 6 G networks, like
computational output, rigorous algorithms, equipment creation and resource management.
1| P a g e
MACHINE LEARNING BASED NETWORK INTELLIGENTIZATION
FOR AUTOMATICALLY- CONFIGURABLE CELLULAR
COMMUNICATION SYSTEMS
Drawings:
-: -------------- --- M
is!SeniccPrvi~mc +
I -- ----- ----- ---k t indus mi
RC U I kFlIW tIoiII SDN CenIetOdm1oud Edgc Cloud
- I M:-d1ih fig ICnniItg
~I C -n'ieg r, euiF
~i DAs taitin
Figure 1: The architecture of Al-enabled intelligent 6G networks.
11P ag e
Description
Drawings:
is!SeniccPrvi~mc
+ k t indus mi I -- ----- ----- ---
RC U I kFlIW tIoiII SDN CenIetOdm1oud Edgc Cloud
- I M:-d1ih fig ICnniItg
~I C -n'ieg r, euiF
~i DAs taitin
Figure 1: The architecture of Al-enabled intelligent 6G networks.
11P ag e
Description
Field of the Invention:
This invention relates to Machine Learning Based Network Intelligentization for Automatically Configurable Cellular Communication Systems. The architecture of 6 G networks would be large scale, multi-layer, elevated-complex, diverse and monolithic. In conjunction, 6 G networks continue to promote consistent networking to ensure a wide variety of QoS specifications for a massive number of devices and manage massive volumes of data extracted from physical surroundings. Al technologies with strong computational skills, learning skills, optimizing capabilities and intelligent identification capabilities can be used in 6 G networks to sensibly execute output management, information exploration, specialized learning, structure alignment and difficult decision-making.
Background of the invention:
Existing research has studied the implementation of machine learning (ML) technologies in wireless communications and identify the numerous contributions of ML to cellular connectivity issues including radio access network and spectrum sensing. However, those articles do not discuss configuration problems and network architecture obstacles for associating ML strategies for wireless network implementations. The next version of the mobile networking infrastructure (5 G) would deal with a vast range of wired smartphones at base stations, a huge rise in traffic flow and a wide variety of technologies with diverse specifications and criteria.
The diversity of systems and software enables technology infrastructure much more complicated. Nevertheless, traditional machine learning methods remain restricted to natural data processing of the actual shape. For several decades, the development of a machine learning system or a pattern recognition system involved significant experience in the field and meticulous innovation to
1| P a g e develop a feature extractor. During that stage, the actual data could be transformed into an accurate description to be utilized as feedback to the training mechanism.
According to Ahmed, a few of the experiments that have incorporated deep learning and augmented training to explore the issue of available resources in wireless networks. Many challenges and shortcomings associated with resource utilization, such as throughput maximization, interruption minimization, and energy conservation, have been investigated. While the research is concentrated on the topic of resource utilization, this paper provides a more general comprehensive analysis of the various deep learning models being used 5 G networks. We also discuss other challenges in the 5 G networks that include the use of various deep learning frameworks.
In accordance with Zhang a detailed survey is conducted on the use of deep learning in Mobile broadband networks. Zhang focused more broadly on solving concerns relevant to generic wireless networks like accessibility analysis, wireless sensor network localization, WSN data analysis, and others. Our comprehensive analysis focuses on 5 G wireless networks and potential environments, implementations, and threats. Deep learning models presented in the studied work discuss basic cellular network issues including channel state information, transfer management, spectrum allocation. The situations discussed in the work we choose are also connected to the 5 G networks and affect the suggested deep learning approach.
In accordance with, the consideration is given to the ability of Al to solve the challenge of channel allocation in communication technology is considered. It has been observed that the A-based approach has displayed greater efficiency than that of randomized heuristic and genetic algorithms (GAs). The collection of radio access technologies (RAT) using the Hopfield neural networks as a decision-making mechanism thus incorporating the skill of Al reasoning and the application of multi-parameter decisions using the IEEE 802.21 protocol possibilities is described.
According to patent W02017214271A, an effective management and configuring cellular networking systems and carrier network problems is essential for wireless telephone providers to offer reliable coverage to their customers. Typically, often cellular telecommunications providers use main efficiency measures or other forms of network output data to assess the output of the mobile network and alleviate perceived issues. For example, upon issuance of a problem receipt
21Page for call quality control problems, call quality for one or perhaps more access points may be resolved by supplying premium features per customer on requests. Even then, those techniques can have substantial lead-time gaps that minimise customer service and can trigger investment and operational expenditures.
According to K. F. Poon, an A-based framework that optimizes graph theory-based problem frameworks for the home fiber infrastructure is implemented to simplify the development process. Mixed integer linear programming (MILP), ant colony optimization and a genetic algorithm were used to analyze the challenges. The concepts of Al for radio access network (RAN) preparation involves a new cellular, radio frequency and 5 G wireless network spectrum. The research is carried out by the aggregation of input information from various outlets, by learning-based grouping, estimation and clustering models, and the extraction of applicable information to inform decisions taken by the 5 G network.
The 6G is expected to carry network softwarization to a different level, notably towards network Intelligentization. In 5 G, the "non-radio" element has emerged further and more critical and has become the main force underlying the latest "softwarization" initiative. More precisely, two main G developments are Software-Defined Networking (SDN) and Network Functions Virtualization, which have shifted existing communication infrastructure to software-based virtual networks. They also permit cellular networks, which will offer a valuable virtualization feature to encourage several virtual networks to be built on common infrastructure facilities. However, as the network is growing more dynamic and interdependent, softwarization is not expected to be appropriate for 6G. In specific, to implement Al-based systems, network organizations need to embrace a range of technologies, including messaging, information caching, computation, and even wireless power transmission.
Objects of the Invention:
The main objective is to implement the Machine Learning Based Network Intelligentization for Automatically- Configurable Cellular Communication Systems.
The second objective is to implement prevalent machine learning methods used in cellular connections are adequately outlined, considering their key fundamentals and generalized
31 Pa g e implementations that are categorized as supervised learning, unsupervised learning, reinforcement learning, NNs and transfer learning. The Neural Networks and transfer learning are illustrated individually due to their growing relevance to wireless communication technology.
Summary of the Invention:
In relation, 6 G will approve new radio connectivity interfaces like THz telecommunications and smart structures. Further assistance would also be required advanced automation features, such sensing, data processing, analysis, and storage. Both above problems call for architecture and that is versatile, resilient and, most significantly, smart. Current technologies, such as SDN, NFV and network slicing, would require to be progressively improved to address obstacles. Al techniques with strong computational skills, learning skills, optimizing capabilities and smart identification capabilities may be used in 6 G networks to intelligently execute output management, information exploration, specialized learning, configuration alignment and difficult decision-making. With the support of Al, we introduce the A-enabled smart architecture for 6 G networks, which is broadly composed of four layers: smart sensing layer, data mining and analytics layer, smart control layer and smart device layer. Sensing and tracking are typically other very rudimentary activities of 6 G networks, where 6 G networks appear to intelligently track and track information from real environments by large equipment or masses of human beings. Al-enabled sensing and detection is responsible of intelligently capturing vast quantities of dynamics, distinct and versatile information by specifically communicating with the physical world, like, but not limited to, radio frequency utilization tracking, environmental control, spectrum sensing, intruder detection and disturbance detection.
Detailed Description of the Invention:
Figure 1: The architecture of Al-enabled intelligent 6G networks.
Figure 2: The design of Centralized machine learning system with softwarization and virtualization
Figure 3: The layer structure for machine learning based applications
Figure 4: Four steps for developing intelligent algorithms
41Page
Detailed Description of the Invention:
The architecture of 6 G networks would be large-scale, multi-layer, high-complex, diverse, and highly integrated. In addition, 6 G networks intend to promote consistent integration to ensure a wide variety of QoS standards across many users, as well as handle a massive volume of information created from the physical world. Figure 2 shows the Centralized machine learning system with softwarization and virtualization. In a centralized machine learning system with softwarization and virtualization, a sophisticated central controller method a machine learning algorithm to recognize worldwide network enhancement premised on the relevant data acquired. Centralized control, even then, varies depending on continuously data obtained, which creates massive signaling and overhead computing particularly in a large-scale 5G/6 G network. In addition, the end-to - end interruption of incorporated data and control massages may contribute to unpredictable events, control transmission delay and synchronization issues. The major difference among centralized global precision and high operating costs for 6 G ultra-low interruption networks should be properly examined.
Figure 3 illustrates the OSI layer structure of 6G. The OSI framework is built with a layer hierarchy and the communicating features are separated into fragments that are difficult to optimize worldwide. Many machine learning techniques are adopted to mitigate network operations independently, including unsupervised learning for both ends and internet traffic aggregating in the network layer, deep learning channel distribution in the media access control layer, recurrent Neural Network (RNN) based traffic flow simulation in the transmission layer, and (RL) based reinforcement learning Adaptive Computation offloading algorithm in the application layer.
Figure 4 mention the four steps of Al based Intelligentization. An initial major step should be machine-based problem definition, which has a significant effect on training techniques and difficulties. As we know, unsupervised learning is typically embraced as an agglomeration concern, whereas supervised learning is for the intent of linear regression and classification. If the dilemma be the Markov Decision Process, reinforcement learning is the most effective approach. After choosing how to prepare, the next is to select the right machine learning algorithms and
1P a g e describe its input and output that is complicated and important. Defining the cost function is therefore important as it influences the duration of preparation and the failure of estimation. The third step is to maximize the performance of the machine learning model that has been developed. It comprises of several things, to mention some, the configuration change, the option of activation functions, the initialization operations and the activation function.
61Page
1. Machine Learning Based Network Intelligentization for Automatically- Configurable Cellular Communication Systems consist of Enlist a diminished cellular power transmission configuration; Obtaining adjacent cell details from sensor nodes; In addition to the decision that the existing specification of the wireless base station is in dispute with the adjacent cell data, the reconfiguration of the wireless base station centered at least in part on the adjacent cell relevant data; In accordance to the determination that the reconfiguration was effective, the cellular transmitted capacity increased gradually. 2. An apparatus for use with a wireless base station, comprising: at least one processor; memory storing a program of instructions, wherein the memory storing the program of instructions is configured. Control the wireless base station based on input to reduce transmit power cellular mode that causes the wireless base station to receive neighboring cell information from user equipment; 3. In addition to claim 1 automatically- Configurable Cellular Communication System consist of machine learning methodologies like supervised learning, unsupervised learning and reinforcement learning for network Intelligentization. 4. From claim 3, the automatically- Configurable Cellular Communication System comprises of Intelligent Sensing Layer Data Mining and Analytics Layer Intelligent Control Layer
1 Pag e
Claims (1)
- Smart Application Layer 5. According to claim 3, 6 G networks ought to promote ultra-high consistency and ultra low latency information providers. In turn, complex 6 G networks give rise to frequency feature ambiguity, which ensures that stable and precise detecting is very difficult to accomplish. Al techniques can provide reliable, real-time and reliable spectrum sensing, where fuzzy SVM and non-parallel hyper-plane SVM are robust to environmental uncertainty, and CNN-based collaborative detection can boost sensing precision with low computational complexity. 6. From claim 4, Data Mining and Analytics Layer is a key activity that intends to manage and analyses vast quantities of raw information recorded from a vast range of devices in 6 G networks, and to achieve semantic extraction and exploration of information. Huge data obtained from physical systems may be diverse, non-linear, and high-dimensional, such that data mining and modeling can be used in 6 G networks to solve the complexities of accessing large volumes of information, and to solve the complexities of accessing large volumes of data. 7. From claim 4, the intelligent control layer consists mainly of training, automation and decision-making, where this layer uses relevant information from lower layers to allow mass agents to intelligently learn, optimize and determine the most relevant dual-function behaviour to help different social network utilities. 8. From claim 4, Smart Application Layer handles all operations of smart devices, interfaces and infrastructures in 6 G networks via Al approaches to realize the capacity of the application to self-organize.2|PageMACHINE LEARNING BASED NETWORK INTELLIGENTIZATION FOR AUTOMATICALLY- CONFIGURABLE CELLULAR COMMUNICATION SYSTEMSDrawings: 2020103373Figure 1: The architecture of AI-enabled intelligent 6G networks.1|PageFigure 2: The design of Centralized machine learning system with softwarization and virtualizationFigure 3: The layer structure for machine learning based applications2|PageFigure 4: Four steps for developing intelligent algorithms3|Page
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
AU2020103373A AU2020103373A4 (en) | 2020-11-11 | 2020-11-11 | Machine learning based network intelligentization for automatically- configurable cellular communication systems |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
AU2020103373A AU2020103373A4 (en) | 2020-11-11 | 2020-11-11 | Machine learning based network intelligentization for automatically- configurable cellular communication systems |
Publications (1)
Publication Number | Publication Date |
---|---|
AU2020103373A4 true AU2020103373A4 (en) | 2021-01-28 |
Family
ID=74192031
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
AU2020103373A Ceased AU2020103373A4 (en) | 2020-11-11 | 2020-11-11 | Machine learning based network intelligentization for automatically- configurable cellular communication systems |
Country Status (1)
Country | Link |
---|---|
AU (1) | AU2020103373A4 (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
DE202022104275U1 (en) | 2022-07-28 | 2022-08-25 | Ahmed Alemran | System for intelligent resource management for distributed machine learning tasks |
CN115225540A (en) * | 2022-05-02 | 2022-10-21 | 东北大学 | Software defined network-oriented data plane fault detection and recovery method |
-
2020
- 2020-11-11 AU AU2020103373A patent/AU2020103373A4/en not_active Ceased
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN115225540A (en) * | 2022-05-02 | 2022-10-21 | 东北大学 | Software defined network-oriented data plane fault detection and recovery method |
CN115225540B (en) * | 2022-05-02 | 2023-07-18 | 东北大学 | Data plane fault detection and recovery method for software defined network |
DE202022104275U1 (en) | 2022-07-28 | 2022-08-25 | Ahmed Alemran | System for intelligent resource management for distributed machine learning tasks |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Kato et al. | Ten challenges in advancing machine learning technologies toward 6G | |
Wang et al. | Artificial intelligence enabled wireless networking for 5G and beyond: Recent advances and future challenges | |
Xu et al. | Trust-aware service offloading for video surveillance in edge computing enabled internet of vehicles | |
Yu et al. | Intelligent edge: Leveraging deep imitation learning for mobile edge computation offloading | |
Slama | Prosumer in smart grids based on intelligent edge computing: A review on Artificial Intelligence Scheduling Techniques | |
Liu et al. | Deep learning meets wireless network optimization: Identify critical links | |
CN109814951A (en) | The combined optimization method of task unloading and resource allocation in mobile edge calculations network | |
Faisal et al. | Machine learning approaches for reconfigurable intelligent surfaces: A survey | |
Samek et al. | The convergence of machine learning and communications | |
AU2020103373A4 (en) | Machine learning based network intelligentization for automatically- configurable cellular communication systems | |
Ding et al. | A GAN-based intrusion detection model for 5G enabled future metaverse | |
Li et al. | Rlops: Development life-cycle of reinforcement learning aided open ran | |
WO2023219177A1 (en) | Integrated sensing and communications empowered by networked hybrid quantum-classical machine learning | |
Rui et al. | CLPM: A cooperative link prediction model for industrial Internet of Things using partitioned stacked denoising autoencoder | |
Tervonen et al. | A review of the cognitive capabilities and data analysis issues of the future industrial Internet-of-Things | |
CN114943284A (en) | Data processing system and method of behavior prediction model | |
Liu et al. | Hastening stream offloading of inference via multi-exit dnns in mobile edge computing | |
Hlophe et al. | AI meets CRNs: A prospective review on the application of deep architectures in spectrum management | |
Hafi et al. | Split Federated Learning for 6G Enabled-Networks: Requirements, Challenges and Future Directions | |
Trappey et al. | Patent landscape and key technology interaction roadmap using graph convolutional network–Case of mobile communication technologies beyond 5G | |
US20230368065A1 (en) | Integrated Sensing and Communications Empowered by Networked Hybrid Quantum-Classical Machine Learning | |
Lockhart et al. | Scission: Context-aware and performance-driven edge-based distributed deep neural networks | |
Reddy et al. | A Beginner's Guide to Federated Learning | |
Dai et al. | Offloading dependent tasks in edge computing with unknown system-side information | |
Bellavista et al. | Edge Cloud as an Enabler for Distributed AI in Industrial IoT Applications: the Experience of the IoTwins Project. |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
FGI | Letters patent sealed or granted (innovation patent) | ||
MK22 | Patent ceased section 143a(d), or expired - non payment of renewal fee or expiry |