Wireless System with ML-Based Dynamic Resource Allocation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current wireless communication systems face challenges in optimizing network performance across heterogeneous wireless systems, including 4G/5G/WiFi, and managing resource allocation efficiently to ensure high Quality of Service (QoS) and throughput, especially in dynamic and diverse environments such as satellite, mobile, and unlicensed band communications.
Innovation Solution
The implementation of machine learning algorithms that predict future network conditions by analyzing data on network resource usage, signal quality, and traffic patterns, allowing for dynamic adjustments of network parameters, including Dynamic Spectrum Sharing (DSS) and energy-efficient strategies, with AI-driven methods for resource block allocation and bandwidth management.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional resource allocation methods are used in heterogeneous wireless systems, then system complexity is reduced, but network performance optimization and QoS assurance deteriorate
Solution Approach 1:
The patent implements self-service through machine learning models that automatically predict network conditions and resource allocation requirements without human intervention. The system trains on historical network data and autonomously makes resource allocation decisions, allowing the network to serve itself and optimize performance dynamically.
Solution Approach 2:
The patent replaces traditional mechanical resource allocation mechanisms with AI-based predictive systems. Instead of using fixed rules and manual configuration, the system uses machine learning algorithms to predict future network states and automatically adjust resource allocation, substituting conventional control mechanisms with intelligent automation.
2Productivity
If dynamic resource allocation is implemented to optimize QoS, then network efficiency improves, but signaling overhead increases
Solution Approach 1:
The patent applies preliminary action by using machine learning models to predict future network conditions and resource allocation needs before they actually occur. By forecasting traffic patterns and network states in advance, the system can pre-allocate resources and prepare configurations, reducing the need for frequent signaling updates and minimizing overhead.
3Productivity
If machine learning models are deployed for predictive resource allocation, then throughput increases, but computational requirements and energy consumption increase
Solution Approach 1:
The patent segments the machine learning workload by deploying lightweight models at the network edge (in network devices) while using more complex models for centralized training and analysis. This segmentation allows throughput optimization at the edge with minimal energy consumption, while heavy computational tasks are performed centrally where energy resources are more abundant.
4Measurement precision
If feedback mechanisms are implemented for continuous refinement, then prediction accuracy improves, but system complexity and processing time increase
Solution Approach 1:
The patent implements continuous refinement through ongoing feedback mechanisms where the machine learning models continuously learn from new network data and adjust their predictions. The system maintains continuous training and updating processes, allowing the models to improve accuracy over time while operating continuously without interruption to network service.
Data Source
AI summary
A method of facilitating network access for local Terminal Equipment (TE) involves utilizing a Mobile Terminal (MT) that is pre-equipped with a list of authorized TE identities and a user identity module. Upon receiving an identity authentication signal from the TE, the MT assesses the TE's identity against the authorized list through multifactor verification. If the TE is verified as authorized, the MT retrieves a unique identifier from the user identity module and conveys this identifier to the TE. Subsequently, the TE employs the provided identifier to access the network.