Machine-Learning Device Group Offloading for Cellular-to-Wi-Fi QoS
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for offloading mobile devices between cellular and Wi-Fi networks fail to meet performance requirements due to high computing resources and lack of quality of service prioritization, especially in multi-service environments where different services have different network requirements, leading to unpredictable interference and inadequate network capability to fulfill service demands.
Innovation Solution
A computer-implemented method using machine learning to determine groups of devices that need to be transferred from a cellular network to Wi-Fi based on their respective quality of service policies, ensuring these policies are satisfied in the new network, utilizing a neural network for classification and reinforcement learning for faster training and scalability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If existing methods are used for Wi-Fi offloading, then device transfer between networks is achieved, but quality of service requirements cannot be satisfied and performance is insufficient
Solution Approach 1:
The patent changes the decision parameters from simple connectivity status to multi-dimensional parameters including QoS policies, device group characteristics, network conditions, and service requirements. This enables intelligent selection of devices for offloading based on their specific QoS needs rather than uniform treatment, resolving the contradiction between reliability and productivity.
Solution Approach 2:
The patent segments devices into groups based on their QoS requirements and characteristics before performing offloading. By dividing the device population into distinct groups with similar needs, the system can apply targeted offloading strategies that satisfy QoS requirements while optimizing overall performance.
2Loss of time
If machine learning with neural networks is used for device classification, then training time is reduced and scalability is improved, but computing resources are increased
Solution Approach 1:
The patent performs preliminary training of neural networks during system setup or offline periods, so that once trained, the models can quickly classify devices in real-time without requiring significant computational resources. This preliminary action separates the heavy training phase from the lightweight inference phase, resolving the time-resource contradiction.
Solution Approach 2:
The patent uses pre-trained neural network models that can be copied and deployed across multiple nodes without requiring retraining. By copying the learned patterns from the training phase, the system achieves fast device classification with minimal ongoing computational resource consumption.
3Speed
If devices are transferred without considering QoS policies, then offloading speed is increased, but network performance deteriorates due to unpredictable interference
Solution Approach 1:
The patent implements feedback mechanisms where QoS policy information is continuously monitored and fed back into the offloading decision process. This feedback loop ensures that devices are selected for offloading based on their actual QoS needs and network conditions, maintaining both speed and reliability.
Solution Approach 2:
The patent applies local quality analysis by evaluating individual device QoS requirements and network conditions specific to each device's location and service needs. This localized approach ensures that offloading decisions are optimized for each device's specific requirements rather than applying uniform rules, thereby maintaining network performance while achieving fast offloading.
Data Source
AI summary
A computer-implemented method, performed by a first node. The first node determines, using machine learning, one or more first groups of devices, out of a plurality of groups of devices. The one or more first groups of devices are to be transferred from a first network node operating with a first communication access technology to one or more second network nodes operating with a second communication access technology. Each of the groups the plurality has one or more respective policies pertaining to quality of service for device connection. The determining is based on the one or more first groups of devices having a highest probability to have their one or more respective policies satisfied in the second communication access technology. The first node sends an indication of the result of the determination.