UAV-Assisted MEC Request Scheduling via Virtualization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
5G communication networks face challenges such as severe signal attenuation, small coverage, and susceptibility to obstacle interference due to their short wavelength, which affect service delay and network quality in IoT environments.
Innovation Solution
A method for request scheduling in a UAV-assisted MEC network involves dividing a target area into blocks, determining feasible UAV deployment points, dispatching UAVs with small cell base stations, and using a round-robin policy to allocate user requests, with adjustments based on network pressure and historical data input into a multi-timescale LSTM neural network for predictive redeployment.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If 5G small cell base stations are deployed adjacent to IoT devices for mobile edge computing, then service delay is reduced and processing speed is improved, but network coverage remains limited and signal attenuation is severe
Solution Approach 1:
The patent introduces UAVs (unmanned aerial vehicles) as mobile edge computing nodes operating in three-dimensional space above the ground. This vertical dimension expansion allows the network to overcome the limited ground-based coverage of 5G small cells while maintaining low latency processing capabilities. The UAVs fly at elevated positions to provide broader coverage areas and bypass ground-based signal blockages.
Solution Approach 2:
The system dynamically deploys and relocates UAVs based on real-time network pressure and user request patterns. The UAVs can move to different locations to serve varying demand hotspots, transforming the static ground-based small cell infrastructure into a dynamic, adaptable network that can expand coverage flexibly without permanent infrastructure installation.
2Reliability
If UAVs are deployed to expand network coverage and reduce service delay, then network quality is improved, but system complexity increases due to deployment optimization requirements
Solution Approach 1:
The patent creates virtual copies of UAV resources through virtualization technology. Multiple virtual UAVs can be instantiated from a single physical UAV, allowing the system to manage resource allocation and scheduling more flexibly. This virtualization simplifies the control system by abstracting the complexity of physical UAV management while maintaining improved network quality through multiple service points.
Solution Approach 2:
The system dynamically adjusts UAV deployment parameters such as flight altitude, horizontal position, and service radius based on real-time network conditions and predicted user request patterns. By changing these parameters adaptively, the system optimizes network quality without requiring complex permanent infrastructure, managing complexity through software-based parameter control rather than hardware complexity.
3Productivity
If multiple UAVs are dispatched to handle high network pressure, then service capacity is improved, but resource allocation complexity and scheduling difficulty increase
Solution Approach 1:
The patent divides the target service area into multiple blocks and assigns specific UAVs to serve particular blocks or groups of blocks. This segmentation of the service region simplifies resource allocation by creating manageable zones with dedicated UAV resources, reducing the scheduling complexity that would arise from managing all UAVs across the entire area simultaneously while maintaining high service capacity through parallel operations in different blocks.
Solution Approach 2:
The system uses historical user request data and machine learning models to predict future network pressure patterns and pre-position UAVs in anticipated high-demand areas before actual demand occurs. This preliminary action allows the system to proactively manage resource allocation, reducing the complexity of reactive scheduling while ensuring adequate service capacity is available when needed.
Data Source
AI summary
A method for request scheduling in an unmanned aerial vehicle-assisted mobile edge computing network: determining multiple feasible UAV deployment points based on obstruction information in a target area; randomly selecting U feasible UAV deployment points from the multiple feasible UAV deployment points as UAV deployment points; dividing each UAV into multiple virtual UAVs; allocating user requests in a central queue of a UAV-assisted MEC to the virtual UAVs; using a round-robin policy to schedule the user requests allocated to each virtual UAV; after a specified time period t, if a network pressure mitigation condition isn't met, inputting historical user request data of each feasible UAV deployment point into a trained MT-LSTM neural network model, obtaining a to-be-processed data volume of each feasible UAV deployment point in a next specified time period t; re-determining UAV deployment points; redeploying UAVs carrying small cell base stations based on the new UAV deployment points.