SAGIN Layered Offloading With Adaptive Slicing for Traffic Shifts
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing space-air-ground integrated networks (SAGIN) face challenges in resource management, limited computation capabilities, and high complexity in computation offloading, which hinder efficient service provision for computation-intensive applications due to dynamic network environments and limited resources in satellites and unmanned aerial vehicles.
Innovation Solution
A traffic-aware lightweight layered offloading framework is introduced, dividing the SAGIN into communication access platforms (CAP) and computation offloading platforms (COP), using network slicing to manage resources, with adaptive slicing enabled by deep reinforcement learning (DRL) and policy distillation to optimize task offloading and resource allocation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If computation resources are deployed at ground base stations for MEC, then network bandwidth pressure is reduced, but coverage and service continuity are limited due to fixed infrastructure
Solution Approach 1:
The patent segments the MEC infrastructure into multiple distributed edge nodes deployed across ground base stations, air vehicles (UAVs, aircraft), and satellites. This segmentation allows the system to maintain ground-based computation efficiency while extending coverage through aerial and space-based nodes, resolving the contradiction between reducing bandwidth pressure and improving coverage adaptability.
Solution Approach 2:
The patent transitions from traditional two-dimensional ground-based MEC to a three-dimensional space-air-ground integrated MEC architecture. By adding vertical dimensions (air and space layers), the system maintains the bandwidth pressure reduction benefit of ground MEC while achieving global coverage and service continuity through multi-dimensional node distribution.
2Adaptability or versatility
If satellites and UAVs are equipped with computation resources for SAGIN, then global coverage is achieved, but resource costs increase due to limited onboard capabilities
Solution Approach 1:
The patent introduces ground base stations as intermediary nodes that provide computation support for satellite and UAV tasks. When onboard computation resources on satellites or UAVs are insufficient, tasks are offloaded to ground-based edge nodes with stronger computation capabilities. This intermediary mechanism enables global coverage while managing resource costs by leveraging ground infrastructure for heavy computation.
Solution Approach 2:
The patent merges the computation resources of ground base stations, air vehicles, and satellites into a unified MEC system. This combination allows the system to achieve global coverage by pooling resources across different platforms, where each platform contributes its strengths while sharing the overall computation burden, thereby controlling resource costs.
3Productivity
If deep reinforcement learning is used for adaptive slicing, then resource allocation is optimized, but system complexity increases
Solution Approach 1:
The patent segments the network slicing management into multiple independent deep reinforcement learning agents, each responsible for specific slices or network segments. This segmentation reduces the complexity of the overall system by dividing the complex resource allocation problem into smaller, more manageable sub-problems while maintaining optimized resource allocation through distributed decision-making.
4Quantity of substance
If network slicing is implemented for resource virtualization, then resource utilization improves, but management complexity increases
Solution Approach 1:
The patent introduces an intelligent controller as an intermediary that manages network slicing using deep reinforcement learning. This controller automatically performs slice creation, resource allocation, and dynamic adjustment based on real-time network conditions, thereby improving resource utilization while reducing management complexity by replacing manual processes with automated AI-driven management.
Data Source
AI summary
Provided is a traffic-aware lightweight layered offloading framework for an adaptive slicing-enabled space-air-ground integrated network (SAGIN), where the adaptive slicing-enabled SAGIN is divided into a communication access platform (CAP) and a computation offloading platform (COP), and resources on each of the CAP and the COP are managed by network slicing; an edge service provider (ESP) provides computation offloading while performing resource allocation; for the resource allocation, a dynamic traffic change is captured by using ProbSparse self-attention, and adaptive network slicing is executed in accordance with predicted traffic and a system load; and for the computation offloading, a communication process and a computation process are separated to allocate a sub-channel as required in accordance with a channel state, then a virtual machine is allocated to a task through a lightweight computation offloading algorithm, and a converged policy is extracted as a lightweight neural network for online inference.