Sep 2026· IEEE Transactions on Mobile Computing· Vol 25, pp. 13863-13879· 0 citations· 47 references
Abstract
As Satellite Edge Computing (SEC) emerges as a computing backbone of 6 G IoT, its feasibility is threatened by frequent handovers that disrupt task continuity and cause deadline violations in AI-driven workloads. Conventional communication-centric handover strategies overlook the urgency of computational tasks, further exacerbating these issues. To address this gap, we propose the <bold>C</bold>omputation <bold>A</bold>ware <bold>C</bold>onditional <bold>H</bold>andover <bold>O</bold>ptimization (CACHO) framework, a non-intrusive solution fully compatible with 3GPP Conditional Handover (CHO). The framework integrates model partition driven delayed execution to enable seamless migration via inter-satellite tensor transmission, timeout-risk predictive proactive triggering to protect at-risk tasks from overloaded satellites, and task-load-aware target selection to balance computational workloads and service popularity. These modules are orchestrated within standardized CHO workflows without altering the core logic, ensuring deployability in existing satellite network systems. Simulations based on real Starlink constellation data demonstrate that our approach reduces task timeout rates by 29.95–51.02<inline-formula><tex-math notation="LaTeX">$\%$</tex-math><alternatives><mml:math><mml:mo>%</mml:mo></mml:math><inline-graphic xlink:href="xu-ieq1-3674471.gif"/></alternatives></inline-formula> compared to the 3GPP standard CHO and several state-of-the-art baselines, while maintaining handover success rates and minimizing extra handovers needed. Beyond performance gains, this work provides the first blueprint for embedding task life-cycle awareness into handover workflows, bridging mobility management and computation guarantees, and advancing the feasibility of robust SEC for mission-critical 6G IoT applications.
Vehicular edge computing (VEC), a key enabler for the Internet of Things (IoT) in intelligent transportation, addresses onboard processing constraints through collaborative task offloading among vehicles, facilitating latency-sensitive applications such as autonomous driving. However, developing efficient offloading strategies remains particularly challenging in high-density vehicular networks, where intensive computational demands coexist with severely constrained intervehicle communication ranges due to signal blockage. To handle this, we propose M4O, a mobility-aware task offloading framework supporting multihop, multiuser, and multitask offloading optimization. M4O intelligently integrates vehicle mobility patterns and enables relay-assisted offloading to enhance system effectiveness and robustness. The framework employs a dual-algorithm approach: the advantage actor–critic (A2C) for indivisible tasks and the hybrid proximal policy optimization (H-PPO) for divisible tasks, both optimized to minimize the temporally coupled composite cost of time and resources. Extensive experiments demonstrate that the deep reinforcement learning (DRL)-based solutions of M4O deliver stable and efficient offloading strategies, outperforming existing benchmarks by significant margins in cost efficiency. Our code is available at https://github.com/Zhouym1028/M4O
Momiao Zhou, Yimin Zhou, Yanshi Sun et al.· IEEE Internet of Things Jour...· 0 citations
: The rapid growth of electric vehicle (EV) charging infrastructures has introduced new challenges in monitoring abnormal load behaviors under strict latency and resource constraints. Conventional anomaly detection approaches either rely on centralized processing or incur excessive false alarms, limiting their practical applicability in large-scale deployments. This paper proposes a hierarchical edge-fog anomaly detection framework that integrates lightweight edge-level filtering with a fog-level Temporal Convolutional Network (TCN) detector. The edge component suppresses non-informative patterns, while the fog layer performs temporal modeling on selectively forwarded data. This design enables controllable reduction of fog-level processing load. Under corrected end-to-end evaluation on real-world EV charging load data, the hierarchical pipeline should be interpreted as a system operating point rather than a uniformly superior detector. Relative to fog-only TCN-AE inference, the selected routing policy reduces fog workload by 34.9% and shortens average detection delay from 93.6 to 75.6 h, but increases false alarms per day from 0.88 to 7.29 and lowers F1 from 0.547 to 0.455. Sensitivity experiments over routing thresholds reveal a consistent trade-off among fog workload, alert burden, detection delay, and retained anomaly evidence. Additional routing diagnostics show that the primary source of performance degradation is information loss induced by filtering, rather than weakness of the fog detector on the forwarded subset. These findings suggest that hierarchical edge intelligence is a practical but calibration-sensitive direction for scalable anomaly monitoring in EV charging infrastructures.
H. Jeong· Computers, Materials & C...· 0 citations
In an era dominated by data-driven solutions, Deep Neural Networks (DNNs), which have been proven to be pivotal tools in extensive applications across various domains, are evolving in terms of both depth and architecture to meet the escalating demands of contemporary utilizations. Nevertheless, deploying a complex DNN model on mobile devices may result in substantial processing latency and increasing energy consumption. The emerging Mobile Edge Computing (MEC), characterized by the allocation of computing capacity at the access point, enables the partitioning of DNN models so as to conserve energy on mobile devices and mitigate inference latency. Existing DNN partitioning methods typically train prediction models offline to make partition decisions and reduce end-to-end inference latency, which requires a great number of labeled datasets and may incur a prolonged pre-processing duration. In this paper, we develop an online Deep Reinforcement Learning (DRL) based adaptive partition method to dynamically determine optimal partitioning decision so as to jointly accelerate DNN inference and mitigate energy consumption. We run the proposed algorithm in an edge computing scenario consisting of NVIDIA Jetson Nano and an edge server equipped with RTX3090 for four different DNN models, including VGG16, MobileNetV2, ResNet50 and GPT2-medium. Then we collect actual processing latency and energy consumption and compare the performance of the proposed algorithm with state-of-the-art solutions. The experimental results demonstrate that, even under varying channel conditions, DAPart can achieve an average reduction of 38.8% in latency and 36.5% in energy consumption compared with other available methods.
Shubin Zhang, Junrong Ma, Kaikai Chi et al.· ACM Transactions on Sensor N...· 0 citations
A constrained optimization model that supports different management goals through alternative objective functions (latency-aware or power-aware) while enforcing operational constraints, including node capacities, slice-specific latency bounds, and explicit limits on VNF migrations/relocations between scheduling periods is proposed.
R. Moreno-Vozmediano, E. Huedo, R. Montero et al.· Journal of Network and Syste...· 0 citations
Abstract Recent studies show that global Passive Optical Network (PON) deployments will reach over 1.3 billion subscribers by 2030in Optical Network Unit (ONU) device integration. Despite this, legacy systems suffer from scalability bottlenecks and average latency increments of 15–30 % during user density surges. Existing networks face significant challenges, such as latency spikes at the Optical Line Terminal (OLT) during ONU scalability and insufficient reliability in handling anomalies within converged infrastructures. To address these issues, a novel Multi-Armed Bandit (MAB) approach is applied to optimize dynamic bandwidth allocation (DBA) at the ONU layer, enabling increased user density without inducing latency burdens at the OLT. The MAB-based selection strategy efficiently adapts to varying traffic patterns by learning optimal resource assignment policies in real time, ensuring minimal contention delays and better quality of service (QoS). Network fault tolerance and reliability are enhanced through a Deep Q-Auto Encoder (DQAE)-based anomaly detection model trained to recognize and classify failure signatures across optical and packet layers. This unsupervised deep reinforcement learning model integrates reconstruction loss with Q-learning to identify unknown failure states simultaneously and recommend proactive recovery actions. The combined strategy improves user scalability and service stability and establishes a fault-resilient infrastructure suitable for next-generation converged optical networks.
K.Tara Phani, K. Kumari· Journal of optical communica...· 0 citations
The emergence of AI-native 6G networks necessitates efficient and proactive resource management mechanisms to support highly dynamic and data-intensive services. Existing approaches typically employ independent prediction models for mobility, handover, and channel quality, leading to redundant data processing and limited exploitation of cross-layer dependencies. In this paper, we propose a shared data intelligence–driven multi-task prediction framework that jointly models mobility, handover, and channel quality indicator (CQI) within a unified learning architecture. By leveraging a common feature space and a single processing pipeline, the proposed framework simultaneously generates multiple correlated predictions, thereby reducing computational overhead. A correlation analysis using real-world datasets demonstrates that mobility, CQI, and handover events exhibit inherent inter-dependencies, justifying the use of a shared representation. Furthermore, a processing time comparison shows that the proposed approach achieves approximately 10.12% reduction compared to conventional independent prediction models by eliminating redundant feature extraction and repeated model execution. These results validate that shared data intelligence is an effective and scalable solution for efficient multi-task prediction in real-time 6G network environments.
K. Hasan, Seong-Ho Jeong· International Conference on...· 0 citations
Related blog posts
Microsoft Research Blog· microsoft.comAug 31, 2026
What if pathology foundation models could do more with less? GigaPath-Flash and GigaTIME-Flash cut computational demands while maintaining strong performance, opening the door to larger studies and broader exploration. The post GigaPath-Flash and GigaTIME-Flash: Toward population-scale discovery with efficient pathology foundation models appeared first on Microsoft Research.
MIT News · Artificial Intelligence· news.mit.eduAug 31, 2026
With millions of users across the world, Julia has been used to conduct cutting-edge research and to design new drugs, jet engines, heat pumps, and more.
MIT News · Artificial Intelligence· news.mit.eduAug 27, 2026
A new machine-learning framework aims to improve the success rate of computational protein design while moving away from results that reproduce sequences found in nature.