
A computação de borda descentraliza o processamento de dados ao realocar a computação e o armazenamento para mais perto das fontes de dados, minimizando a latência e o uso de largura de banda em redes distribuídas. Ela integra nós de borda, micro data centers e gateways para suportar análises em tempo real. IoT aplicações e inferência de IA na periferia da rede. As considerações arquitetônicas incluem gerenciamento de recursos, orquestração, virtualização e segurança em ambientes de borda heterogêneos. Nossa compilação abaixo apresenta publicações e patentes recentes que detalham inovações em design de infraestrutura de borda, desenvolvimento de protocolos, otimização de cargas de trabalho e estratégias de implantação escaláveis, aprimorando as capacidades e a adoção de sistemas de computação de borda.
Esta é a nossa mais recente seleção de publicações e patentes mundiais em inglês sobre Computação de Borda, entre diversos periódicos científicos online, classificadas e focadas em computação de borda, computação em névoa, nó de borda, redução de latência, computação distribuída, análise de borda, IoT de borda, IA de borda, data center de borda, micro data center, orquestração de borda, segurança de borda, rede de borda, processamento em tempo real, dispositivo de borda, infraestrutura de borda, nuvem de borda, localização de dados, cache de borda, gateway de borda, gerenciamento de recursos de borda, virtualização de borda e carga de trabalho de borda.
Unified iot edge framework for lifecycle management of artificial intelligence solutions across multiple iot edge devices
Patent published on the 2026-05-21 in US under Ref US20260141312 by ZEDEDA INC [US] (Pitchai Pandi Chandran [in], Sankaran Hariharasubramanian Cinthamani [us])
Abstract: A computer-implemented method may be used to implement artificial intelligence (AI) over an Internet of Things (IoT) Edge Network. The method may include, at one or more storage devices, storing an AI model. The method may further include, at one or more hardware processing devices, using the AI model to develop an AI Edge Solution suitable for achieving a business purpose through the IoT Edge Network. The method may further include, at one or more communication devices, deploying the AI Edge So[...]
Our summary: This framework manages AI solutions across multiple IoT edge devices. It stores AI models on storage devices and processes them on hardware devices. The AI Edge Solution is deployed to various Edge Computing Devices within the IoT Edge Network.
IoT Edge, AI Solutions, Lifecycle Management, Edge Computing
Patent
Adaptive hyperdimensional computing for noise-resilient on-device time series forecasting
Patent published on the 2026-05-21 in US under Ref US20260141286 by UNIV CALIFORNIA [US] (Moreno Ivannia Gomez [us], Yu Xiaofan [us], Rosing Tajana Simunic [us])
Abstract: [0000] Some inventive concepts described herein relate to an Adaptive Hyperdimensional Forecasting System that provides a robust and lightweight solution for on-device time series forecasting by integrating Hyperdimensional Computing with a Kalman Filter. The system can combine the resilience of the Kalman Filter to sensor noise with the computational efficiency of HDC, making it highly suitable for edge deployment. The system can encode historical data into high-dimensional hypervectors and app[...]
Our summary: The Adaptive Hyperdimensional Forecasting System integrates Hyperdimensional Computing with a Kalman Filter for noise-resilient time series forecasting. It encodes historical data into high-dimensional hypervectors and uses an iterative Kalman process for weight estimation. This lightweight solution is optimized for edge devices, ensuring efficient and accurate real-time forecasts in resource-constrained environments.
Adaptive Hyperdimensional Computing, Time Series Forecasting, Kalman Filter, Edge Deployment
Patent
Multi-tenant distributed computing system and computer-implemented method for adaptive quality management in artificial intelligence (ai) processing
Patent published on the 2026-05-07 in WO under Ref WO2026094029 by ODEH IHAB [AE] (Odeh Ihab [ae], Al Shraiedeh Mahmoud [jo])
Abstract: The present invention provides a computer system and a computer-implemented method for adaptive quality management in artificial intelligence (AI) processing. The system comprises a memory unit storing machine-readable instructions and a processor configured to execute the instructions to store, in a persistent correction memory, records of prior corrections to AI outputs; retrieve a correction for a current input based on similarity between the input and stored records; generate, by an audit-tr[...]
Our summary: The invention describes a system for adaptive quality management in AI processing. It uses a persistent correction memory to store prior corrections and retrieve relevant adjustments for current inputs. The system optimizes AI workloads and updates parameters based on effectiveness, enhancing error reduction and auditability.
multi-tenant, distributed computing, adaptive quality management, artificial intelligence
Patent
Reusing programming environments in a distributed computing environment
Patent published on the 2026-05-07 in US under Ref US20260126968 by PALANTIR TECHNOLOGIES INC [US] (Correia JoÃo [gb], Lafleche Jonathan [gb], Raschkowski Willi [gb], Ayad John [gb])
Abstract: [0000] A system and related methods for creating a new programming environment from existing programming environments is disclosed. The system is programmed to retrieve relocatable packages of existing programming environments for a programming language from a remote repository and locally set up the existing programming environments as components of a new programming environment for the programming language. The system is programmed to further adjust an environment variable for the programming [...]
Our summary: The system creates new programming environments by reusing existing ones. It retrieves relocatable packages from a remote repository and sets them up locally. Symbolic links are used to manage library versions and facilitate updates.
programming environments, distributed computing, relocatable packages, symbolic links
Patent
Enhanced caching operations based on a coherency metric
Patent published on the 2026-04-23 in US under Ref US20260111394 by NETAPP INC [US] (Phatak Shirish [us], Kovvuri Vaishnav Kumar [us], Nair Sunil Sadashivan [in])
Abstract: A global policy-driven framework for managing, revising, and implementing data coherency policies in a distributed data storage system. An edge node of the distributed data storage system receives a request from an application to store a data object in the system. The edge node, in response to receiving the request to store the data object, generates a coherency policy request, which is then submitted to the primary node of the system. The primary node of the system generates a coherency policy [...]
Our summary: The content describes a framework for managing data coherency in distributed storage systems. An edge node generates a coherency policy request upon receiving a storage request. The primary node responds with a coherency policy, allowing the edge node to store the data object accordingly.
caching, coherency, distributed systems, data management
Patent
System, edge device, and control method thereof
Patent published on the 2026-04-23 in US under Ref US20260111770 by CANON KK [JP] (Kitagata Kei [jp])
Abstract: One or more systems, edge devices, control methods, and storage mediums are provided herein. In a case where query data is input from an operation unit of a client terminal or a first edge device, a determining unit of the first edge device determines which of a reasoning unit of the first edge device, a second device, and a first application service should perform a reasoning process on the query data. In a case where it is determined that the second device should perform the reasoning process,[...]
Our summary: The content describes a system involving edge devices and control methods for processing query data. A determining unit decides where the reasoning process should occur among available devices. If the second device is chosen, query data is transmitted directly to it, bypassing the first application service.
edge device, control method, reasoning process, query data
Patent
Adaptive Multi-Camera Fusion and Calibration for Large-Scale Multi-Vehicle Cooperative Simulation Scenarios
Published on 2026-02-03 by Hui Zhang, Chenyu Xia, Huantao Zeng @MDPI
Abstract: In the development of multi-vehicle cooperative hardware-in-the-loop (HIL) simulation platforms based on machine vision, accurate vehicle pose estimation is crucial for achieving efficient cooperative control. However, monocular vision systems inevitably suffer from limited fields of view and insufficient image resolution during target detection, making it difficult to meet the requirements of large-scale, multi-target real-time perception. To address these challenges, this paper proposes an eng[...]
Our summary: This paper presents a multi-camera cooperative vision detection method for large-scale vehicle simulations. It enhances vehicle pose estimation by projecting images onto a unified plane and optimizing image stitching. Experimental results show high accuracy and real-time performance, achieving a 93.41% identification rate and 30 frames per second output.
Multi-Camera Fusion, Vehicle Pose Estimation, Real-Time Processing, Image Mosaicking
Publication
A Lightweight Wheat and Corn Recognition Model Integrating Multiple Attention Mechanisms
Published on 2026-02-03 by Hao Yao, Ji Zhu, Yancang Li, Haiming Yan, Wenzhao Feng, Luwang Niu, Ziqi Wu @MDPI
Abstract: This study is grounded in the macro-context of smart agriculture and global food security. Due to population growth and climate change, precise and efficient monitoring of crop distribution and growth is vital for stable production and optimal resource use. Remote sensing combined with deep learning enables multi-scale agricultural monitoring from field identification to disease diagnosis. However, current models face three deployment bottlenecks: high complexity hinders operation on edge device[...]
Our summary: This study presents MAF-RecNet, a lightweight model for crop recognition that integrates multiple attention mechanisms to enhance efficiency and generalization with limited data. It addresses deployment challenges in smart agriculture by optimizing structure and improving feature discrimination. The model demonstrates high accuracy and strong cross-task generalization, contributing to the practical application of AI in precision agriculture.
Crop Recognition, Attention Mechanisms, Deep Learning, Smart Agriculture
Publication