To further improve the accuracy of underground positioning, a method of underground Ultra-Wide Band positioning based on clustering and TDOA is proposed. First, the TDOA method is used to measure, Chan algorithm is se...
详细信息
This article investigates the application of meta-learning in training reinforcement learning agents for fast adaptation in rapidly changing dynamic environments. The proposed meta-learning approach demonstrates promi...
详细信息
ISBN:
(数字)9798350370249
ISBN:
(纸本)9798350370270
This article investigates the application of meta-learning in training reinforcement learning agents for fast adaptation in rapidly changing dynamic environments. The proposed meta-learning approach demonstrates promising results in multi-agent environments, outperforming baseline algorithms and achieving stable convergence. However, in single-agent settings, the meta-algorithms exhibit similar or inferior performance compared to baseline algorithms, potentially due to overfitting. Nevertheless, the meta-learning algorithms perform well in non-dynamic single-agent environments, showcasing their ability to adapt effectively. The findings highlight the potential of incorporating meta-learning approaches in multi-agent systems to enhance adaptive behaviours and optimize performance in dynamic environments.
Federated learning (FL) has considerably emerged as a promising solution to enhance user privacy and data security by enabling collaboratively multi-party model learning without exchanging confidential data. Neverthel...
详细信息
This paper introduces an architecture that combines IoT-driven data collection with Retrieval-Augmented Generation (RAG) for real-time environmental monitoring, analysis, and prediction. Using edge computing for local...
详细信息
ISBN:
(数字)9798350364637
ISBN:
(纸本)9798350364644
This paper introduces an architecture that combines IoT-driven data collection with Retrieval-Augmented Generation (RAG) for real-time environmental monitoring, analysis, and prediction. Using edge computing for local processing and centralized LLMs for complex analysis, the system delivers scalable and timely insights into environmental issues like climate change and disaster prevention. Data from IoT sensors is preprocessed at edge devices and sent to a central server, where embeddings are generated, stored in a vector database, and analyzed by the LLM for fast query handling. This system provides real-time feedback through dashboards, focusing on overcoming challenges such as network delays, scalability, and security to deliver reliable environmental monitoring for non-expert users.
Mobile Ad hoc Networks (MANETs) are the integral part of advanced IoT applications, offering flexibility as well as scalability in dynamic network conditions. However, their distributed nature and reliance on wireless...
详细信息
Multi-Sequence Alignment (MSA) is considered an NP problem in bioinformatics. Compared to traditional techniques, nature-inspired techniques produce accurate results. In this article, Improved Chemical Reaction Optimi...
详细信息
The precise control of an exoskeleton, especially for pediatric subjects, is a challenging task due to the presence of parametric perturbations and external disturbances (PPED). The sliding mode control is a well-know...
详细信息
Modern deep Reinforcement Learning (RL) techniques are quite good at choosing the best possible rules to maximise rewards. By using rich visual information for policy selection, this method combined with Deep Learning...
详细信息
ISBN:
(数字)9798350394474
ISBN:
(纸本)9798350394481
Modern deep Reinforcement Learning (RL) techniques are quite good at choosing the best possible rules to maximise rewards. By using rich visual information for policy selection, this method combined with Deep Learning methodologies shows promise for difficult tasks. In this work, we have used Double DQN and also three variants of Duelling DQN to implement the ice hockey environment. The Duelling DQN outperformed the Double DQN and gave better results.
In our swiftly digitizing world, the importance of low-earth-orbit satellite (LEO SAT) communication is escalating. Advancements in 5G, IoT, and AI underscore the need for a robust communication network. Research reve...
In our swiftly digitizing world, the importance of low-earth-orbit satellite (LEO SAT) communication is escalating. Advancements in 5G, IoT, and AI underscore the need for a robust communication network. Research reveals that nearly 40% of the global population faces Internet service challenges due to limited high-performance access by 2022. This technology not only addresses coverage limitations but also mitigates latency issues. Multinational corporations like SpaceX and Amazon are already exploring its potential. As we approach the 6G era, users’ computational demands surge, prompting satellite edge computing with onboard processing. Our paper proposes a multi-hop code offloading model using LEO SAT, reducing energy consumption, and introduces a dynamic optimization algorithm for LEO SAT - mobile edge computing. The research aims to enhance satellite-based communication networks, with a focus on optimizing satellite edge computing efficiency.
暂无评论