A Mobile-Agent-Based Adaptive Data Fusion Algorithm for Multiple Signal Ensembles in Wireless Sensor Networks

Distributed Compressed Sensing (DCS) is an emerging field that exploits both intra- and inter-signal correlation structures and enables new distributed coding algorithms for multiple signal ensembles in wireless sensor networks. The DCS theory rests on the joint sparsity of a multi-signal ensemble. In this paper we propose a new mobile-agent-based Adaptive Data Fusion (ADF) algorithm to determine the minimum number of measurements each node required for perfectly joint reconstruction of multiple signal ensembles. We theoretically show that ADF provides the optimal strategy with as minimum total number of measurements as possible and hence reduces communication cost and network load. Simulation results indicate that ADF enjoys better performance than DCS and mobile-agent-based full data fusion algorithm including reconstruction performance and network energy efficiency.

Automatic Voltage Control of Differential Power Grids Based on Transfer Learning and Deep Reinforcement Learning

In terms of model-free voltage control methods,when the device or topology of the system changes,the model’s accuracy often decreases,so an adaptive model is needed to coordinate the changes of input.To overcome the defects of a model-free control method,this paper proposes an automatic voltage control(AVC)method for differential power grids based on transfer learning and deep reinforcement learning.First,when constructing the Markov game of AVC,both the magnitude and number of voltage deviations are taken into account in the reward.Then,an AVC method based on constrained multiagent deep reinforcement learning(DRL)is developed.To further improve learning efficiency,domain knowledge is used to reduce action space.Next,distribution adaptation transfer learning is introduced for the AVC transfer circumstance of systems with the same structure but distinct topological relations/parameters,which can perform well without any further training even if the structure changes.Moreover,for the AVC transfer circumstance of various power grids,parameter-based transfer learning is created,which enhances the target system’s training speed and effect.Finally,the method’s efficacy is tested using two IEEE systems and two real-world power grids.

Edge Intelligence for Smart Grid:A Survey on Application Potentials

With the booming of artificial intelligence(AI),Internet of Things(IoT),and high-speed communication technology,integrating these technologies to innovate the smart grid(SG)further is future development direction of the power grid.Driven by this trend,billions of devices in the SG are connected to the Internet and generate a large amount of data at network edge.To reduce pressure of cloud computing and overcome defects of centralized learning,emergence of edge computing(EC)makes the computing task transfer from the network center to the network edge.When further exploring the relationship between EC and AI,edge intelligence(EI)has become one of the research hotspots.Advantages of EI in flexibly utilizing EC resources and improving AI model learning efficiency make its application in SG a good prospect.However,since only a few existing studies have applied EI to SG,this paper focuses on the application potential of EI in SG.First,the concepts,characteristics,frameworks,and key technologies of EI are investigated.Then,a comprehensive review of AI and EC applications in SG is presented.Furthermore,application potentials for EI in SG are explored,and four application scenarios of EI for SG are proposed.Finally,challenges and future directions for EI in SG are discussed.This application survey of EI on SG is carried out before EI enters the largescale commercial stage to provide references and guidelines for developing future EI frameworks in the SG paradigm.