Cooperative planning is one of the critical problems in the field of multi-agent system gaming.This work focuses on cooperative planning when each agent has only a local observation range and local communication.We pr...Cooperative planning is one of the critical problems in the field of multi-agent system gaming.This work focuses on cooperative planning when each agent has only a local observation range and local communication.We propose a novel cooperative planning architecture that combines a graph neural network with a task-oriented knowledge fusion sampling method.Two main contributions of this paper are based on the comparisons with previous work:(1)we realize feasible and dynamic adjacent information fusion using GraphSAGE(i.e.,Graph SAmple and aggreGatE),which is the first time this method has been used to deal with the cooperative planning problem,and(2)a task-oriented sampling method is proposed to aggregate the available knowledge from a particular orientation,to obtain an effective and stable training process in our model.Experimental results demonstrate the good performance of our proposed method.展开更多
Defects may occur in photovoltaic(PV)modules during production and long-term use,thereby threatening the safe operation of PV power stations.Transient thermography is a promising defect detection technology;however,it...Defects may occur in photovoltaic(PV)modules during production and long-term use,thereby threatening the safe operation of PV power stations.Transient thermography is a promising defect detection technology;however,its detection is limited by transverse thermal diffusion.This phenomenon is particularly noteworthy in the panel glasses of PV modules.A dynamic thermography testing method via transient thermography and Wiener filtering deconvolution optimization is proposed.Based on the time-varying characteristics of the point spread function,the selection rules of the first-order difference image for deconvolution are given.Samples with a broken grid and artificial cracks were tested to validate the performance of the optimization method.Compared with the feature images generated by traditional methods,the proposed method significantly improved the visual quality.Quantitative defect size detection can be realized by combining the deconvolution optimization method with adaptive threshold segmentation.For the same batch of PV products,the detection error could be controlled to within 10%.展开更多
基金Project supported by the National Natural Science Foundation。
文摘Cooperative planning is one of the critical problems in the field of multi-agent system gaming.This work focuses on cooperative planning when each agent has only a local observation range and local communication.We propose a novel cooperative planning architecture that combines a graph neural network with a task-oriented knowledge fusion sampling method.Two main contributions of this paper are based on the comparisons with previous work:(1)we realize feasible and dynamic adjacent information fusion using GraphSAGE(i.e.,Graph SAmple and aggreGatE),which is the first time this method has been used to deal with the cooperative planning problem,and(2)a task-oriented sampling method is proposed to aggregate the available knowledge from a particular orientation,to obtain an effective and stable training process in our model.Experimental results demonstrate the good performance of our proposed method.
基金Supported in part by the National Natural Science Foundation of China under Grant 51977117.
文摘Defects may occur in photovoltaic(PV)modules during production and long-term use,thereby threatening the safe operation of PV power stations.Transient thermography is a promising defect detection technology;however,its detection is limited by transverse thermal diffusion.This phenomenon is particularly noteworthy in the panel glasses of PV modules.A dynamic thermography testing method via transient thermography and Wiener filtering deconvolution optimization is proposed.Based on the time-varying characteristics of the point spread function,the selection rules of the first-order difference image for deconvolution are given.Samples with a broken grid and artificial cracks were tested to validate the performance of the optimization method.Compared with the feature images generated by traditional methods,the proposed method significantly improved the visual quality.Quantitative defect size detection can be realized by combining the deconvolution optimization method with adaptive threshold segmentation.For the same batch of PV products,the detection error could be controlled to within 10%.
