AUTHOR=Li Wenjing , Zhang Nan , Liu Zhu , Ma Shiqian , Ke Huaqiang , Wang Jinfa , Chen Ting TITLE=A trusted decision fusion approach for the power internet of things with federated learning JOURNAL=Frontiers in Energy Research VOLUME=Volume 11 - 2023 YEAR=2023 URL=https://www.frontiersin.org/journals/energy-research/articles/10.3389/fenrg.2023.1061779 DOI=10.3389/fenrg.2023.1061779 ISSN=2296-598X ABSTRACT=Data is a treasure in the information society. With artificial intelligence approaches, data can be used to make precise decisions. Electricity power represents the political, economic, and development status of a country. There is a large amount of private data in the particular domain of power Internet of Things, and the owners of electricity data with boundaries do not want their data to be leaked. By intercepting electricity power data, many aspects of a country can be analyzed. The value brought by electricity power data cannot be ignored. There are currently various approaches such as research on secure communication protocols to allow data to be secured in transmission or proposing new security solutions in terms of authorization methods. Most of these approaches are secured from communication and they are suitable for data security sharing of traditional approaches. But for artificial intelligence approaches that feed big data into deep learning for precise decision-making within power Internet of Things, it is essential to secure the data while building the model. In this paper, a federated learning approach based on homomorphic encryption is introduced to power Internet of Things for decision systems. Federated learning is introduced to provide a way of data security sharing. Still, data augmentation and transfer learning are used to overcome too little local training data. The paper also attempts to incorporate the specialized nature of traditional manual decision-making in the power field by fusing expert and model values after stratifying the requirements. Experiments are conducted to simulate the decision requirements in the field of power Internet of Things (e.g., electrical material identification), using image recognition as an example, and the experimental results show that the models in this paper can achieve high accuracy rates and that the fusion approach is feasible.