Spatial-Temporal Federated Transfer Learning with multi-sensor data fusion for cooperative positioning

Xiaokang Zhou, Qiuyue Yang, Qiang Liu, Wei Liang, Kevin I-Kai Wang, Zhi Liu, Jianhua Ma, Qun Jin

Abstract: Highlights•A three-layer ST-FTL architecture for more efficient global model aggregation and lightweight local training.•A multi-attribute based spatial–temporal clustering algorithm for suitable source domain and model weight selection.•A convolutional-gated unit for feature refinement in more effective model initialization and weight aggregation.•A lightweight fusion model with a Siamese network structure for local data augmentation, feature selection and fusion.