An advanced Neuro-space mapping (Neuro-SM) multiphysics parametric modeling approach for microwave passive components is proposed in this paper. The electromagnetic (EM) domain model, which represents the EM responses with respect to geometrical parameters, is regarded as a coarse model. The multiphysics domain model, which represents the multiphysics responses with respect to both geometrical parameters and multiphysics parameters, is regarded as a fine model. The proposed model is constructed by the input mapping, the output mapping and the coarse model. The input mapping is used to map multiphysics parameters to EM parameters. The output mapping is introduced to further narrow the gap between the output of the coarse model and the multiphysics data. In addition, a three-stage training method is proposed for efficiently developing the proposed multiphysics model. The proposed technique, which combines the efficiency of EM analysis and the accuracy of multiphysics analysis, can achieve better accuracy with less multiphysics data than existing modeling methods. The developed Neuro-SM multiphysics model provides accurate and fast predictions of multiphysics responses. Therefore, the design cycle of microwave passive components is shortened while the modeling cost is significantly reduced. Two microwave filter examples are utilized to demonstrate the accuracy of the proposed parametric modeling technique.

Radio frequency identification (RFID) is a promising sensing solution for recognizing human activity. Existing RFID-based human activity recognition (RHAR) sensing systems either rely on actively worn tags or have poor recognition accuracy and computational complexity. In this paper, we present a novel device-free RHAR sensing method called less-complex and accurate recognition method (LARM). A novel dictionary-based learning is proposed to distinguish different activities from RFID sensory data. Rather than traditionally using several dictionaries, the proposed approach merely learns a background dictionary for the scenario with no activities performed. Then, using the background dictionary as a guide, a reconstructed offset matrix (RCOM) is created as activity fingerprints. It reduces computational complexity while remaining highly accurate. Meanwhile, a feature selection approach based on stability is developed for RCOM optimization, which improves the discriminability of sensor-derived fingerprints even more. Finally, a k-nearest neighbor-based recognition method is suggested to label the testing sample. The similarities between a testing sample and the fingerprints are precisely appraised via a feature selection matrix, pinpointing the optimal candidate fingerprints for activity classification. Experiment comparisons with the existing algorithms demonstrate the superiority of the proposed LARM.