A Novel Device-Free Localization Approach Based on Deep Dictionary Learning

As an emerging technology, device-free localization (DFL) has a wide range of application scenarios in the field of the internet of things. However, most of the existing DFL methods take the mode of learning features from raw data, and then perform to achieve localization using classification, which has inferior localization performance. To improve the localization accuracy, this study proposes an accurate and effective localization technique based on deep dictionary learning with sparse representation (DDL-DFL). The method extracts the in-depth features of the data through multi-layer dictionary learning and stacks the features of each layer for classification. Furthermore, we propose a data augmentation method, which can be applied to scenarios with fewer sensor nodes to increase the data dimension and strengthen the essential features to improve the accuracy of localization. We evaluate the performance of the DDL-DFL algorithm on collected laboratory datasets, and the results are superior to existing localization algorithms. In addition, the DDL-DFL algorithm with data augmentation is conducted on the laboratory datasets with a low dimension of data, and the localization performance has been significantly improved.