**5. Discussion**

In the evaluation for this paper, 24 positions with an interval of 0.6m are sampled in a typical indoor area. When the range of sensing area is fixed, increasing the number of sampling locations will improve the perception effect. If the sensing area continues to expand, such as in a larger room, the perception effect will be decreased to some extent. When the sensing target is far away from the transmitter and receiver, the influence on signal transmission will be weakened. Theoretically speaking, the sensing performance decreases with the increase of the distance between the sensing target and the sensing device. For the number of activities, there may be more activities in practice scenarios. As for the experimental settings in most literature, five to eight activities are usually recognized in a typical smart home control scenario. If the number of activities continues to increase, the recognition accuracy will decrease to a certain extent, because some actions may have similar features and be easily confused. This is still a challenging issue for Wi-Fi-based human activity recognition, which will be further explored in future work.

## **6. Conclusions**

In this paper, Wi-Fi-based multi-location human activity recognition technique is explored. A novel AP-DCN-based method that fully leverages the amplitude and phase information is presented. The complex convolution layer, complex batch normalization layer, and complex ReLU activation function are leveraged for feature representation. Furthermore, considering the unbalanced sample number at different locations, a perception method based on DCN-TL is proposed. To verify the performance of the method, a dataset involving five activities at 24 positions in an office is built. The experiment results indicate that the AP-DCN-based method can achieve an average accuracy of 96.85% for five people with only five training samples at each of the 24 locations. Furthermore, the proposed method is also applicable to the training samples with a low sampling rate and fewer subcarriers. In the case of unbalanced number of data samples at different locations, the recognition accuracy is 94.02%. Therefore, it is concluded that the presented method is feasible for multi-location human activity recognition with limited data samples, which promisingly promotes the generalization performance of the device-free sensing system.

**Author Contributions:** Conceptualization, X.D., C.H., W.X. and T.J.; methodology, X.D.; software, X.D.; validation, X.D.; formal analysis, X.D., C.H., W.X., Y.Z. and J.Y.; investigation, X.D. and Y.Z.; resources, W.X., C.H. and T.J.; data curation, X.D.; writing—original draft preparation, X.D.; writing— review and editing, X.D., C.H., W.X., Y.Z., J.Y. and T.J.; visualization, X.D.; supervision, T.J., C.H. and W.X.; project administration, C.H., W.X. and T.J.; funding acquisition, T.J., Y.Z. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work is supported by the National Natural Sciences Foundation of China (No. 62071061), and Beijing Institute of Technology Research Fund Program for Young Scholars.

**Institutional Review Board Statement:** Not applicable

**Informed Consent Statement:** Not applicable

**Data Availability Statement:** Not applicable

**Conflicts of Interest:** The authors declare no conflict of interest.
