Towards a Diffraction-based Sensing Approach on Human Activity Recognition

FUSANG ZHANG, Peking University; Institute of Software, Chinese Academy of Sciences, China
KAI NIU, Peking University, China
JIE XIONG, University of Massachusetts, Amherst, USA
BEIHONG JIN, Institute of Software, Chinese Academy of Sciences; University of Chinese Academy of Sciences, China
TAO GU, Royal Melbourne Institute of Technology University, Australia
YUHANG JIANG, Peking University, China
DAQING ZHANG, Peking University, China

Key words

1. how wireless signals are affected by human activities taking transceiver location and environment settings into consideration;
2. a new deterministic sensing approach to model the received signal variation patterns for different human activities;
3. a proof-of-concept prototype to demonstrate our approach and a case study to detect diverse activities.

In particular, we propose a diffraction-based sensing model to quantitatively determine the signal change with respect to a target’s motions, which eventually links signal variation patterns with motions, and hence can be used to recognize human activities.

本文旨在通过对信号的深入理解解决人体干扰导致 的信号pattern不一致不稳定的问题——在第一菲尼尔区内建立了菲涅尔区衍射模型。

contribution

• This work investigates a diffraction-based sensing model to explain the reason behind the performance instability which is a major issue for existing sensing system. This work sheds light on a new deterministic approach for wireless activity sensing.
• We analyze the signal propagations with respect to target location in the FFZ, and develop a mathematical model to establish the quantitative relationship between the signal variation and target location. We verify this model through benchmarking experiments and reveal several important properties for target sensing.
• Based on the proposed sensing model, we build a proof-of-concept system for detecting repetitive motion activities. We employ the unique signal variation induced by each activity to recognize the activity. Our experiments demonstrate that the system is effective and robust against ambient environmental changes. In addition, the proposed model improves the recognition accuracy of machine learning systems by above 10%.

Diffraction-based sensing model

1. 基于菲涅尔区理论

能量的70%集中在第一菲涅尔区（FFZ）
2. 衍射模型
3.