Movement Detection Using A Reciprocal Received Signal Strength Model

Received signal strength measurements of commodity radios can be utilized for sensing the surrounding environment. This work harnesses the signal strength measurements for estimating time periods when a person is stationary and moving. A novel reciprocal signal strength model is presented, and an energy detector is developed. It is shown that the decision threshold can be calculated in closed form for the proposed model. In addition, the observation time window can be minimized to one communication cycle which equals 58 milliseconds in our case. Using real-world experimental data from two different environments, it is demonstrated that movement can be correctly detected over 99% of the time.