Estimation of Arbitrary Resident Locations Using Data Obtained from an Infrared Sensor Array

The goal of the series of our studies is to build a resident-friendly smart home while protecting residents' privacy. According to our previous studies, an array of 15(5 x 3) binary infrared sensors can recognize some pre-defined locations of a resident in high accuracy. The purpose of this study is to investigate the possibility of estimating the coordinates of arbitrary locations, including those not defined during training. Estimation of the coordinate (x, y) is equivalent to regression of two functions x (D) and y(D), where D is a 15-dimensional vector observed by the sensor array. Experimental results illustrate that the estimation is relatively accurate for the y-axis. However, we could not obtain highly accurate results for the x-axis this time. We are going to examine the reasons and improve the performance in the next step.