Multivariate Regression Analysis of Skin Temperature Rises for Millimeter-Wave Dosimetry

This study presents a fast and simple estimation method for dosimetry assessment of a clothed human body exposed to electromagnetic fields in millimeter-wave bands. A multivariate regression algorithm was given instead of conventional numerical methods for the first time, which utilized the exact value of skin surface temperature rise that was computed by the Monte-Carlo method by changing the oblique incidence wave angle, the cross-polarization power ratio, and the air gap spacing from the cloth to the body. The detailed formula for bioheat transfer analysis considering the abovementioned key parameters is first introduced. Then, a polynomial function is derived by multivariate regression analysis using the simulation result to estimate the skin temperature rises under various exposure conditions and scenarios. In comparison with those results obtained by the traditional methods, the proposed regression algorithm significantly reduced the expense of computational resources. The results estimated using the developed function show a maximum error of less than 0.013 °C for the skin tissues of different body parts and different cloth materials, demonstrating that the proposed method is easy and efficient for the estimation of a high degree of skin temperature rise in a relatively complicated radio-wave exposure environment in millimeter-wave bands.