Machine learning approach to determine the clothing level using an IR sensor in the vehicle cabin

Authors

A. Kirmas - ika, RWTH Aachen University

Summary

In the current model-based climate control, the individual thermal state of each occupant has no influence on the output variables. For this reason, research is conducted towards a model that allows prediction of the thermal state of the vehicle occupants. This thermal state can include a number of variables, such as clothing level, sweat rate, metabolism and the resulting thermal comfort.

In this paper, the sub-aspect of the clothing level is presented in more detail. For this purpose, the interior of the test vehicle was equipped with IR sensors and cameras, from which a thermal profile of the clothing and the skin can be determined. In addition, data from the climate control ECU is recorded to supplement the thermal images with vehicle data to describe the thermal environment within the vehicle cabin. This data will be recorded for different clothing levels, thermal boundary conditions and driving situations and transferred to a machine learning algorithm. This then allows an estimation of the clothing level and a more precise adaptation of the climate control to the user's needs.