Optimal Experimental Design for Human Thermoregulatory System Identification

R ollins et al. (2006) proposed a block-oriented modelling (BOM) approach for obtaining human thermoregulatory models for individual subjects. The objective of this approach is to obtain a library of model structures that map input changes such as humidity, ambient temperature, exercise, and so on, to dynamic thermoregulatory responses such the core temperature, skin temperature, muscle temperature, sweat rate and so on. These model structures will be categorized by human attributes and environmental conditions. To obtain this library of model structures a huge amount of experimentation in environmental chambers will be required to determine how the structures vary over the enormous input space. Thus, it is critical that experimental time is minimized. Using a D-optimality criterion, this article presents an experimental design approach that reduced the experimental time by 70% in comparison to the design in (Rollins et al., 2006).