A trial of linear dynamic model extension to improve prediction accuracy of core temperature change for persons with spinal cord injury

Persons with spinal cord injury (SCI) experience thermoregulatory disturbances because of the damage to their nervous systems. Their bodies often accumulate heat while exercising under high temperatures, and their core temperature continues to rise unconsciously. To prevent heat accumulation of such patients. We proposed a linear dynamic model for people suffering from SCIs focusing heat inputs that are air temperature and exercise load, and we predicted the core temperature with the error of 0.2 °C. In this paper, we verified the fitting error by adding the parameters, and its prediction errors were evaluated changing the condition of the parameter identification using the model with added parameters and previous model. As a result, it was confirmed that the fitting error was improved by adding the model parameters, but longterm prediction accuracy was not improved. Using the linear dynamic model identified by the data sampled from several conditions is effective to predict the core temperatures of persons with SCI.