In order to study the static and dynamic comfort of tight sportswear in winter, the subjective comfort was aimed to be evaluated by collecting sensory data such as humidity feeling, cold feeling and other perceptions. In this paper, the experiment was divided into standing, squatting, jumping, jogging, walking and so on.
Through particle swarm optimization-cuckoo search model, the sensory factors that affect the overall comfort were optimized, and it was found that there were great differences in the overall comfort factors under different motions. Then, analytic hierarchy process was used to sort the optimized sensory indicators in each experimental stage, and the influence degree of sensory indicators was studied. Finally, by the long short-term memory (LSTM) model, taking comfort senses of standing, squatting, jumping and jogging as input parameters, and regarding comfort senses of walking, lifting legs and resting as output parameters, the prediction model was founded.
The results showed that there were certain differences between the prediction value and the real subjective evaluation value, but most of the predicted values were consistent with the real values on the sensory level, and the overall prediction level was good, which meant that the LSTM model had more accurate prediction ability for subjective evaluation and could be extended to other sports.
The research results could provide scientific methods for the design of tight-fitting sportswear in winter.
