A novel approach for slip resistance evaluation of winter footwear based on probability of slipping and cost analysis

Abstract

This paper proposes a slip-prediction model to estimate the probability of footwear slipping on ice. The mode is based on slips that are induced when human subjects are challenged by walking in a test chamber with full control over environmental conditions. This technique will help to evaluate the slip resistance quality of winter footwear in an ecologically valid manner. Four styles of winter footwear were tested by 8 participants on wet and dry ice surfaces with different slopes. The tests were repeated three times on different days. The probability of slipping was estimated for each incline angle using logistic regression analysis with three predictors: angle, ice condition and walking direction. In order to rank the quality of the winter footwear, a comprehensive cost analysis was performed to estimate the misclassification cost from the logistic regression. This method makes no assumption regarding the distribution of the data. The probability cutoff for each footwear style is obtained by minimizing the misclassification cost. The results show that increasing the incline angle significantly increases odds of slipping for all four styles (p<0.01). In addition, as indicated by the results, different styles of footwear perform differently on the various winter surfaces. Therefore, this method is able to rank the slip-resistance property of the footwear based on the probability of slipping.