A Predictive Statistical Model for Shoe-Floor-Fluid Coefficient of Friction

Project Summary:Slip and fall accidents are a major and growing source of occupational injuries. Slip-resistant shoes with a highcoefficient of friction (COF) are effective at a reducing slipping risk. However, neither experts nor industry hasagreed upon a consistent set of criteria for labeling a shoe as slip-resistant. A consequence of this lack ofstandardization is that significant variability exists across shoes that are labeled slip-resistant. Furthermore,independent testing of shoe COF is expensive, which may limit its use by employers and employees. Theproposed research aims to address this problem by developing a predictive model for shoe-floor-contaminantCOF based on shoe parameters that can be measured with little cost. The overall objective of this R03 studyis to train and validate a statistical model for predicting the COF of footwear against a floor surface in thepresence of a liquid contaminant. The feasibility of this approach is supported by preliminary experimental andmodeling work conducted by the principal investigator. The proposed research will accomplish this goal withtwo aims: Aim 1: Develop and mechanically validate a statistical model for predicting the shoe COF based onthe characteristics of the shoe outsole; Aim 2: Validate the findings of Aim 1 using unexpected slips of humansubjects. To accomplish Aim 1, a shoe tribometer that mimics the under-shoe conditions during slipping willmeasure coefficient of friction across fifty shoes, three floor surfaces (quarry tile, vinyl tile and ceramic tile), andthree contaminant conditions (water, detergent solution and canola oil). ANOVA methods and a ten-fold cross-validation method will be used to identify the most predictive model and quantify its accuracy. The predictionvariables will include outsole hardness, contact area, tread orientation, floor roughness and fluid viscosity.Regression methods will be used to test the hypothesis that the model predicts shoe-floor COF (H1.1). Also,ANOVA methods will be used to test the hypotheses that shoe features (hardness, tread contact area andtread orientation) influence hysteresis COF (H1.2) and adhesion COF (H1.3). For Aim 2, thirty individuals willbe unexpectedly slipped to determine if the developed model can predict actual slipping risk. The modeldeveloped in Aim 1 will be used to predict slips based on measured outsole design features. Subjects will berandomly assigned across these outsole types during two unexpected slips. A logistic regression analysis willtest the hypothesis that the model predicted slips across the subject population based on shoe treadmeasurements (H2). This proposed research promotes Research to Practice (R2P) concepts by makingassessment of shoe slip-resistance more accessible to employees and employers. The outputs of thisresearch will be newly developed mathematical relationships between footwear features and slip-resistance.The outcomes will be improved shoe design and selection policies that lead to a reduction in slip and fallaccidents. This research will address NORA Strategic Goals for Manufacturing (Goal 2), Wholesale andRetail Trade (Goal 2) and Services (in 9 different research goals).