A method for classifying snow using ski-mounted strain sensors

Abstract

Qualitative characteristics of snow can vary substantially from day to day as well as across aspects and elevations. This makes it challenging for skiers to choose the right ski, and for ski manufacturers to design and evaluate their ski's performance across diverse real-world conditions. Here, we explore a proof-of-concept approach for automatically assessing qualitative aspects of snow type while skiing using strain sensors mounted to the top surface of an alpine ski. We show that with two strain gauges and an inertial measurement unit it is feasible to correctly assign one of three qualitative labels (powder, slushy, or icy/groomed snow) to each 10 s segment of a trajectory. Our dataset included two general skiing styles—wide turns and tight turns—and for our data the algorithm we developed was able to correctly assign the qualitative labels with 97% accuracy. Our algorithm uses a combination of a data-driven linear model of the ski-snow interaction, dimensionality reduction, and a Naïve Bayes classifier. Comparisons of classifier performance between strain gauges suggest that the optimal placement of strain gauges is halfway between the binding and the tip/tail of the ski, in the cambered section just before the point where the unweighted ski would touch the snow surface. A key component of our approach is to use the relative dynamics between the inertial measurement unit under the boot and the strain gauges placed farther away, making it possible the distinguish dynamics that are a result of the snow rather than the skiers input. The ability to classify snow, potentially in real-time, using skis opens the door to applications including improved quantitative characterization of ski performance across a range of conditions, snow specific recommendations for ski selection and technique coaching, and future development of skis with automated stiffness tuning based on the snow type.