ANC Workshop - Andrew Wood

Speaker: Andrew Wood

Title: Improvements to Hiking Speed Predictions using GPS Analysis

Abstract: Hikers and hillwalkers typically use the gradient in the direction of travel (walking slope) as the main variable in predicting walking speeds along a route. We used 4500 km of public GPS data from walking / hiking tracks (filtered to remove breaks and non-walking sections), to build a model for predicting walking speed where the gradient of the terrain (hill slope) is considered alongside the walking slope. The resulting generalised linear model (GLM) was found to more accurate than the most commonly used existing functions (Naismith's rule and Tobler's Hiking function), with a lower root-mean-square-error . Furthermore, we found an increase in RMSE between the GLM and existing functions as hill slope increases, indicating the importance of including this new variable in models.