Abstract
Walking is a fundamental aspect of human movement, and understanding how irregular surfaces impact gait is crucial. Existing gait research often relies on laboratory settings with ideal surfaces, limiting the applicability of findings to real-world scenarios. While some irregular surface datasets exist, they are often small or lack biomechanical gait data. In this paper, we introduce a new irregular surface dataset with 134 participants walking on surfaces of varying irregularity, equipped with inertial measurement unit (IMU) sensors on the trunk and lower right limb (foot, shank, and thigh). Collected during the North American Congress on Biomechanics conference in 2022, the dataset aims to provide a valuable resource for studying biomechanical adaptations to irregular surfaces. We provide the detailed experimental protocol, as well as a technical validation in which we developed a machine learning model to predict the walking surface. The resulting model achieved an accuracy score of 95.8%, demonstrating the discriminating biomechanical characteristics of the dataset's irregular surface gait data.