Abstract
The purpose of this paper is to provide researchers with a description of a data library representing human interaction with medium- to large-sized objects in everyday life. The library includes motion capture data characterizing human and object motion, as well as data for characterizing haptic interaction with the object via force and torque measurements via a load cell and inertial measurement unit (IMU) readings of the object accelerations. Objects include a box, luggage, briefcase, walker, shopping cart, wheelbarrow, and door. The data collected includes multiple types of interactions with each object, such as manipulating the object and walking while interacting with the object (e.g., pulling, pushing, carrying, operating, etc.). Data processing techniques for synchronizing data, deriving human biomechanics, and segmenting trials are presented. Examples of how the data in the library can be manipulated and processed further are provided. This includes combining ten wheelbarrow lifts of one subject together and analyzing the knee motion, object acceleration, and load cell readings (force and torque) with mean trajectories and standard deviations of the trajectories. From there, the range of motion can be extracted, such as for the hip, knee, and ankle joint minimum angles, maximum angles, and range of motion. A comparison of walking with and without a wheelbarrow is presented using spatiotemporal parameters and cyclograms to demonstrate their differences. The database is available on AddBiomechanics, SimTK, and GitHub.