Dynamically Consistent Human Movement Prediction for Interactive Vehicle Occupant Package Design
Proc. of Digital Human Modeling Symposium, Tokyo, Japan
A differential kinematics-based algorithm is realized for predicting the posture and movement of a vehicle occupant performing operational tasks. The algorithm is developed in OpenSim, a widely used open-source software system that lets users develop models of musculoskeletal structures and create kinematic or dynamic simulations of movement. The proposed algorithm obtains a dynamically consistent inverse kinematics solution that accounts for the dynamic and anatomical characteristics of the system such as inertial properties and joint limits. The predicted movement trajectory is continuous and achieves a highly accurate prediction of the motion undertaken by a human to perform a task. The method offers a 30-fold increase in computation speed over OpenSim's optimization-based inverse kinematics algorithm, enabling interactive, adaptive prediction of movement during operation of tasks in a wide variety of vehicle design configurations. The formulation is being extended to achieve additional objectives, such as enforcing collision and obstacle avoidance constraints and maximizing kinematic and dynamic dexterity. This advancement in motion prediction technology enables the development of software systems for interactive, biomechanics-based automotive design.