G V L A B
Research
Human motion analysis
***human motion identification
Human motion analysis such as estimation of joint angles and dynamic parameters is of crucial importance for a field of the evaluation of rehabilitation or the humanoid robots' interacting or imitating.
The joint angles are normally estimated by the inverse kinematics which uses the marker positions obtained from motion capture. However, the kinematic data (marker positions) are sensitive to marker misplacements or soft tissue artifacts. Because of this, the inverse kinematics result might not be consistent with dynamic laws governing the system whole body motion. In this work, the a new method to perform a dynamically consistent inverse kinematics calculation is proposed by using Kinematic and dynamic data together. This method would be helpful to estimate joint torques or analyze other dynamic state since it is possible to estimate the dynamics parameters such as mass, centre of mass, and inertia and the kinematic parameters at the same time.

HRP2


copyright Gentiane Venture 2009