Fast Inverse Dynamics Algorithm with Decomposed Computation of Gradient forWire-Driven Multi-Body Systems and its Application to Estimation of Human Muscle Tensions
This paper proposes a fast computationmethod for computing the inverse dynamics of a multibodysystem driven by wires, and applies the method to estimatemuscle tensions of a complex human musculoskeletalmodel. It basically utilizes the existing inverse dynamicsmethod to solve the quadratic programming consideringboth mechanical and physiological consistency. Thequadratic programming usually needs to compute the gradientof the evaluation function of the optimization. Inthe proposed method, by making use of the structure ofa multi-body system and its muscle-tendon wires, the gradientcomputation can be decomposed into kinematics anddynamics computation of rigid bodies and wires respectively,and the computation complexity of the optimizationcan be reduced.
inverse dynamics wire-driven multi-body system musculoskeletal model quadratic programming
K.Ayusawa Y.Nakamura
University of Tokyo,Japan
国际会议
上海
英文
1-7
2011-10-14（万方平台首次上网日期，不代表论文的发表时间）