Optimal design of musculoskeletal models using force field data

We propose an optimal design framework for a musculoskeletal model. A six degrees of freedom generalized force field at an end effector is measured by activating a specific muscle in-situ. Given that we have a musculoskeletal simulation model with a set of undetermined design parameters, the framework chooses the optimal parameters which minimize the difference between simulated and measured force field data. To ensure generality, the framework is model independent, which means that any type of musculoskeletal model can be applied. As a generalized method, a PID (Proportional-Integral-Derivative) controller and an empirical Jacobian are used to correctly simulate the force field without suffering from numerical issues such as stability and indeterminacy. Case studies of the muscles of the rat hind limb show a satisfactory capability of the framework.