Strange effects of activation dynamics on musculoskeletal trajectory optimization

Abstract

Muscle activation dynamics is usually described with a nonlinear differential equation, such that the activation occurs faster than deactivation. When a muscle excitation input switches rapidly between two values, this produces a "pumping" effect that raises the activation above the mean excitation input. This effect will favor rapid switching strategies when excitation is used in the optimization objective for human movement. The optimal strategy has infinitely fast switching, and is difficult to obtain with direct collocation methods. This problem can be eliminated by using activation, rather than excitation, in the optimization objective.