Abstract

In this research, we propose a model to calculate muscular fatigue at joint level, more precisely at muscle groups level. The model is used for the computerized generation of realistic human postures. Our approach splits each single degree of freedom joint into two coordinated half-joints, thus the name of half-joint pair. In Anatomy, both groups constituting the half-joint pair are said to be antagonist. Fatigue model parameters are joint strength and the current joint torque, which are used to calculate a value of normalized torque. The normalized torque is used to compute the maximum holding time that the posture can be sustained (in an evolving static context). The model integrates time as an explicit variable in an Inverse Kinematics framework in such a way that fatigue evolution over time can be exploited for posture optimization and reachable volume characterization.

In the postural optimization we introduce a hysteresis activation pattern for each half-joint to set a fatigue reduction constraint whenever necessary. They can be named “hard constraints” as they have to be ensured with a higher priority than all the other Inverse Kinematics tasks. The hysteresis activation pattern analyses half-joint fatigue level and when it is above the fatigue threshold, the joint variation is constrained to reduce the half-joint torque by a small increment compatible with the corresponding time increment. Our approach achieves fatigue minimization, exploiting active and passive torques at joint level. We have used a factor, named “muscular tonus”, which represents the proportion of active torque that is being used in the fatigue reduction process. At a higher level, we use the model to identify postures or reachable spaces using the fatigue factor.

Contributions of the model are the reduced number of parameters used and the consideration of the eass of time in the own model. Our approach has a low computational cost, being suited for PC platform. It has been validated in an Inverse Kinematics framework for realistic posture generation. During the simulation, joints fatigue values are updated so that the system reacts when unbearable fatigue values are reached. The fatigued posture is then adjusted searching for a less fatigued one. We have proven experimentally that strategies followed by subjects were similar to those achieved by the simulation framework.