The strain on the body resulting from the use of an exoskeleton can be assessed via musculoskeletal simulations of joint reaction forces (JRFs). JRFs are the forces transferred from one bone to another within a joint. Parameters like the human musculoskeletal model as well as the simulation software and tools can influence the simulated JRFs. Until now, many studies used simple torque actuators to represent the exoskeleton’s support in their simulations. The results were then often utilized to optimize the exoskeletons’ physical design and controls, which is why a correct computation of the JRFs is important. This thesis investigates the effect that the representation of an exoskeleton has on the JRFs. This is done by implementing and simulating them for different virtual models of the same soft, cable-driven elbow exosuit prototype developed by Harbauer et al. (2020). The exosuit is implemented in 12 different models. Each of them goes through a simulation workflow including the computed muscle control (CMC) and joint reaction analysis (JRA) tools of OpenSim to acquire the JRFs. The process is automated using MATLAB. A pure elbow flexion motion of lifting a 5 kg weight is supported by the different models with equivalent support. Three control strategies are implemented for this. The exoskeleton representation is found to affect the simulated JRFs in different ways. Several relevant parameters in the modeling approach are identified and analyzed. Using a coordinate actuator tends to underestimate the forces compared to path actuators. The bone(s) that a path actuator is attached to can play a role for the JRFs. The level of detail in the representation of the cable path along the forearm also has an influence, but is less relevant when the path points are attached to one bone only. For the selected model, motion and exosuit, the exosuit’s cable path going over the user’s shoulder toward the motors does not have a relevant effect on the elbow JRFs. The simulation results indicate that using an ulna-attached single- or double-stringed path actuator to represent the LfE exosuit prototype is a reasonable choice for JRF computation with the presented workflow. Finally, the simulated elbow JRF values are compared to the results of other publications to evaluate the quality of the simulation workflow, providing confidence in the results. It is furthermore confirmed that the underlying human model, as well as the exoskeleton’s control strategy have a large influence on the produced JRFs.     «

The strain on the body resulting from the use of an exoskeleton can be assessed via musculoskeletal simulations of joint reaction forces (JRFs). JRFs are the forces transferred from one bone to another within a joint. Parameters like the human musculoskeletal model as well as the simulation software and tools can influence the simulated JRFs. Until now, many studies used simple torque actuators to represent the exoskeleton’s support in their simulations. The results were then often utilized...     »