Assessing effects of exoskeleton misalignment on knee joint load during swing using an instrumented leg simulator

Abstract

Background: Exoskeletons are working in parallel to the human body and can support human movement by exerting forces through cuffs or straps. They are prone to misalignments caused by simplified joint mechanics and incorrect fit or positioning. Those misalignments are a common safety concern as they can cause undesired interaction forces. However, the exact mechanisms and effects of misalignments on the joint load are not yet known. The aim of this study was therefore to investigate the influence of different directions and magnitudes of exoskeleton misalignment on the internal knee joint forces and torques of an artificial leg.

Methods: An instrumented leg simulator was used to quantify the changes in knee joint load during the swing phase caused by misalignments of a passive knee brace being manually flexed. This was achieved by an experimenter pulling on a rope attached to the distal end of the knee brace to create a flexion torque. The extension was not actuated but achieved through the weight of the instrumented leg simulator. The investigated types of misalignments are a rotation of the brace around the vertical axis and a translation in anteroposterior as well as proximal/distal direction.

Results: The amount of misalignment had a significant effect on several directions of knee joint load in the instrumented leg simulator. In general, load on the knee joint increased with increasing misalignment. Specifically, stronger rotational misalignment led to higher forces in mediolateral direction in the knee joint as well as higher ab-/adduction, flexion and internal/external rotation torques. Stronger anteroposterior translational misalignment led to higher mediolateral knee forces as well as higher abduction and flexion/extension torques. Stronger proximal/distal translational misalignment led to higher posterior and tension/compression forces.

Conclusions: Misalignments of a lower leg exoskeleton can increase internal knee forces and torques during swing to a multiple of those experienced in a well-aligned situation. Despite only taking swing into account, this is supporting the need for carefully considering hazards associated with not only translational but also rotational misalignments during wearable robot development and use. Also, this warrants investigation of misalignment effects in stance, as a target of many exoskeleton applications.

Keywords: Exoskeletons; Joint load; Joint misalignments; Rehabilitation; Safety.

Fig. 1

Overview of the setup. Left: Schematic representation of the setup and marker placement, lateral view. The ILS is shown in green, the orthosis in blue, the frame in grey/black, the sensors in orange and the markers in yellow. Markers XULM2, FTFM, AxisMed, XULMJoint, XLLMJoint, XLLM2 and LowerLegMed are not shown as they are placed on the medial side. Detailed information about marker placement can be found in the Annex. Middle: Photos of the setup. Right: Visualization of internal rotational misalignment (A), posterior translational misalignment (B) and distal translational misalignment (C) of a leg with respect to a knee brace, where the red dashed line and crosses represent the orthosis center of rotation and the green dashed line and crosses represent the leg center of rotation

Fig. 2

Schematic representation of the setup and measures used for calculating the forces and torques at the joint location as well as the actuation torque created by the pulling force

Fig. 3

ILS joint angle, flexion torque and joint forces and torques over the course of a typical trial in aligned setup. The top graph shows the flexion angle of the ILS where 0 deg is full extension (black) and the flexion torque generated through the pulling force applied to the orthosis (red). The middle graph shows the forces in the ILS joint where positive Fx is forward directed force, positive Fy is pointing from medial to lateral, and positive Fz is compressive force. The bottom graph shows the torques in the ILS joint where positive Mx is adduction torque, positive My is flexion torque, and positive Mz is external rotation torque. See also Fig. 2 for axis orientation

Fig. 4

Hysteresis plots of ILS joint forces and torques over the ILS flexion angle. The amounts of rotational misalignment (rot. MA) are represented by the graph colors with darker blue shades representing stronger internal rotation and darker red shades representing stronger external rotation

Fig. 5

Hysteresis plots of ILS joint forces and torques over the ILS flexion angle. The amounts of anteroposterior translational misalignment (transl. MA) are represented by the graph colors with green–blue shades representing posterior translation and orange-red shades representing anterior translation

Fig. 6

Hysteresis plots of ILS joint forces and torques over the ILS flexion angle. The amounts of proximal/distal translational misalignment (transl. MA) are represented by the graph colors with green–blue shades representing distal translation and orange–red shades representing proximal translation

Fig. 7

Behavior of misalignment over flexion/extension cycles of each trial. The amounts of misalignment (rot. / transl. MA) are represented by the graph colors and the starting value of each misalignment setting is marked with a circle. Green to blue shades represent internal rotation, posterior translation and distal translation respectively (left to right), while orange–red shades represent external rotation, anterior translation and proximal translation respectively

Fig. 8

Behavior of flexion torque over flexion/extension cycles of each trial. The amounts of misalignment (rot. / transl. MA) are represented by the graph colors. Green to blue shades represent internal rotation, posterior translation and distal translation respectively (left to right), while orange–red shades represent external rotation, anterior translation and proximal translation respectively