Development of an Iterative Learning Control Architecture to Evaluate Ex-Vivo Shoulder Biomechanics During Simulated Active Motion

The shoulder is highly unconstrained and is susceptible to injury and disease. While shoulder injuries and pathologies are common, their effects on joint biomechanics are not well understood. Ex-vivo simulators allow shoulder biomechanics to be studied in controlled environments to improve our understanding of shoulder function. To date, the most sophisticated simulators only perform simple planar motions at quasi-static speeds, limiting their utility. The aim of this research was to develop and validate a shoulder motion simulator capable of multiplanar motion at elevated speeds.\nThe developed simulator used an iterative learning tendon excursion control architecture with open-loop execution and independent muscle control to perform multiplanar motion at elevated speeds. Motion tracking errors were below 1.5° and the observed muscle force relationships were consistent with previous literature. Following its validation, the simulator was used in a number of biomechanical studies.\nThe first study measured surface strains along the scapular spine and acromion in the native shoulder and showed that surface strains were significantly affected by the deltoid load vector. The second study mapped the three-dimensional moment arms of eight major muscles across a multiplanar range of motion and showed that the muscles in the shoulder had multi-dimensional functions that changed with joint orientation. Study three investigated the effect of rotator cuff deactivation pattern on shoulder function. Consistent with existing clinical data, deactivation patterns involving three tendons resulted in functional deficits while all deactivation patterns were associated with increases in deltoid force and scapula strain. The final study investigated the effect of humeral offset on reverse total shoulder arthroplasty biomechanics. While humeral offset had a significant effect on muscle function and scapula strains, it did not influence joint compression or centre of pressure.\nThis research represents a significant advancement in ex-vivo simulation technology. It encompasses the development of a novel iterative learning control architecture capable of faster and more complex motions than has been previously reported. The studies conducted in this work have expanded our understanding of shoulder biomechanics and the developed simulator will continue to assist researchers and clinicians in improving patient outcomes.