The primary purpose of this thesis was to develop, validate and implement a novel ergonomics tool for manual arm strength (MAS) prediction. In Chapter 2, an empirical study was conducted to: 1) fill in gaps in our MAS database, and 2) examine the relationships between MAS and shoulder/elbow moments, to help identify important sources of variance for future predictive modeling attempts. Chapter 3 focused on the evaluation of artificial neural network (ANN) and traditional multiple regression approaches for MAS prediction, and revealed that ANNs provided a more accurate and generalizable prediction of MAS for our specific dataset. Chapter 4 drew on the data and findings of Chapters 2 & 3, and described the development of the ‘Arm Force Field’ (AFF) method for MAS prediction. The AFF method can be used to predict the MAS for any percentage of the population, given only the simple inputs of force vector direction, hand location (relative to the right shoulder), and torso orientation. In Chapter 5, a theoretical examination of the relative changes in wrist strength, due to interacting forearm and wrist postures, was conducted. That study resulted in a set of regression equations that can be used to predict wrist strength correction factors in complex wrist and forearm postures, allowing for more accurate estimations of the limiting joint once the MAS is calculated. An example of the AFF method’s implementation is provided and discussed in Chapter 6. The four studies, presented in this thesis, add to the current knowledge related to strength prediction in ergonomics, and the AFF method has the potential to be easily integrated within digital human models, for more valid estimates of manual force capabilities for the population. / Dissertation / Doctor of Philosophy (PhD)

Introduction: External support behaviours, which include leaning (supporting with the non-task hand) or bracing (supporting with the body), are frequently employed by workers in manufacturing settings. However, current ergonomic assessment tools are limited by our limited understanding of these behaviours. Recent studies have investigated these behaviours, however, the designs of these studies are limited in their applicability to real-world scenarios. The purpose of this study was to assess how different task parameters affect the prediction of external support behaviours, as well as the effect of support on task hand, and body, kinematics and kinetics, in a minimally constrained experimental design. Methods: Female participants (n = 18) performed a series of one-handed maximal exertions (in the six orthogonal directions), and one precision task, in four hand Locations. Trials either featured support (as chosen by the participant), or no support. Results & Discussion: Three logistic regression models were developed, with inputs from individual and task characteristics, and they correctly predicted the occurrence of leaning, bracing, or simultaneous leaning and bracing, 74-86% of the time. Leaning and/or bracing were found to provide: 1) oppositional forces to increase task hand force generation, 2) balance, by countering destabilizing moments about the feet, and 3) a reduction in moment arm of the task hand force, with respect to the upper body joints, by bringing the shoulder closer to the task hand. Participants were able to exert 64.8% more force at the task hand as a result of support. Leaning hand placement depended on the task force direction and location. However, the positioning of the leaning hand varied very little. Finally, the precision condition showed that fine motor demands may also affect external support choice. / Thesis / Master of Science in Kinesiology