Estimating centre of mass trajectory and subject-specific body segment parameters using optimisation approaches

Whole body dynamics analyses are compromised by various error sources including body segment parameter (BSP) and ground reaction force (GRF) measurement errors. This research employed nonlinear optimisation techniques, attempting to account for such errors and, thus, improve dynamical representation of whole body movement activities. The first experiments demonstrated new optimisation-based integration approaches (IA optimisation methods) for determining whole body centre of mass (CM) trajectory based on double numerical integration of acceleration data derived exclusively from GRF measurements. The zero-point-to-zero-point (ZPZP) method of representing CM horizontal trajectory (King and Zatsiorsky, 1997; Zatsiorsky and Duarte, 2000) was modified by including a GRF measurement offset error term and other design variables in an optimisation process for determining CM trajectory relative to centre of pressure data. Much smoother, more realistic CM trajectory was produced by the new ZPZP IA optimisation method. New IA optimisation techniques for estimating CM trajectory during jumping activities were also demonstrated. The vertical dimension methods were all appropriate for determining transient jump performance parameters commonly calculated in jumping assessments (Hatze, 1998), including CM jump height, work and power. The final experiment presented methods of optimising inverse dynamics analyses by selecting optimal GRF measurement offset error terms and BSPs. Feasible and realistic GRF offset error terms were invariably produced. However, approximately 50% of all estimated BSPs were unrealistic under most tested conditions. Improved modelling and more contemporary motion capture technology may improve results, and may ultimately lead to the development of a versatile, relatively non-invasive and subject-specific BSP estimation method