This study used optimisation procedures in conjunction with an 8-segment torque-driven computer simulation model of the takeoff phase in springboard diving to determine appropriate subjectspecific strength parameters for use in the simulation of forward dives. Kinematic data were obtained using high-speed video recordings of performances of a forward dive pike (101B) and a forward 2½ somersault pike dive (105B) by an elite diver. Nine parameters for each torque generator were taken from dynamometer measurements on an elite gymnast. The isometric torque parameter for each torque generator was then varied together with torque activation timings until the root mean squared (RMS) percentage difference between simulation and performance in terms of joint angles, orientation, linear momentum, angular momentum, and duration of springboard contact was minimised for each of the two dives. The two sets of isometric torque parameters were combined into a single set by choosing the larger value from the two sets for each parameter. Simulations using the combined set of isometric torque parameters matched the two performances closely with RMS percentage differences of 2.6% for 101B and 3.7% for 105B. Maximising the height reached by the mass centre during the flight phase for 101B using the combined set of isometric parameters and by varying torque generator activation timings during takeoff resulted in a credible height increase of 38 mm compared to the matching simulation. It is concluded that the procedure is able to determine appropriate effective strength levels suitable for use in the optimisation of simulated forward dive performances.
IntroductionComputer simulation models of various sports movements have been developed in order to investigate the mechanics of optimum performance (Alexander, 1990;King and Yeadon, 2004;Koh et al., 2003). Crucial to the development of such models is that they must be accurate representations of the biomechanical systems being studied. To assess the accuracy of a simulation model and show that it is sufficiently complex to represent the biological system, subject-specific values of inertia, strength and visco-elastic parameters are required so that the output of the model can be compared to recorded performances by the same subject in a quantitative manner . Subject-specific strength parameters are difficult to determine and have typically been calculated from isovelocity torque measurements on the subject over a range of joint angles and angular velocities at each joint Yeadon, 2002, King et al., 2006;Mills et al., 2008). The limitations of this technique are that it requires access to the subject and it can be difficult for the subject to produce maximal torque values during each trial on the dynamometer.Recently, torque-driven computer simulation models have been used to provide insights into springboard diving takeoff techniques. Cheng and Hubbard (2004) used a two-segment model with theoretically assigned muscle parameters to optimise knee extension strategies for maximum ...