ABSTRACT
Motion capture is becoming a popular technique for animating a virtual actor by a real person in the game industry. It is also used for biomechanical motion analysis in clinical and sports applications. In recent years, this technique is applied for analyzing and predicting complex human motions in ergonomic simulation by digital human fi gures. Indeed, a design engineer needs to simulate the motions of the future operators/users and to make ergonomic evaluation of a product in a very early phase. Due to limited knowledge on human motion control strategies, the data-based approach is the only one that is capable of simulating complex task-oriented movements today. For this, a well-structured motion database is needed (see for instance Park et al., 2004; Monnier et al., 2006; Wang et al., 2006). Realistic motion reconstruction is the fi rst step for humanlike motion simulation. Th e quality of reconstructed motions directly aff ects the realism of simulated motions. As a large amount of motion data is necessary for simulating diff erent human activities, a fast motion reconstruction method has to be developed. Th is chapter aims to give an overview of diff erent motion reconstruction methods, especially for whole-body motion reconstruction from external markers trajectories.
