An automatic method to reconstruct human movement by considering a reduced number of information: Application to the automobile ingress movement.

Measuring the automobile ingress movement is a stage towards the definition of a human physical model exploitable in new vehicles design. A part of information describing the movement is not measurable due to experimental constrains. A 2 stages method is proposed to reconstruct human movement. The first stage aims at building, automatically, a human model adapted to the anthropometry of the subject without manual reshaping on the subject. The second stage uses the human model and carries out a trajectory tracking based on a global optimization of the whole body under articular constraints. Two reconstructions of an automobile ingress movement, realized by considering the whole markers and a reduced number of markers, are compared. The articular coordinates resulting from these 2 reconstructions are correlated at more than 0.86, their average difference is lower than ±4.7° and their RMS is lower than 7°. The proposed method, thus, constitutes a good alternative to calculate the articular coordinates with a reduced number of information to study the automobile ingress movement.