The aim of the work presented in this paper is to realise the model for the control of a robotic interface for equilibrium assistance during sit-to-stand transfer. Here, it is thought that interactive robotic devices, as human-centered robotics, is more comfortable and more efficient that traditionnal technical devices. One supposes the need of virtual model of the pathology. This model, called observer, aims at being used in the smoothing control part of this assisting device. A usefull property of this observer should be a prediction ability of a postural instability. This article presents a study of different neural solutions : a Neural Trajectory Generator and a Neural Predictive Observer (NPO). Records used for the learning have been done from healthy people that stand up normally and quickly. Some tests will also be done in patients with cerebellar syndrome disease. The presented experimental results show the good accuracy of these approaches whatever the speed of the movement.