Synthesis of Responsive Motion Using a Dynamic Model

Abstract

Synthesizing the movements of a responsive virtual character in the event of unexpected perturbations has proven a difficult challenge. To solve this problem, we devise a fully automatic method that learns a nonlinear probabilistic model of dynamic responses from very few perturbed walking sequences. This model is able to synthesize responses and recovery motions under new perturbations different from those in the training examples. When perturbations occur, we propose a physics-based method that initiates motion transitions to the most probable response example based on the dynamic states of the character. Our algorithm can be applied to any motion sequences without the need for preprocessing such as segmentation or alignment. The results show that three perturbed motion clips can sufficiently generate a variety of realistic responses, and 14 clips can create a responsive virtual character that reacts realistically to external forces in different directions applied on different body parts at different moments in time.

Augmented Motion History Volume for Spatiotemporal Editing of 3D Video in Multi-party Interaction Scenes

“…Ye et al [7] have proposed a fully automatic method that learns a nonlinear probabilistic model of dynamic responses from very few perturbed sequences. Their model is able to synthesize responses and recovery motions under new perturbations different from those in the training examples.…”

Conditional stochastic simulation for character animation

“…Most recent applications of Gaussian processes include motion editing, 8 motion synthesis of a responsive virtual character 15 , and stylecontent separation. 16 Contrary to these previous works, our method does not require any global optimization procedure as it can perform sequentially, thus making it fully suitable for real time systems, even with a large number of characters such as in a crowd.…”

Continuous character control with low-dimensional embeddings

“…The Gaussian process latent variable model (GPLVM) is a nonlinear generalization of principal component analysis that learns such a space [Lawrence 2005]. Prior methods have applied the GPLVM to motion data [Grochow et al 2004;Urtasun et al 2005;Wang et al 2008;Ye and Liu 2010] for tasks such as human tracking in video and constrained motion synthesis. However, no prior method has used the model for interactive, taskdriven control, and prior applications have largely been limited to modeling homogeneous datasets such as forward walking.…”