Output details
11 - Computer Science and Informatics
Goldsmiths' College
Learning Finite State Machine Controllers from Motion Capture Data
<24> Published in the highest ranked games journal (SCImago), the paper develops the methodology presented in Gillies et al (2008), allowing more complex behaviour models. Rather than the reactive models used in the previous paper, this research includes models with state, in which behaviour can evolve over time and in response to interaction. The models were tested on different human behaviour datasets. Cited by researchers in Concordia and Hong Kong Polytechnic. This work led to the EPSRC funded project “Performance-Driven Expressive Virtual Characters” (EP/H02977X/1). All project reviewers gave it 6 out of 6 in their overall assessment.