“…The adaptive VDHMM Proceedings of the 6th International Conference on Automation, Robotics and Applications, Feb 17-19, 2015, Queenstown, New Zealand matrix, such that the mean of state duration will be best approximated (maximum likelihood). In this section, the adaptive training results are presented based on the training set {T training } consisting of a randomly picked combination: T 2 ,T 3 ,T 5 ,T 8 ,T 9 , T 10 and T 11 .…”