Online Intention Recognition for Computer-Assisted Teleoperation
MetadataShow full item record
An online intention recognition algorithm for computer-assisted teleoperation is introduced. The algorithm is able to distinguish between phases of a typical object manipulation task. It adopts a new advanced feature extraction algorithm which extracts features from haptic data and uses a Hidden Markov Model for stochastic classification. The method is implemented and validated on a real hardware setup. The obtained results reveal a robust and fast intention recognition.