Application of trajectory segmentation techniques for operator skill evaluation

This paper compares trajectory segmentation techniques applied to the analysis of human operators' spatial control performance. The goal of the paper is to determine the effectiveness of three levels of analysis: traditional statistics, kinematic segmentation and finally dynamic clustering. The analysis emphasizes the broad range of behaviors associated with human spatial control skills. Such methods are necessary for the development of rating systems and models suited for applications involving human-machine interactions, such as in robotic surgery or teleoperation of agile vehicles. The data used for the analysis was collected using a miniature helicopter in an interactive flight test facility. Outcomes of the three different techniques are compared and discussed.