Classification of Robot Service during Sit-to-Stand through Segments Coordination

Service robots had become widespread, but reputation towards these robots remained low because of the suboptimal service. Sit-to-Stand (STS) transition is frequently performed during daily life and essential for the elderly. However, little research had contributed to measuring and classifying robot service during STS transition. In this paper, we aimed to propose a method to classify the suboptimal service of robot through segments coordination and machine learning method. Six subjects participated in our research. STS experiments were performed under two conditions according to robot service duration of 2 and 5 s. Relative Phase (RP) was utilized to evaluate segment coordination and identify suboptimal robot service. Support Vector Machine (SVM) was used to classify robot service. The results showed that the absolute maximum RP became 1.6 times larger for suboptimal robot service. SVM based classification showed high accuracy rate (87.9%) to classify suboptimal robot service. It was concluded that suboptimal robot service resulted from poor human movement coordination. Suboptimal robot service can be classified by using machine learning method.