KSAS (Kenpo Set Assisting System) is a mobile application that is able to assist in the learning of a set of movements of American Kenpo Karate, known as Blocking Set I. This was selected this martial arts because it entails many of the characteristics common to other psychomotor activities like coordination of different parts of the body, the use of objects or the interaction with other practitioners. This application is able to teach the order in which the movements are to be executed, using the mCMAR2 framework (Motion Capture, Modeling, Analysis, Response and Report).

As shown in the video, KSAS is able to capture the movements of the arm using the inertial sensors of an Android device, obtaining a set of 18 values each 20ms (Phase 1 – Motion Capture). The captured movements are then modelled as temporal series and smoothed using the Exponentially Weighted Moving Averages algorithm (EWMA) (Phase 2 – Motion Modeling). A Long Short-Term Memory (LSTM) neural network has been trained using data captured from 20 volunteers to analyze the movements and is then used as a classifier of movements (Phase 3 – Motion Analysis). Using information extracted from the analysis, the application is able to give verbal indications and feedback to the user, as well as haptic feedback (Phase 4 – Motion Response). Finally, a report is generated using information from the execution with the purpose of showing the progress of the learner to the teacher and, if desired, to other learners (Phase 5– Motion Reports).

Points of Contact: Alberto Casas-Ortiz (acasas148@alumno.uned.es) & Olga C. Santos (ocsantos@dia.uned.es), Computer Science School, UNED. Madrid, Spain