作者: Pavel Dohnálek , Petr Gajdoš , Tomáš Peterek
DOI: 10.1007/978-3-319-00542-3_49
关键词:
摘要: Human physical activity monitoring is a relatively new problem drawing much attention over the last years due to its wide application in medicine, homecare systems, prisoner etc. This paper presents Orthogonal Matching Pursuit as method for recognition and proposes modification that significantly increases accuracy. Both methods show promising results both total differentiation between certain activities even without necessity of prior data preprocessing. The were tested on raw sensor data.