Wearable activity recognition for automatic microblog updates
Hao Tian, Pang Lei, Li Xingjuan, Shusong Xing · 2009
Activity recognition system based on MEMS sensors and wearable computer can produce updates on microblog automatically and frequently. Miniature sensors like the accelerometers made with MEMS technologies can measure body movements continuously while GPS constrains activity recognition within a small set of possible activities. Context information about a person's activities is made into short text messages shared among group members in order to keep social networking. The work for identifying daily activities from inertial sensor data with hidden Markov models (HMMs) is described.