الفهرس | Only 14 pages are availabe for public view |
Abstract This thesis is concerned with the problem of monitoring and tracking human activities. Real-time recognition and monitoring of the human daily activities attracted lots of attention from researchers due to its crucial role in emerging elds such as pervasive healthcare and ambient assisted living. However, despite being an active eld, recent benchmarks show a diculty in providing robust monitoring of the dierent various activities under realistic everyday life conditions. Through this thesis, we present the ubiMonitor system as an accurate real-time activity monitor using low-cost o-the-shelf three body-worn 3D accelerometers. For a robust monitoring, ubiMonitor (1) intelligently fuses the accelerometers and the acceleration in each of the three axes to provide key discriminative features for the dierent activities based on their physical characteristics, (2) employs a novel hierarchical activity recognition scheme, and (3) applies an eective temporal- and context-based postprocessing stage to remove falsely detected activities and enhance the overall system accuracy . Experimental results using real traces from dierent subjects show that ubiMonitor can achieve an overall accuracy more than 95% with a median latency less than 3 msec. This is better than state-of-the-art by 23.4% in the recognition accuracy with a reduction of 70% in the sensors used. |