Classification of long-term motions using a two-layered hidden Markov model in a wearable sensor system
Si-Jung Ryu, Jong-Hwan Kim · 2011
This paper proposes a classification system for long-term motions in a wearable sensor system with 3-axis accelerometers. The long-term motion is defined as a sequence of short-term motions so that the overall classification algorithm processes short-term motions in the first layer and then classifies long-term motions in the second layer. The hidden Markov model is employed in each layer as a classification algorithm. The wearable sensor system consists of two 3-axis accelerometers which are attached to both forearms. Raw data from the accelerometers are pre-processed and forwarded to the classification algorithm designed with the hidden Markov model. For comparison, other algorithms such as artificial neural networks, support vector machine, k-nearest neighbor algorithm and k-means clustering, are tested. In experiments, eight kinds of short-term motions are randomly selected from daily life to test the performance of the proposed system and to compare its performance with that of existing algorithms. Also, three long-term motions which consist of short-term motions are selected and tested to demonstrate the effectiveness of the proposed algorithm.