Human Action Recognition Research Based On EEMD And Fuzzy LS-SVM

There has been an exceptional development of microelectronics and computer systems in the past decade, enabling sensors and mobile devices with unprecedented characteristics. Their high computational power, small size, and low cost allow people to interact with the devices as part of their daily living. That was the genesis of Ubiquitous Sensing, the main purpose of which is extracting knowledge from the data acquired by pervasive sensors. Particularly, the recognition of human activities has become a task of high interest within the field, and it play an important role in promoting the development of health, sports, military, security, and so on.The first works on human activity recognition (HAR) date back to the late’90s. However, there are still many issues that motivate the development of new techniques to improve the accuracy. Some of these challenges are (1) the selection of the attributes to be measured,(2) the design of feature extraction and inference methods,(3) the flexibility to support new users without the need of re-training the system, and (4) the implementation in mobile devices meeting energy and processing requirements.The recognition of human activities has been approached in two different ways, namely using external and wearable sensors. In the former, the devices are fixed in predetermined points of interest, so the inference of activities entirely depends on the voluntary interaction of the users with the sensors. In the latter, the devices are attached to the user. Within the wearable sensors, inertial sensor is one of the most common and most effective devices used in recognizing human activities; these sensors provide a reliable and objective measurement of physical activity.A novel approach for recognizing human activities investigated in this article, which combines ensemble empirical mode decomposition (EEMD) and fuzzy least-squares support vector machine (Fuzzy LS-SVM). First, the inertial data were recorded with a number of wireless sensors, and then the time-domain signal features and the time-frequency signal features based on EEMD were extracted from these data. Feature sets obtained in above step are processed by SBMLR to get a lower dimensions but the most discerning feature subset. Finally, the reduced feature vectors are then classified by the fuzzy LS-SVM technique.