Department of Computer Science and Technology, Harbin Institute of Technology, Heilongjiang 150001, China.
Sensors (Basel). 2020 May 31;20(11):3117. doi: 10.3390/s20113117.
With the development and popularity of micro-electromechanical systems (MEMS) and smartphones, sensor-based human activity recognition (HAR) has been widely applied. Although various kinds of HAR systems have achieved outstanding results, there are still issues to be solved in this field, such as transition activities, which means the transitional process between two different basic activities, discussed in this paper. In this paper, we design an algorithm based on standard deviation trend analysis (STD-TA) for recognizing transition activity. Compared with other methods, which directly take them as basic activities, our method achieves a better overall performance: the accuracy is over 80% on real data.
随着微机电系统（MEMS）和智能手机的发展和普及，基于传感器的人体活动识别（HAR）得到了广泛应用。尽管各种 HAR 系统已经取得了优异的成果，但在这个领域仍有一些问题需要解决，例如本文所讨论的过渡活动，即两种不同基本活动之间的过渡过程。在本文中，我们设计了一种基于标准差趋势分析（STD-TA）的算法来识别过渡活动。与直接将过渡活动视为基本活动的其他方法相比，我们的方法取得了更好的整体性能：在真实数据上的准确率超过 80%。
