基于BSN和CHMMs的人体日常动作识别方法研究

Research on human daily activity recognition method based on BSN and CHMMs

应用人体传感器网络(body sensor networks,BSN)识别人体日常动作可以有效地提高对老年人、慢性病人,以及术后病人等特殊人群的医疗监护质量.为此建立了一个基于BSN的人体日常动作监督平台,应用采集到的加速度信号识别9个常见的人体日常动作.针对动作识别过程中存在的多传感器数据融合问题,提出一种基于耦合隐马尔可夫模型(coupled hidden Markov models,CHMMs)的动作识别方法.实验结果显示,与已有动作识别方法相比,提出的基于CHMMs的动作识别方法的识别正确率有明显的提高.

Body sensor networks （BSN） may offer continuous monitoring of human activities in a range of healthcare areas, including caring the elderly, helping chronic patients, and monitoring the recovery of post-operative patients. A monitoring platform based on BSN is established for recognizing 9 human daily activities using acceleration signal. A activity recognition method based on coupled hidden Markov models （CHMMs） is proposed for multi-sensor data fusion. The experimental results show that compared with previous methods, the proposed method can achieve satisfactory performance for human daily activity recognition.