In order to evaluate the ability of human standing balance scientifically, we in this study proposed a new evaluation method based on the chaos nonlinear analysis theory. In this method, a sinusoidal acceleration stimulus in forward/backward direction was forced under the subjects' feet, which was supplied by a motion platform. In addition, three acceleration sensors, which were fixed to the shoulder, hip and knee of each subject, were applied to capture the balance adjustment dynamic data. Through reconstructing the system phase space, we calculated the largest Lyapunov exponent (LLE) of the dynamic data of subjects' different segments, then used the sum of the squares of the difference between each LLE (SSDLLE) as the balance capabilities evaluation index. Finally, 20 subjects' indexes were calculated, and compared with evaluation results of existing methods. The results showed that the SSDLLE were more in line with the subjects' performance during the experiment, and it could measure the body's balance ability to some extent. Moreover, the results also illustrated that balance level was determined by the coordinate ability of various joints, and there might be more balance control strategy in the process of maintaining balance.
The requirement for unconstrained monitoring of heartbeat during sleep is increasing, but the current detection devices can not meet the requirements of convenience and accuracy. This study designed an unconstrained ballistocardiogram (BCG) detection system using acceleration sensor and developed a heart rate extraction algorithm. BCG is a directional signal which is stronger and less affected by respiratory movements along spine direction than in other directions. In order to measure the BCG signal along spine direction during sleep, a 3-axis acceleration sensor was fixed on the bed to collect the vibration signals caused by heartbeat. An approximate frequency range was firstly assumed by frequency analysis to the BCG signals and segmental filtering was conducted to the original vibration signals within the frequency range. Secondly, to identify the true BCG waveform, the accurate frequency band was obtained by comparison with the theoretical waveform. The J waves were detected by BCG energy waveform and an adaptive threshold method was proposed to extract heart rates by using the information of both amplitude and period. The accuracy and robustness of the BCG detection system proposed and the algorithm developed in this study were confirmed by comparison with electrocardiogram (ECG). The test results of 30 subjects showed a high average accuracy of 99.21% to demonstrate the feasibility of the unconstrained BCG detection method based on vibration acceleration.
With the rapid improvement of the perception and computing capacity of mobile devices such as smart phones, human activity recognition using mobile devices as the carrier has been a new research hot-spot. The inertial information collected by the acceleration sensor in the smart mobile device is used for human activity recognition. Compared with the common computer vision recognition, it has the following advantages: convenience, low cost, and better reflection of the essence of human motion. Based on the WISDM data set collected by smart phones, the inertial navigation information and the deep learning algorithm-convolutional neural network (CNN) were adopted to build a human activity recognition model in this paper. The K nearest neighbor algorithm (KNN) and the random forest algorithm were compared with the CNN network in the recognition accuracy to evaluate the performance of the CNN network. The classification accuracy of CNN model reached 92.73%, which was much higher than KNN and random forest. Experimental results show that the CNN algorithm model can achieve more accurate human activity recognition and has broad application prospects in predicting and promoting human health.