论文部分内容阅读
为提高可穿戴多传感数据远程联合重构性能,提出了一种基于分布式压缩感知的可穿戴多传感加速度数据联合重构新方法。该方法首先对可穿戴多传感原始数据压缩编码,将数据融合传送至远端服务器;然后,基于可穿戴传感数据的时空相关性,构建块稀疏贝叶斯学习联合重构算法,实现压缩数据解码,准确重构各传感原始数据;最后,新方法对美国加州伯克利大学可穿戴多传感运动数据进行分析。实验结果表明,对不同编码采样率,文章所提方法重构性能明显优于传统的算法,并且能够准确解码压缩数据,有望在远程医疗环境下推广应用。
In order to improve the performance of remote joint reconstruction of wearable multi-sensor data, a new joint wearable multi-sensor acceleration data reconstruction method based on distributed compression sensing is proposed. In this method, the wearable multisensor raw data is first compressed and encoded, and the data is fused and transmitted to the remote server. Then, based on spatio-temporal correlation of the wearable sensor data, a joint sparse Bayesian learning reconstruction algorithm is constructed to compress Data decoding, accurately reconstruct the raw data of each sensor; Finally, the new method is used to analyze the wearable multisensory data of the University of California at Berkeley. Experimental results show that the reconstruction performance of the proposed method is obviously better than that of traditional algorithms for different sampling rates, and it can accurately decode compressed data and is expected to be widely used in telemedicine environment.