针对可穿戴心电设备低功耗采集传输与高精度重构问题，展开全新压缩感知与重构技术研究，主要内容包括：⑴嵌入式心电压缩感知。研究构建满足RIP条件，适合可穿戴式设备的随机二值观测矩阵，实现心电压缩感知，从源头上解决大数据量、高功耗问题。⑵基于综合框架的高精度压缩感知重构。包括基于样本和奇异值分解的符合信号特征的稀疏字典表达和基于多通道信号时空域关联特征先验知识的块贝叶斯学习，以解决心电信号高精度重构问题。⑶直接压缩感知域心电信号模式识别。不经传输到远程终端和重构，完成压缩采样信息直接识别，以解决可穿戴设备健康监视高效实时预警问题。.特色和创新在于可穿戴式心电设备全新压缩感知低功耗技术与基于综合框架的高精度重构。项目属于数学、信息科学与医学交叉领域应用基础研究，由可穿戴式心电设备低功耗技术需求驱动，且可广泛应用于其他可穿戴设备、物联网传感节点等大数据量、低功耗技术领域。

The project researches on key techniques of low-power compressive sensing (CS), transmission and reconstruction of wearable electrocardiogram (ECG) device to solve its low-power consumption demand. 1) Compressed sampling technique based on binary random observation matrix. Binary random observation matrix is easy to be realized in embedded wearable devices for CS. How to construct a binary observation matrix which satisfies RIP conditions is one of the essential contents of the project. 2) Comprehensive reconstruction techniques of CS ECG signal. Include: sample and KSVD based sparse dictionary construction in accordance with signal characteristics, and block sparse Bayesian learning based on spatio-temporal correlation of multi-channel ECG signals and heuristic knowledge to improve the reconstruction precision. 3) Explore the novel pattern recognition techniques in CS domain for abnormal ECG signal to enrich the CS theory and fill the technique gap..The project characteristic and innovations lie in two aspects: the completely novel compressed sampling based low-power consumption technique and the comprehensive scheme based high-precision reconstruction technique. The project belongs to the aim-oriented applied background research with the combination of mathematics and information science. It is driven by the wearable ECG device low-power consumption demand and will be widely used in other wearable devices and sensor network nodes.