Lossless and Lossy Direct Compression Design With Multi-Signal Symptom Detection for Low-Temperature Wearable Devices

摘要

In this paper, the lossless and lossy direct compression design (LLDC) is proposed for the wearable devices. The electrocardiogram (ECG), blood pressure (BP), and respiration (RESP) can be applied to the compression design for cardiovascular diseases. The proposed LLDC can detect the abnormal symptoms and dynamically change the compression modes with different sample intervals and data precisions. In our experiments, the symptoms are detected, and the transmitted power and temperature for the wearable devices are reduced. The compression ratios for ECG, BP, and RESP are up to 7.94, 6.06, and 6.51 with percentage root-mean-square difference less than 6.5%. For ECG, BP, and RESP, the energies of the wearable devices are reduced by 50%–87%, 38%–89%, and 41%–89%; the temperatures are reduced by 2.6 °C–4.2 °C, 1.3 °C–3 °C, and 1.7 °C–3.6 °C, respectively.