Approach to weak signal detection via over-sampling and bidirectional saw-tooth shaped function in wearable devices

Abstract

Many biological signals that reflect the human health are relatively weak, and high-resolution Analog-Digital Converters (ADCs) are required to measure them. The over-sampling technology is a method for increasing the accuracy, equal to the accuracy of high resolution ADCs, when an ADC has low resolution. For high resolution ADCs, their accuracy from the over-sampling technology would break its maximum limitation. During our signal collection, there are two items which affect our results: 1. measurement error; 2. non-linear characteristics at the end points. In over-sampling method of detecting weak signal of wearable devices by using the unidirectional saw-tooth shaped function, the magnitude of the shaped function is as close as possible to the integer multiple of the quantization step size to reduce the additional measurement error. To reduce the difficulty of the adding shaped function, which features nonlinearity of the starting point, a new method of detecting the weak signal by adding the bidirectional saw-tooth shaped function is proposed. The experimental results show this method is feasible for improving the ADC resolution.