A new data-dependent energy reduction algorithm for successive approximation register (SAR) analog-to-digital convert (ADC) is presented in this paper. The proposed algorithm starts with a less significant bit (LSB) window with N-bit length, which is configurable depending on signal characteristics. By using less significant bit to more significant bit (L2M) successive extending (SE), the signal window is self-adaptive to cover the input signal within boundary. The proposed technique leads to less mean bit trials per sample, suggesting higher energy efficiency in many data-dependant ADC applications. Furthermore, this algorithm can be implemented based on conventional charge redistribution SAR ADC without any change in analog circuits. According to MATLAB simulation, the proposed technique is able to reduce mean bit trials effectively in biomedical signal detection applications. The simulation results show 37.9%, 32.3% and 18.2% less mean bit trials than using conventional SAR algorithm in processing electrocardiogram (ECG), electroencephalogram (EEG) and in electro-myography (EMG) signals respectively.

提出了一种新的基于数据的能量减少算法，用于逐次逼近寄存器（SAR）模数转换（ADC）。所提出的算法从具有N位长度的较低有效位（LSB）窗口开始，该窗口可以根据信号特性进行配置。通过使用低有效位到高有效位（L2M）连续扩展（SE），信号窗口可以自适应以覆盖边界内的输入信号。所提出的技术导致每个样本的平均比特试验更少，这表明在许多数据相关的ADC应用中，能量效率更高。此外，该算法可以基于常规电荷重新分配SAR ADC来实现，而无需在模拟电路中进行任何更改。根据MATLAB仿真，所提出的技术能够有效地减少生物医学信号检测应用中的平均比特试验。仿真结果显示，与传统的SAR算法相比，在处理心电图（ECG），脑电图（EEG）和肌电图（EMG）信号方面，平均比特试验分别减少了37.9％，32.3％和18.2％。