论文题目:20.2 A 67μW/Channel, 0.13nW/Synapse/b Nose-on-a-Chip for Noninvasive Diagnosis of Diseases with On-Chip Incremental Learning
作者:Dexuan Huo, Yu-Hsien Lin, P. K. Shihabudeen, Jilin Zhano, Tao Li, Chi-Rong Chou, Zhihua Wang, Kea-Tiong Tang, Hong Chen
期刊:ISSCC 2025
年份:2025.16-20 Feb.
卷(期)及页码:pp.350-352
摘要:
Portable electronic noses (E-noses) are proposed to detect possible pathological changes in the body by analyzing the patient's exhaled gas [1]. As a diagnostic tool for early disease detection, this approach helps to reduce the need for tissue sampling, greatly alleviating pain and discomfort for the patient while avoiding the risk of infection or complications. However, the composition of exhaled gas varies greatly depending on the environment with changing conditions and patients using different devices in different locations (such as hospital, health center, home and so on), which makes the E-nose difficult to maintain a satisfying accuracy in detecting diseases for different patients. Besides, with the increasing demand for intelligence of E-noses, the circuit complexity and the amount of gas data to be processed are constantly growing, as a result, the overall power consumption issue poses a great challenge for power-constrained portable E-noses, as shown in Fig. 20.2.1 (middle).