Abstract
This paper presents a signal acquisition and processing chip. The chip connects with a chemiresistive sensor array and directly acquires array data. A built-in neural network cluster enhances the signal processing ability of the proposed chip and enlarges the distance between various odour groups. When noise is induced in the training phase, the cluster is robust against background interference. The chip is space-efficient and features a flexible hardware structure. The chip was fabricated using the TSMC 90-nm CMOS process. The chip area is 2.0 × 2.6 mm2. The supply voltage is 1.0 V. The power consumption is approximately 4 mW. The measurements revealed that the chip excellently identified the odours of methanol, ethanol, and a mixture of 50% methanol + 50% ethanol. Gas-sensing systems based on this signal acquisition and processing chip can be extensively implemented in diverse applications. © 2014 IEEE.