Real-Time Edge Neuromorphic Tasting From Chemical Microsensor Arrays.

Nicholas LeBow,Josep Maria Margarit-Taulé,Pengfei Sun,Bodo Rueckauer,Cecilia Jiménez-Jorquera,Meritxell Rovira,Shih-Chii Liu

doi:10.3389/fnins.2021.771480

Abstract

Liquid analysis is key to track conformity with the strict process quality standards of sectors like food, beverage, and chemical manufacturing. In order to analyse product qualities online and at the very point of interest, automated monitoring systems must satisfy strong requirements in terms of miniaturization, energy autonomy, and real time operation. Toward this goal, we present the first implementation of artificial taste running on neuromorphic hardware for continuous edge monitoring applications. We used a solid-state electrochemical microsensor array to acquire multivariate, time-varying chemical measurements, employed temporal filtering to enhance sensor readout dynamics, and deployed a rate-based, deep convolutional spiking neural network to efficiently fuse the electrochemical sensor data. To evaluate performance we created MicroBeTa (Microsensor Beverage Tasting), a new dataset for beverage classification incorporating 7 h of temporal recordings performed over 3 days, including sensor drifts and sensor replacements. Our implementation of artificial taste is 15× more energy efficient on inference tasks than similar convolutional architectures running on other commercial, low power edge-AI inference devices, achieving over 178× lower latencies than the sampling period of the sensor readout, and high accuracy (97%) on a single Intel Loihi neuromorphic research processor included in a USB stick form factor.

Highlights

Liquid analysis systems that assess process quality in sectors like food, beverage, and chemical manufacturing are in rising demand
We evaluated both k-nearest neighbors (k-NN) and convolutional neural networks (CNNs) accuracies initially using all nine sensor readings from the test dataset
Our electronic tongue uses rate-based, deep spiking convolutional networks to fuse dynamic, electrochemical microsensor readings. It performs with high accuracy (97%) and high energy efficiency (0.1 mJ per inference), and it can run the neural networks in real time (5.6 ms, over 178× lower than the sampling period of the sensor readout employed by the system) on a single Loihi chip

Summary

Introduction

Liquid analysis systems that assess process quality in sectors like food, beverage, and chemical manufacturing are in rising demand. Driven by increasingly strict regulations, and by the need to boost productivity and to reduce costs, industry has promoted the development of automated systems for monitoring physicochemical properties of products in their manufacturing cycle. To allow process control when and wherever required, such systems must be small, energetically autonomous, and able to operate in real time. In this context, the use of chemical multisensor arrays as “electronic tongues” stands out due to their capability of recognizing quantitative and qualitative composition of complex solutions. Artificial tongues use an array of chemical sensors (i.e., the artificial taste cells) selective—but not specific—to different solution properties. The multivariate sensor responses are read out in the electrical domain and modeled by appropriate Machine Learning methods

Objectives

Methods

Results

Conclusion