The Sniffbot: A biohybrid robot for active sensing-based odor localization and discrimination

Abstract

The detection, identification and localization of volatile compounds are of critical importance for various applications, ranging from gas leak detection to drug and explosive sensing. Current technologies--such as gas chromatography-mass spectrometry and e-noses--are limited by slow analysis, low mobility, and reduced sensitivity and adaptability, making them unsuitable for real-time odor localization in real-world settings. Here, we present Sniffbot: an autonomous, mobile biohybrid robotic sensory system that overcomes these challenges by harnessing the extraordinary olfactory capabilities of the desert locust antenna, an advanced olfactory sensor, that generates odorant-specific electrophysiological responses to numerous odorants. Our Sniffbot platform consists of a compact robotic vehicle onto which we have assembled: (i) a sensing module, comprising a locust antenna and a miniaturized electrophysiology system; (ii) a "sniffing" module, which actively samples air in the environment, creating a timed airflow over the antenna, preventing the antenna from becoming habituated to odorant stimuli; and (iii) a decision-making module that analyzes the sensory input in real time to navigate or identify odors. Sniffbots movements are controlled by an odorant-search algorithm coupled with the sniffing module. This enables Sniffbot to detect and localize odors independently of wind-induced odorant gradients, and thus to be used in challenging windless environments. The Trident, a novel search algorithm, outperforms several commonly used algorithms in localizing the odorant source. We further demonstrate Sniffbots ability to discriminate a target odor among others. Our results demonstrate the potential of augmenting biological sensors with autonomous robotic components for next-generation chemical sensing and environmental monitoring.