An application of mobile robotics for olfactory monitoring of hazardous industrial sites

Purpose – The purpose of this paper is to present a mobile robot with an olfactory capability for hazardous site survey. Possible applications include detection of gas leaks and dangerous substances along predefined paths, inspection of pipes in factories, and mine sweeping. Design/methodology/approach – The mobile sentry is equipped with a transducer array of tin oxide chemical sensors, compliant with the standard interface IEEE 1451, which provides odour-sensing capability, and uses differential drive and spring-suspended odometric trackballs to move and localize in the environment. The monitoring strategy comprises two stages. First, a path learning operation is performed where the vehicle is remotely controlled through some potential critical locations of the environment, such as valves, pressure vessels, and pipelines. Then, the robot automatically tracks the prerecorded trajectory, while serving as an electronic watch by providing a real-time olfactory map of the environment. Laboratory experiments are described to validate the approach and assess the performance of the proposed system. Findings – The approach was shown to be effective in experimental trials where the robot was able to detect multiple odour sources and differentiate between sources very close to one another. Research limitations/implications – One limitation of the methodology is that it has been specifically designed for odour detection along a welldefined path in a highly structured environment, such as that expected in the industrial field. The problem of detection of leakages outside the search path is not addressed here. Practical implications – This mobile robot can be of great value to detect hazardous fluid leakages in chemical warehouses and industrial sites, thus increasing the safety level for human operators. Originality/value – The paper describes a mobile robotic system, which employs an odour-sensing capability to perform automated monitoring of hazardous industrial sites. A dynamic model of the mobile nose is also discussed and it is shown that it well describes the behaviour of the system.