Humanitarian Demining Using an Insect Based Chemical Unmanned Aerial Vehicle

Nowadays, there are about 100 million active landmines distributed around the world as the result of earlier conflicts (Habib 2007). These cheap and simple to manufacture weapons have a long lasting effect that cause injury to the civil population even decades after the conflict has ended. Their removal is very expensive, dangerous and time consuming, and it has a high social and economical impact. Hence, although demining is a necessity for the affected populations to recover, it poses fundamental technical challenges. At the present time, the use of animals, i.e. biological sensors, still provides the highest accuracy and safety standards in demining tasks. A number of animal species including dogs, rats and bees have been successfully trained to detect and localize landmines following the minute chemical trails of leaking explosive compounds (Fjellanger et al. 2002; Bromenshenk et al. 2003; Verhagen et al. 2006). Insects, and in particular moths, are highly optimized chemical detection systems that are extremely proficient at the detection and localization of different chemical compounds, in particular pheromones, at very long distances, i.e. several hundreds of meters. Moths use pheromone signals for sexual communication and it has been shown that males are able to detect and distinguish minute amounts of female pheromones (as little as 104 molecules·cm–3) against a background of other chemicals in very irregular and unpredictable plumes (Wyatt 2003). As a consequence, the evolutionary pressure to detect pheromones has generated specific neural and behavioral adaptations to deal with this specific problem. Nonetheless, the chemical search task is not reduced to a unique olfactory process but is a multi-modal task that includes the integration of complex behavioral strategies with visual, olfactory and wind sensing information (Kennedy & Marsh 1974; Ludlow 1982; Charlton & Carde. 1990). In this chapter we analyze the relationship between the chemical detection and localization problem and its biological solution, and we will show how our understanding of the biological solution can be exploited to construct efficient autonomous chemo-sensing Unmanned Aerial Vehicles (cUAV). Firstly, we describe a blimp-based technology for a cUAV. Subsequently, we investigate the computational and behavioral principles underlying the opto-motor system of the fly and the locust, and we show that relying solely