Toward a fully autonomous hovercraft visually guided thanks to its own bio-inspired motion sensors

Based on a biorobotic approach developed in our laboratory over the past 25 years, we have designed and built several terrestrial and aerial vehicles controlling their position and speed on the basis of optic flow cues (Franceschini et al. 1992, Viollet and Franceschini 1999, Ruffier and Franceschini 2005, Franceschini et al. 2007, Kerhuel et al. 2010). In particular, in our project on the autonomous guidance of Micro-Air Vehicles (MAVs) in confined indoor and outdoor environments, we have developed a vision-based autopilot, which is called LORA III (Lateral Optic flow Regulation Autopilot, Mark III). This autopilot, based on the dual optic flow regulation, allows an air vehicle to travel along a corridor by automatically controlling both its speed and its clearance from the walls (Serres et al. 2008a). The optic flow regulation is a feedback control based on an optic flow sensor, which strives to maintain a perceived optic flow at a constant set-point by adjusting a thrust (Ruffier and Franceschini 2005). The LORA III autopilot consists of a dual optic flow regulator in which each regulator has its own optic flow set-point and controls the robot’s translation in one degree of freedom: a bilateral optic flow regulator controls the robot's forward speed, while a unilateral optic flow regulator controls the side thrust, making the robot avoid the walls of the corridor (Serres et al. 2008b). This autopilot draws on former studies which aimed to understand how a honeybee might be able to center along a corridor (Srinivasan et al. 1991), to follow a single wall (Serres et al., 2008b), and to adjust its speed according to the corridor width (Srinivasan et al. 1996).