Flying in 3D with an Insect based Visual Autopilot

• Portelli G.
• Serres Julien
• Ruffier F.
• Franceschini N.

• Optic flow OF
• Elementary Motion Detector
• Honeybee
• Navigation
• Insect autopilot
• Biomimetics
• Biorobotics

Flying insects rely on Optic Flow (OF) cues to avoid collisions, control their speed, control their height, and land. Recent studies have shown that the principle of “OF regulation” may account for various behaviors observed in freely flying insects. The aim of the present study was to suggest a visually guided autopilot enabling an insect to navigate in 3D, and to test its robustness to natural images. Using computer-simulation experiments, we simulated a bee that flies through a tunnel wallpapered with natural images, by controlling both its ground speed and clearance all four sides: the lateral walls, the ground, and the ceiling. The simulated bee can translate along three directions (the surge, sway, and heave axes): it is therefore fully actuated. The new visuo-motor control system, called ALIS (AutopiLot using an Insect based vision System), is a dual OF regulator consisting of two interdependent feedback loops: the speed control loop (along the surge axis) and the positioning control loop (along both the sway and heave axes), each of which has its own OF set-point. The experiments show that the simulated bee navigates safely along a straight tunnel, while compensating for the major OF perturbations caused by, e.g., a tapering of the tunnel or the lack of texture on one wall. The minimalistic visual system used here (only eight pixels) is robust to naturally contrasted stimuli and tunnels, and is sufficient to control both the clearance from the four sides and the forward speed jointly, without requiring to measure any speeds or distances. Besides, the ALIS autopilot accounts remarkably for the quantitative results of ethological experiments performed on honeybees flying freely in straight or tapered corridors.