IPAB Workshop - 21/07/2022

Title:  Insect-inspired control for aerial robot navigation

Abstract:  Insects have long been a source of inspiration for flying robot control, and offer the possibility of reliable, lightweight autonomous navigation solutions. Insects such as bees are known to be able to navigate between sites of interest in large local neighbourhoods that span several kilometres, using relatively low-resolution visual input and computational processing performed in a relatively small brain. We describe several hardware and software developments resulting in an aerial biorobot that we use to test insect navigation models, featuring a quadcopter airframe, Pixhawk flight controller and an active mechanical view stabilisation system. This biorobot is first used to embody a recently proposed anatomically constrained path integration circuit. It uses a biologically plausible matched visual filter to obtain ground speed during holonomic flight, and combines this with directional information in the neural circuit to maintain an estimate of its starting position. We show this real-world system, tested outdoors, has a homing error drift rate of only 1.5m per 100m flown. A new design for a polarisation-sensing system, based on the receptor layout of the insect eye, has also been implemented and can be used to obtain an accurate 360 degree estimate of heading. We have also used the biorobot to investigate whether flying insects could use visual route following to overcome the drift issues associated with path integration. We propose a new approach using downward views and a similarity algorithm based on oriented bandpass filters that plausibly represent processing in the insect visual system. We show this method can be effective to zero in on the target location across distances of at least 30m, even in seemingly featureless environments such as empty arable fields.