Herr Dr. Hanno Gerd Meyer

Flying insects like for example flies or bees heavily rely on their visual system when performing behavioural tasks like course control, collision avoidance or spatial navigation. Compared to current approaches in machine vision the visual systems of flying insects are - despite the relatively small number of neurons involved - superior in terms of speed, flexibility and resource efficency, though they only have a poor spatial resolution. Hence, from an engineer‘s perspective the underlying mechanisms that allow such tiny animals to interact with their environment are of great interest.

Our goal is to develop a bio-inspired camera system allowing robotic platforms to perform visually-guided orientation behaviour based on physiological and computational principals found in insects. In cooperation with the Cognitronics group at CITEC, we develop and implement algorithms on dynamically reconfigurable hardware in combination with low power microprocessors to extract motion information contained in the optic flow of images a robotic agent perceives during self-motion. This visual motion information is subsequently used in bio-inspired algorithms controlling the agent's orientation behaviour in unfamiliar terrain.