Reverse Engineering of Insect Flight Control

Rafał Żbikowski

Cranfield University (Defence Academy Shrivenham)

Great Britain

Micro air vehicles (MAVs) are small flying vehicles developed to reconnoitre in confined spaces. This requires power-efficient, highly-manoeuvrable, low-speed flight with stable hover. All of these attributes are present in insect flight and hence the focus on reproducing the functionality of insect flight by engineering means. One key aspect is the excellent manoeuvrability of insects which calls for examination of their flight control. Both modern aircraft and flying insects exhibit complex flight dynamics, whose handling requires generating much information in real time. In conventional control little measurement is made, so that the required information is recovered by involved calculations. I believe that insects do the converse: they measure much and compute little.

This sensor-rich feedback control is a new approach to tackle complexity of animal/vehicle motion, a particularly challenging problem in the context of manoeuvrable flight. Insect vision seems to be the key sensor system, as not only does it recognise patterns (objects), but can perceive motion. The compound eyes, composed of up to 6,000 ommatidia, are a rich sensor system, each ommatidium measuring light intensities within a small solid angle, but with a high temporal resolution. Neural processing of the compound eyes produces a global, detailed view of the relative motion of the insect with respect to its surroundings.

This sensor-rich feedback control paradigm relies on extensive, distributed measurement of the quantities of interest in space and time. In insects motion representation covers the whole of space, and involves several overlapping patches of the space. It seems that the complex underlying differential equations of motion need not be integrated numerically, as their solutions are readily available through interpolation of the detailed measurements. Also, this redundant representation may allow fine control of high agility manoeuvres with little computation.

This new paradigm will have a large impact not only on the vast applications of automatic control, but also on the sensor, instrumentation, and measurement communities. Further, detailed and multidisciplinary research is needed to reverse engineer insect flight control completely.

Dr Rafał Żbikowski is Reader in Control Engineering at the Department of Aerospace, Power & Sensors at Cranfield University (Defence Academy Shrivenham), England. He has published over 60 papers on adaptive and nonlinear intelligent control, flapping wing micro air vehicles, missile guidance and control, and co-operative control of swarms of vehicles. A control engineer with a strong mathematical background, he initiated the insect-like flapping wing micro air vehicles research in the UK and has led the effort since 1998. He is a Visiting Professor (sponsored by the US Air Force) at the University of Florida and reviews for several scientific journals, including Nature.

Further information can be found at: http://www.dcmt.cranfield.ac.uk/daps/guidance/microairvehicles/view