Thanks to the ecological, physiological, neurological and engineering knowledge obtained during this project, we will be in a position to advise system engineers dealing with chains of ‘sensing- perception-action’ on the degree of ‘success’ needed at each stage. We do not know may such examples of complex technological systems for which each component has a given and predetermined contribution to the fitness of the whole structure.

The highly innovative use of technology such as Laser Doppler vibrometry, micromechanical testing in MEMS, modeling using Finite Element Methods and model testing using high speed video technology and flow visualization technique will enable us to carry a truly comprehensive analysis of air velocity sensing hairs at a new scale. In particular, we should be able to test some experimental data and analytical predictions obtained by previous authors who usually tackled the problem considering a single or a few hairs only. This step will also enable us to extend the analysis to the scale of an array of hairs and provide experimental evidence for the need (or not) of a large number of sensors. This is new ground.

We expect major advances in the construction of arrays of MEMS-based sensors (thresholds, cross-sensitivity, array topologies) and their interface with ‘wetware’. For example, we have few ideas, as yet, on how to build 3D MEMS and this project may provide good analogies. We also have little information as of the degree of dissimilarity between sensors (usually regarded as a nuisance) to obtain a good signal representation. Our project may provide useful hints in this respect, as cricket hairs are dissimilar on purpose.

Information preprocessing localised as near as possible to the receptors is at the heart of the very quick escape reactions of crickets. Such peripheral information preprocessing is a process which we will implement in the demonstrator. While robot crickets have already been made in the last few years, none of them had MEMS based sensing technology. Furthermore, none of them was of hybrid type. The number of sensing units was at least an order of magnitude smaller than we intend to have. Given the actual problems with systems based on many sensors, it is unlikely that the known robot crickets could be scaled up to carry more complex tasks, such as recognition of specific wind signature and appropriate decisions. In order to do such tasks, hybrid systems based on energy-efficient MEMS technology seems a visionary long term goal our project can approach.

While miniaturization leads to many benefits for society, in particular for a friendly Information Technology Society, the advances proposed in this project could also be used in other fields, less directly to human beings. For example, arrays of MEMS sensors could be used in agronomy and plant sciences to monitor air flow in canopies, at the level of leaves and branches, and adapt irrigation levels or identify potential water losses due to transpiration. These are very serious challenges, given the scale of global change one has to expect. Other applications include air flow pattern observation in places which are small or difficult to reach.