In addition to aero, I do neuro, gero and euro.
First, the cover. Electrification of training aircraft seems financially feasible. Benefits are operational and environmental, much less maintenance, much less fuel consumption, little noise and no carbon footprint. With additional design features such as distributed propulsion and easier control. We are currently designing a case study to see whether the theoretical transpires. Read more here.
Under the hood. I have had some experience about nanotechnology, and even managed to publish a few interesting research, as part of a team, during my postdoc. Then, I turned away, selling all the stock at their peak in 2011. Huh? Indeed. Because, I believe nano is not sustainable as we know it today. Nano is the clean room and bio, the only sustainable nano, is the bedroom - not the epitome of precision and calculation. The bedroom is so efficient that you must really take necessary precaution lest you really need the manufacture. How so? Bio has a few tricks on her sleeve, but her greatest is that it does not invest in super precise nano-structures, instead churns out a fairly large number of below par gadgets, and, capitalising on the fantastic calculus of multidimensional algebra(*) it reaches much further. Just as the sole point of a graphic is information, the point of a sensing unit is the information fed-forward. Biology does not manufacture to brag(**). In this line, I argue that a 'bedroom' equivalent of a Bragg mirror is as good, even better, a sensor for detecting airborne organic/inorganic molecules. In another show case, I argue that neuromorphic computation rescues an otherwise hopeless photonic nose. And currently I devise a method to use the humble FTIR system as an electronic nose by leveraging a neural backend. I skimmed through these ideas in the Odyssey from nano to neuro talk given in 2018.
(*) a friend has recently made a good deal of this, read it here.
(**) see the humble nose for a case study, here.
We are repurposing an FTIR spectrometer as a fully implemented electronic nose with sampling, measurement and data evaluation. The infrared spectra is considered a gestalt chemical snapshot of an odorous agent, or rather, a continuous, actively sampled dynamic snapshots of the close environment. We read-in the infrared spectra considering it as the response array of pseudo-sensors, assuming each wavenumber data itself is a sensor reading from a broadly tuned non-specific sensor . We are considering the full pipeline, including a possible active sniff cycle, possibly chemotaxis as well as an ultimate behavioral response in response to the presence of a chemical agent. We call this technology deep spectroscopy or neuromorphic spectroscopy too.
Development of the essential supernose technology
R. Garifullin, Özlem İnci, Asma Djehiche, Sude Çetin, Kaan Gök, Ozan Aslan, Hande Uyan, Süleyman Kınıkoğlu, Ekinsu Bülbül
Ozan Aslan, Hande Uyan, Ekinsu Bülbül The promise of attention based neural network architectures for electronic nose systems
Active sniffing
Mert Kızılkaya, Burak Kaya, Sude Çetin Active sniffing using the FTIR spectrometer
MOX Sensors Fatih Kuzucuk, Haticenur Çetin, Süleyman Kınıkoğlu
Chemotaxis (crawling and airborne)
...
We are developing drones with cognitive control capabilities.
Sniffing drones is about adding a chemosensory modality to drones, and use them for target localization. This is a natural extension of the supernose project.
Thermal soaring is about implementation of a reinforcement learning policy for use in a glider.
Muzaffer Özmen, Yağmur Aksel, Mustafa Şentürk
...