Periodic Reporting for period 1 - COgITOR (A new COlloidal cybernetIc sysTem tOwaRds 2030)
Berichtszeitraum: 2021-06-01 bis 2022-11-30
COgITOR project aims at developing a liquid state cybernetic system prototype, capable of sensing the external environment, of performing simple calculations on the data received by its liquid sensors, and of generating some energy for its own operation. No other autonomous systems exist, made out of liquid state materials. Why is this important? Because a liquid can easily withstand extreme conditions, such as high temperatures, pressures, magnetic fields, and operate where other common robotic technologies fail. Furthermore, the new frontier of liquid electronics offers unexplored domains, new exciting discoveries, and potential applications for the society. The overall objectives are: i) realizing a liquid state pressure sensor that will be placed on the outer skin of the prototype; ii) realizing a liquid state rewritable memory; iii) realizing a liquid state energy harvester to power the system; iv) integrate all of these parts in a single spherical system.
During the first 18 months since the beginning of the project, the consortium has achieved the following progresses: 1) a measurement system to assess the capability of liquids to respond to electrical stimulation, with a broad band resolution, that already allowed us to collect 10 GB of raw data on several liquids under test. Based on the first results, we have started developing a dedicated integrated circuit for measuring the colloids state changes and implement in a single chip all of those functionalities. The chip is currently under final design stages in a low-cost technology. 2) We have discovered a protocol for implementation of learning in liquid systems via a programmable propagation of conductive pathways. Also, we obtained fundamental preliminary results on implementation of reservoir computing in colloid systems, achieving very nice demonstration of in memory computing, of liquid analogue memory equivalent to 6 bits, of liquid state artificial recurrent neural network. The liquid was shown able to recognize 10 digits, from zero to nine. Such results represent a unique achievement in science. 3) Alternative energy supply in the soft cybernetic system was explored with the idea of inducing the conversion of light into heat directly in the liquid phase, avoiding solid interfaces, and to exploit with a mixed composition the resulting thermal gradient into electrical power. 4) Several magnetic and infrared absorbing nanomaterials on the base of Iron(II, III) oxides and Titanium Nitrides as well as Magnetite with different susceptibility and pyromagnetic effects were developed and tested. These technologies and materials will lay the basement for computing subsystems used in later tasks. 5) Self-healing soft skins for protecting the functional colloids were developed.
We can already claim outstanding progresses beyond the state of the art, namely: a liquid state analogue memory equivalent to (at least) 6 bits, a liquid state artificial recurrent neural network that can recognize digits, a liquid state physical reservoir computer, the protocol for implementing learning in a liquid, a liquid state photothermal conversion towards thermal harvesting. The liquid state in-memory-computing system has been patented, and this represents a clear and strong impact at socio-economic level. We expect to file soon another patent, about the microwave impedance spectroscopy holonomic chip (the integrated system we have called Dandelium) that together with the previous patent will create enough value for establishing a spin-off company and transform this radical innovation into economical value. Besides, the formulation of protocols for the synthesis of new nanomaterials by our partner PLASMACHEM has already impacted the economical value of COgITOR. Liquid electronics has become reality and we expect by the end of the project to strengthen its impacts, exploring in particular holonomic liquid state computers and extreme environment liquid cybernetic systems.