В.Н. Коновалов, М.О. Корлякова
14
Approach to developing control systems of micro spacecraft
in neuronetwork basis
© V.N. Konovalov, M.O. Korlyakova
Kaluga Branch of Bauman Moscow State Technical University, Kaluga, 248000, Russia
The article reviews the main directions in hardware implementation of neural networks
and examines their opportunities for using onboard micro spacecraft. The main require-
ments to onboard control systems and approaches to their neuronetwork realization are
considered. The findings of this research illustrate restrictions and requirements to the
resources necessary for onboard systems of micro spacecraft. We show that it is possible
to find the effective solution in simple networks. We modeled separate blocks of a neu-
ronetwork control system in the conditions of limited resources and presented the neu-
ronetwork model of approximating the tabular model of the atmosphere. We provide the
sample task solution in float and discrete formats of synapse. The article delves into ap-
proximation errors in networks of different architecture and shows, that even for options
with the discrete form of inputs and a short format of representing the synapses, the task
solution quality is sufficient.
Keywords:
neuronetworks, micro spacecraft, control system
REFERENCES
Proletarsky A.V.
Nauka i obrazovanie. Elektronnoe nauchno-tekhnicheskoe
izdanie — Science and education. Electronic scientific and technical
periodical
, 2007, no. 3. Available at:
(accessed 12 December 2013).
Efimov V.V.
Mekhatronika, avtomatizatsiya, upravlenie
— Mechatronics,
Automation, Control
, 2003, no. 5, pp. 26–32.
Nikishov A.N., Oleinik A.A., Talimanchuk L.L., Zaitsev A.V., Sukhanov N.V.
Neirokompyutery — Neurocomputers
, 2010, no. 1, pp. 34–39.
Khaikin S
. Neironnye seti: polnyi kurs
[Neural networks: a complete course].
2nd ed., transl. from Engl. Moscow, «Izd. dom Vil'iams, Ltd » Publ., 2006,
1104 p.
Ermolenko A.V., Korlyakova M.O. Primenenie virtual'nykh modelei ob"ektov
dlia obucheniia neirosetei. [Applying virtual models of objects for training
neural networks].
12-ya Nats. konf. po iskusstvennomu intellektu s
mezhdunarodnym uchastiem KII–2010, Tver', Rossiya, 2010: Trudy
konferentsii
. [Twelfth National Conference on Artificial Intelligence with CAI
2010, Tver, Russia, 2010: Proceedings]. Moscow, Fizmatlit Publ., 2010, vol. 2,
pp. 107–115.
Karimov R.R., Kositsyn V.V., Ruban V.I., Blinov V.N., Valeev S.S.,
Islamgulov T.V. Sistema informatsionnoi podderzhki zhiznennogo tsikla
malogo nauchno-obrazovatel'nogo kosmicheskogo apparata [System of
informational life cycle support of small scientific and educational spacecraft].
Sbornik trudov 11-i mezhdunar. konf. CSIT
[Proceedings of the 11th Int. conf.
CSIT], 2009, vol. 3, pp. 22–28.