Radiotekhnika
Publishing house Radiotekhnika

"Publishing house Radiotekhnika":
scientific and technical literature.
Books and journals of publishing houses: IPRZHR, RS-PRESS, SCIENCE-PRESS


Тел.: +7 (495) 625-9241

 

Feature of multi-agent systems scaling on an example of centralized and decentralized group management of small-sized mobile robots

Keywords:

R.A. Prokopovich - Ph.D. (Eng.), Associate Professor, Heading Research Scientist, E-mail: rprakapovich@robotics.by


The purpose of the article is to determine the effectiveness of the robots group under different control methods. The research object is decentralized and centralized control methods of the group of the mobile robots based on the methodology of multi-agent systems (MAS). The algorithms mentioned have been developed earlier by the author in previous articles. A decentralized algorithm is based on the analogy of foraging ants, who while moving mark their path with pheromones. The proposed decentralized algorithm is stochastic. A centralized control algorithm is based on the well-known algorithm of the path search in the maze – the wave algorithm. This algorithm works on the principle of the width graph traversal. The geometric parameters and the number of obstacles were not changed during the experiments. For obtaining more feasible results the location of the obstacles and the target objects were selected randomly. The number of the robots in the group was increased from 1 to 100 in each method. The results of the experiments are illustrated by the performance function graphs of each of the selected methods of the group control. For the centralized control method, which has not been imposed with any restrictions (the ideal case), there is a non-linear increase in performance. The second graph (the centralized method with restrictions) illustrates a significant influence of the threshold value responsible for the largest possible number of simultaneously working robots. With the decentralized control method, it is noticeable that with an increase in the number of the group members the effectiveness enhances almost linearly. Due to the fact that under the centralized method the computation centre has global information about all the inves-tigated areas of the area map and the location of the robots, the group operates more efficiently than under the decentralized control. However, when a certain number of the group members is reached under the centralized control, their performance decreases comparing to the decentralized control method. It was also observed that the performance of the decentralized control algorithms almost always does not depend on external conditions and is more stable and predictable.
References:

 

  1. Kaljaev I.A., Gajjduk A.R., Kapustjan S.G. Raspredeljonnye sistemy planirovanija dejjstvijj kollektivov robotov. M.: JAnus-K. 2002. 292 s.
  2. Tarasov V.B. Ot mnogoagentnykh sistem k intellektualnym organizacijam: filosofija, psikhologija, informatika. M.: EHditorial URSS. 2002. 352 s.
  3. Kaljaev I.A., Gajjduk A.R., Kapustjan S.G. Modeli i algoritmy kollektivnogo upravlenija v gruppakh robotov. M.: FIZMATLIT. 2009. 280 s.
  4. Dorigo M., Maniezzo V., Colorni A. The Ant System: Optimization by a colony of cooperating agents // IEEE Transactions on Systems, Man and Cybernetics. 1996. Part B. V. 26. P. 29–41.
  5. Adamackijj A.I., KHolland O. Rojashhijjsja intellekt: predstavlenija i algoritmy // Informacionnye tekhnologii i vychislitelnye sistemy. 1998. № 1. S. 45–53.
  6. Prokopovich G.A. Ob odnom metode decentralizovannogo upravlenija gruppojj robotov // Robototekhnika. Vzgljad v budushhee: materialy Mezhdunar. sem. SPb.: Politekhnika-servis. 2010. S. 199–202.
  7. Prokopovich G.A., Sychjov V.A. Modelirovanie kollektivnogo povedenija robotov dlja poiskovo-issledovatelskikh zadach // EHkstremalnaja robototekhnika: mat. Mezhdunar. nauch.-tekhn. konf. SPb.: Politekhnika-servis. 2010. S. 237–243.
  8. Prokopovich G.A., Sychjov V.A. Primenenie geneticheskikh algoritmov dlja reshenija zadachi transportirovki gruza gruppojj robotov // Mekhatronika i robototekhnika: sb. dokl. Mezhdunar. molodjozhnojj konf. SPb.: Politekhnika-servis. 2011. S. 76–80.
  9. Sychjov V.A., Prokopovich G.A. EHksperimentalnaja model gruppovogo robota dlja reshenija zadach monitoringa pomeshhenijj // Innovacionnye tekhnologii, avtomatizacija i mekhatronika v mashino- i priborostroenii // Materialy Mezhdunar. nauch.-prakt. konf. Minsk: Biznesofset. 2013. S. 166–167.
  10. Prokopovich G.A. Algoritm prostranstvennogo orientirovanija mobilnogo robota v neznakomojj mestnosti // Sb. materialov Mezhdunar. foruma «Pervyjj shag v nauku – 2007». Minsk: «Izd. dom «Belorusskaja nauka». C. 428–432.
  11. Sychjov V.A., Prokopovich G.A., Gerasjuto S.L. Programmno-apparatnoe obespechenie uchebno-metodicheskogo kompleksa po robototekhnike // EHlektronika-Info. 2014. № 6 (108). S. 26–29.
  12. Sychjov V.A., Gerasjuto S.L. Razrabotka protokola i programmno-apparatnogo obespechenija sistemy centralizovannogo upravlenija gruppojj robotov // Avtomatizacija i robotizacija processov i proizvodstv: mat. resp. nauch.-prakt. sem. Minsk: Biznesofset. 2014. S. 101–102.
  13. Prokopovich G.A., Sychjov V.A., Gerasjuto S.L. Issledovanie osobennostejj masshtabirovanija chislennosti gruppy malogabaritnykh mobilnykh robotov pri centralizovannom i decentralizovannom sposobakh upravlenija // Materialy mezhdunar. nauch.-tekhn. konf. «EHkstremalnaja robototekhnika». SPb.: CNII RTK. 2015. C. 76–81.

 

Sept. 2, 2020
Aug. 27, 2020
June 24, 2020

© Издательство «РАДИОТЕХНИКА», 2004-2017            Тел.: (495) 625-9241                   Designed by [SWAP]Studio