|Article title||COMPUTATIONAL METHODS OF SYNTHESIS OF SYNERGETIC CONTROL FOR ROBOTS GROUP INTERACTION|
|Authors||A. I. Diveev, E. Yu. Shmalko|
|Section||SECTION I. GROUP CONTROL ROBOTS|
|Month, Year||09, 2017 @en|
|Index UDC||004.896, 519.688, 004.023|
|Abstract||The paper presents the results of our research in application of modern numerical methods of symbolic regression for the automatic synthesis of synergistic control for robots group interac-tion. Synergetic control is determined by the presence of such manifolds in the state space that must have the properties of an attractor, in particular terminal manifolds, as well as domains through which solutions of the system do not pass or have repeller properties, for example, the domain of phase constraints. In this paper we formulated a multipoint numerical criterion for ensuring the properties of an attractor for manifolds in the state space. The presented criterion takes into account not only the hit of the solution to the neighborhood of the manifold, but also the behavior of the solution in the neighborhood of the manifold. The proposed criterion turns out to be especially important for one-dimensional manifolds, for which it is necessary not only to get into the neighborhood of a manifold, but also to provide a definite form of variation of the solution in time in the neighborhood of the manifold. The paper presents the results of computational experiments on the effective application of the proposed criterion. As an application object, we consider the problem of synthesizing the control for group of robots. Additional complexity of the task of controlling a group of robots is the presence of dynamic phase constraints that determine the absence of collisions between robots. These dynamic constraints seriously complicate the process of finding the optimal solution for both the control synthesis and optimal control problems. The complexity of the search is mainly dictated by changing the convexity properties of the search area. The paper shows by computational experiments that the solution of the synthesis problem and the problem of optimal control is most effectively performed for control objects that have the property of local stability with respect to a given point in the state space. Therefore, when solving these problems, the proposed two-stage approach is applied when firstly the problem of synthesis of the stabilization system is to be solved in advance to ensure the stability of the object relative to a given point in the state space, and secondly problems of synthesis and optimal control are solved as problems of controlling the stabilization point in the state space. To synthesize control, we use the methods of symbolic regression, which allow us to find the coded mathematical expression of the required synthesizing function using evolutionary algorithms. When searching for optimal trajectories we use various evolutionary and gradient algorithms of nonlinear programming.|
|Keywords||Control of group of robots; synergetic control; symbolic regression methods; attractors.|
|References||1. Kolesnikov A.A. Sinergeticheskie metody upravleniya slozhnymi sistemami: teoriya sistemnogo sinteza [Synergetic control methods of complex systems: theory of system synthesis]. Moscow: Editorial URSS, 2005, 228 p.
2. Kolesnikov A.A. Sinergeticheskaya teoriya upravleniya [Synergetic control theory]. Taganrog: TRTU, M.: Energoatomizdat, 1994, 344 p. ISBN:5-230-24678-2.
3. Kolesnikov A.A. Posledovatel'naya optimizatsiya nelineynykh agregirovannykh sistem upravleniya [Sequential optimization of nonlinear aggregated control systems]. Moscow: Energoatomizdat, 1987, 160 p.
4. Diveev A.I. Chislennyy metod setevogo operatora dlya sinteza sistemy upravleniya s neopredelennymi nachal'nymi znacheniyami [A numerical method for network operator for syn-thesis of a control system with uncertain initial values], Izvestiya RAN Teoriya i sistemy upravleniya [Journal of Computer and Systems Sciences International], 2012, No. 2, pp. 63-78.
5. Diveev A.I. Priblizhennye metody resheniya zadachi sinteza optimal'nogo upravleniya [Approximate methods of solution of synthesis problem of optimal control]. Moscow: Izd-vo VTs RAN. 2015, 184 p.
6. Diveev A.I., Shmal'ko E.Yu. Evolyutsionnye metody vychisleniy dlya sinteza upravleniya gruppoy robotov i poiska optimal'nykh traektoriy ikh dvizheniya [Evolutionary computing methods for control synthesis a group of robots and the search for optimal trajectories of their movemen], Cloud of Science, 2017, Vol. 4, No. 3, pp. 395-414.
7. Diveev A.I., ShmalkoE.Yu. Complete Binary Variational Analytic Programming for Synthesis of Control at Dynamic Constraints, ITM Web Conf. Vol. 10, 2017 Seminar on Systems Analysis, 10. 1051/itmconf/20171002004.
8. Diveev A.I., Shmal'ko E.Yu. Chislennyy sintez sistemy upravleniya gruppoy robotov metodom simvol'noy regressii [Numerical synthesis of control systems by group of robots with the method of symbolic regression], Izvestiya YuFU. Tekhnicheskie nauki [Izvestiya SFedU. Engi-neering Sciences], 2015, No. 10 (171), pp. 29-45.
9. Diveev A.I., Shmal'ko E.Yu. Dvukhetapnyy sintez sistemy upravleniya metodom setevogo operatora [Two-step synthesis of a control system by the network operator method], Vestnik RUDN. Seriya Inzhenernye issledovaniya [Bulletin of the Russian University of friendship of peoples. Series: Engineering research], 2015, No. 1, pp. 91-99.
10. Diveev A.I., ShmalkoE.Yu. Automatic Synthesis of Control for Multi-Agent Systems with Dy-namic Constraints, Preprints, 1st IFAC Conference on Modelling, Identification and Control of Nonlinear Systems June 24-26, 2015. Saint Petersburg, Russia, pp. 394-399.
11. Diveev A.I., ShmalkoE.Yu. Self-adjusting Control for Multi Robot Team by the Network Operator Method, Proceedings of the 2015 European Control Conference (ECC) July 15-17, 2015. Linz, Austria, pp. 709-714.
12. Diveev A.I., ShmalkoE.Yu. Optimal Control Synthesis for Group of Robots by Multilayer Net-work Operator, Proceedings of 3rd International Conference on Control, Decision and Infor-mation Technologies (CoDIT’16). St. Paul’s Bay – Malta on April 6-8, 2016, pp. 077-082.
13. Evtushenko Yu.G. Optimizatsiya i bystroe avtomaticheskoe differentsirovanie [Optimization and fast automatic differentiation]. Moscow: VTs RAN, 2013, 144 p.
14. Diveev A.I., Konstantinov S.V. Evolyutsionnye algoritmy dlya resheniya zadachi optimal'nogo upravleniya [Evolutionary algorithms for solving optimal control problem], Vestnik RUDN. Seriya Inzhenernye issledovaniya [Bulletin of the Russian University of friendship of peoples. Series: Engineering research], 2017, Vol. 18, No. 2, pp. 254-265.
15. Karpenko A.P. Sovremennye algoritmy poiskovoy optimizatsii. Algoritmy, vdokhnovlennye prirodoy [Modern algorithms of search engine optimization. Algorithms inspired by natur]. Moscow: Izd-vo MGTU im. N.E. Baumana, 2014, 448 p.
16. Goldberg, D.E. Genetic Algorithms in Search, Optimization, and Machine Learning. Addison-Wesley. 1989.
17. Kennedy J., Eberhart R. Particle Swarm Optimization, Proceedings of IEEE International Conference on Neural Networks IV, 1995, p. 1942-1948.
18. Mirjalili S.A., Mirjalili S.M., Lewis A. Grey Wolf Optimizer, Advances in Engineering Software, 2014, Vol. 69, pp. 46-61.
19. Pham D.T. et al. The Bees Algorithm – A Novel Tool for Complex Optimisation Problems, Intelligent Production Machines and Systems. 2nd I*PROMS Virtual International Conference 3-14 July 2006. Elsevier Ltd, 2006, pp. 454-459.
20. Panteleev A.V., Letova T.A. Metody optimizatsii v primerakh i zadachakh: ucheb. posobie [Optimization methods in examples and tasks: textbook]. Moscow: Vysshaya shkola, 2005, 544 p.