ANALYSIS AND DATA PROCESSING SYSTEMS

Abstract
The development of recurrent methods in the estimation theory is caused by the requirements of modern production where it is necessary to introduce complex control systems. The complexity of such systems is characterized, first of all, by the need to work under conditions of a priori uncertainty about external environment properties in the operation where human control is complicated. Many search algorithms of unknown parameters of the control object are based on the recurrent approach to an extremum of some chosen quality criterion. One of the most widespread recurrent estimation methods is the recursive least-squares method. In this work the estimation of a dynamic object parameter based on input system and output system signals is executed in the Matlab package. The parameter recurrent estimation procedure is considered for the case when there are no measurements noises, and also for the case when Gaussian noises are supposed to be present. In contrast to passive identification methods a possibility of applying the required input signal to an object is supposed in this work. An input signal is a meander- like signal. The recurrent estimation procedure in the MatLab package is presented by three levels. The top level blocks show the modeling results of the dynamic object, the input signal and measurement noises. Blocks of the following levels form the parameter recurrent estimation procedure. The object parameter estimation results are plotted for the case when measurement noises are absent and also for the case when they are present. In addition, the results of the behavior of the amplification coefficient which is used to estimate the parameter are given. It should be noted that recurrent synthesis in which object parameters are adjusted on receiving results of new measurements is widely used in the adaptive filter design that makes it possible to process large data amounts timely.  

References
1. Åstrŏm K.J. Introduction to stochastic control theory. New York, London, Academic Press, 1970. 298 p. (Russ. ed.: Ostrem K. Vvedenie v stokhasticheskuyu teoriyu upravleniya. Translated from English. Moscow, Mir Publ., 1973. 320 p.).

2. Ljung L. System identification: theory for the user. New Jersey, Prentice Hall, 1987. 384 p. (Russ. ed.: L'yung L. Identifikatsiya sistem. Teoriya dlya pol'zovatelya. Translated from English. Moscow, Nauka Publ., 1991. 432 p.).

3. Eykhoff P. System identification: parameter and state estimation. London, John Wiley & Sons, 1974. 555 p. (Russ. ed.: Eikkhoff P. Osnovy identifikatsii sistem upravleniya: otsenivanie parametrov i sostoyaniya. Translated from English. Moscow, Mir Publ., 1975. 680 p.).

4. Meditch J.S. Stochastic optimal linear estimation and control. New York, McGraw-Hill, 1969. 384 p. (Russ. ed.: Medich Dzh. Statisticheski optimal'nye lineinye otsenki i upravlenie. Translated from English. Moscow, Energiya Publ., 1973. 440 p.).

5. Sage A.P., Melse J.L. Estimation theory with application to communication and control. New York, McGraw-Hill, 1972. 540 p. (Russ. ed.: Seidzh E.P., Melsa Dzh. Teoriya otsenivaniya i ee primenenie v svyazi i upravlenii. Translated from English. Moscow, Svyaz' Publ., 1976. 495 p.).

6. Gupta H.K., Mehra R.K. Computational aspects of maximum likelihood estimation and reduction in sensitivity function calculation. IEEE Transactions on Automatic Control, 1974, vol. 19, no. 7, pp. 774–785.

7. Åstrŏm K.J. Maximum likelihood and prediction error methods. Automatica, 1980, vol. 16, no. 5, pp. 551–574.

8. Mehra R.K. Optimal input signal for parameter estimation in dynamic system – survey and new results. IEEE Transactions on Automatic Control, 1974, vol. AC-19, no. 6, pp. 753–768.

9. Mehra R.K. On the identification of variences and adaptive Kalman filtering. IEEE Transactions on Automatic Control, 1970, vol. AC-15, no. 2, pp. 175–184.

10. Mehra R.K. Optimal input for linear system identification. IEEE Transactions on Automatic Control, 1974, vol. 19, no. 3, pp. 192–200.

11. Goodwin G.C., Payne R.L. Dynamic system identification: experiment design and data analysis. New York, Academic Press, 1977. 291 p.

12. Antsaklis P.J., Michel A.N. Linear systems. New York, McGraw-Hill, 1997. 685 p.

13. Brown R.J., Sage A.P. Error analysis of modeling and bias errorsin continuous time state estimation. Automatica, 1971, vol. 7, pp. 577–590.

14. Goodwin G.C. Optimal input signals for nonlinear-system identification. Proceedings of the Institution of Electrical Engineers, 1971, vol. 118, pp. 922–926.

15. Sage A.P., White C.C. Optimum system control. 2^nd ed. Englewood Cliffs, NJ, Prentice-Hall, 1977 (Russ. ed.: Seidzh E.P., Uait Ch.S. Optimal'noe upravlenie sistemami. Moscow, Radio i sviaz' Publ., 1982. 392 p.

16. Voskoboinikov Yu.E. Kriterii raskhodimosti i algoritm adaptatsii rekurrentnogo algoritma otsenivaniya vektora sostoyaniya [The divergence criterion and the adaptation algorithm of the recurrent algorithm of the state vector estimation]. Nauchnyi vestnik Novosibirskogo gosudarstvennogo tekhnicheskogo universiteta – Science bulletin of the Novosibirsk state technical university, 2015, no. 3 (60), pp. 7–22. doi: 10.17212/1814-1196-2015-3-7-22

17. Denisov V.I., Chubich V.M., Chernikova O.S. Aktivnaya parametricheskaya identifikatsiya gaussovskikh lineinykh nepreryvno-diskretnykh sistem na osnove planirovaniya vkhodnykh signalov i nachal'nykh uslovii [Active parametric identification of Gaussian linear continuous-discrete systems based on designing input signals and initial conditions]. Nauchnyi vestnik Novosibirskogo gosudarstvennogo tekhnicheskogo universiteta – Science bulletin of the Novosibirsk state technical university, 2014, no. 4 (57), pp. 19–30. doi: 10.17212/1814-1196-2014-4-19-30

18. Chubich V.M., Chernikova O.S. Osobennosti zadachi planirovaniya eksperimenta dlya gaussovskikh lineinykh sistem [Peculiarities of the experiment design problem for Gaussian linear systems]. Nauchnyi vestnik Novosibirskogo gosudarstvennogo tekhnicheskogo universiteta – Science bulletin of the Novosibirsk state technical university, 2015, no. 3 (60), pp. 178–191. doi: 10.17212/1814-1196-2015-3-178-191

19. Bolovin E.V. Kriticheskii ekspertnyi analiz metodov identifikatsii parametrov asinkhronnykh dvigatelei [Critical expert analysis of the induction motor parameters identification methods]. Nauchnyi vestnik Novosibirskogo gosudarstvennogo tekhnicheskogo universiteta – Science bulletin of the Novosibirsk state technical university, 2015, no. 1 (58), pp. 7–27. doi: 10.17212/1814-1196-2015-1-7-27

20. Voevoda A.A., Troshina G.V. Otsenivanie parametrov modelei dinamiki i nablyudeniya dlya lineinykh statsionarnykh diskretnykh sistem s ispol'zovaniem informatsionnoi matritsy Fishera [Parameters estimation of dynamics and supervision models for linear stationary discrete systems with use of Fischer information matrix]. Nauchnyi vestnik Novosibirskogo gosudarstvennogo tekhnicheskogo universiteta – Science bulletin of the Novosibirsk state technical university, 2006, no. 3 (24), pp. 199–200.

21. Troshina G.V. Aktivnaya identifikatsiya lineinykh dinamicheskikh diskretnykh statsionarnykh ob"ektov vo vremennoi oblasti. Diss. kand. tekhn. nauk [Active identification of linear dynamic discrete stationary objects in a time domain. PhD eng. sci. diss.]. Novosibirsk, 2007. 171 p.

22. Troshina G.V. Vychislitel'nye aspekty zadachi vosstanovleniya vektora sostoyaniya dlya modeli s netochno zadannymi parametrami [Computing aspects of problem of the state vector recovering for models with inexact given parameters]. Sbornik nauchnykh trudov Novosibirskogo gosudarstvennogo tekhnicheskogo universiteta – Transaction of scientific papers of the Novosibirsk state technical university, 2008, no. 3 (53), pp. 25–34.

23. Voevoda A.A., Troshina G.V. Active identification of linear stationary dynamic object on base of the Fisher information matrix: the steady state. Proceedings of the XII International Conference “Actual problems of electronic instrument engineering (APEIE-2014)”, Novosibirsk, Russia, 2–4 October 2014, pp. 745–749. doi: 10.1109/APEIE.2014.7040785

24. Voevoda A.A., Troshina G.V., Active identification of the inverted pendulum control system [Aktivnaya identifikaciya sistemy upravleniya perevernutij mayatnik]. Proceedings of the 18^th International Conference on Soft Computing and Measurements (SCM'2015). St. Petersburg, LETI Publ., 2015, vol. 1, pp. 153–156.

25. Troshina G.V., Voevoda A.A., Patrin V.M., Simakina M.V. The object unknown parameters estimation for the 'inverted pendulum-Cart' system in the steady state. Proceedings of the 16^th International Conference of Young Specialists on Micro/Nanotechnologies and Electron Devices (EDM 2015), Altai, Erlagol, 29 June – 3 July 2015, pp. 186–188.

26. Voevoda A.A., Troshina G.V. O nekotorykh metodakh fil'tratsii v zadache identifikatsii [About some filtration methods in the identification problem]. Sbornik nauchnykh trudov Novosibirskogo gosudarstvennogo tekhnicheskogo universiteta – Transaction of scientific papers of the Novosibirsk state technical university, 2014, no. 2 (76), pp. 16–25.

27. Voevoda A.A., Troshina G.V. Ob otsenke vektora sostoyaniya i vektora parametrov v zadache identifikatsii [About parameters vector estimation and state vector estimation in identification problem]. Sbornik nauchnykh trudov Novosibirskogo gosudarstvennogo tekhnicheskogo universiteta – Transaction of scientific papers of the Novosibirsk state technical university, 2014, no. 4 (78), pp. 53–68. doi: 10.17212/2307-6879-2014-4-53-68

28. Troshina G.V. Modelirovanie dinamicheskikh ob"ektov v srede Simulink. Ch. 1 [The dynamic objects modelling in Simulink environment. Pt. 1]. Sbornik nauchnykh trudov Novosibirskogo gosudarstvennogo tekhnicheskogo universiteta – Transaction of scientific papers of the Novosibirsk state technical university, 2015, no. 3 (81), pp. 55–68. doi: 10.17212/2307-6879-2015-3-55-68