A.A. Kobzev - Dr. Sc. (Eng.), Professor, Head of Department «Mechatronics and Electronic Systems of Vehicle», Vladimir State University named after A.&N. Stoletovs. E-mail: kobzev42@mail.ru Yu.E. Mishulin - Ph. D. (Eng.), Associate Professor, Department «Mechatronics and Electronic Systems of Vehicle», Vladimir State University named after A.&N. Stoletovs. E-mail: mishulin59@mail.ru N.A. Novikova - Ph. D. (Eng.), Associate Professor, Department «Mechatronics and Electronic Systems of Vehicle», Vladimir State University named after A.&N. Stoletovs. E-mail: natalia5104@mail.ru A.V. Lekareva - Post-graduate Student, Department «Mechatronics and Electronic Systems of Vehicle», Vladimir State University named after A.&N. Stoletovs. E-mail: tasya671@rambler.ru

In some automatic systems (ACS) because of design features of the control object or the operating conditions it is impossible to obtain complete and accurate information about adjustable coordinate. The object of control in such systems is not covered by the feedback, therefore, not considered its dynamic properties, which negatively affects the quality of the work of the ACS. These objects include machining equipment, industrial robots, tracking systems for moving objects, control objects, situated on tracked and wheeled vehicles, mobile robots with attachments, etc. the simulation Results show that the lack of complete information about the coordinate variable dramatically reduces the performance of ACS, especially under changing external influences. Increase the impact of specifying causes in such systems, the increase in the rate of oscillatory and may lead to instability. At harmonic input signals, the increase of the frequency causes a decrease in precision, steady-state dynamic error increases to unacceptable values. One of the possible variants of a rational structure of the ACS, to improve the performance management is the introduction of parallel models that form the signal «more control», you can use 2 of the parallel structures model: 1) the model is similar to the original ACS, but the main feedback variable coordinate; 2) a model that generates a given («reference») transition process under various external influences. Studies have shown that the introduction of parallel models that have a structure similar to the original ACS, but with the object of control, is fully covered by feedback, allows to stabilize the dynamic characteristics only in a limited range of external influences. When significant changes impact of specifying a model can be a source of additional disturbances and often leads to degradation of dynamic performance and stability. The p parallel model of the second type have a higher dynamic performance. As in the first and in the second case, it is necessary to enter in the chain of formation of the «additional control» element with variable transmission coefficient, which significantly complicates the structure of the ACS and control problems, since its value depends on many factors. Introduction in the channel of the reference signal SAU with the parallel element model predict external influences allows to improve dynamic characteristics of the system without the use of a nonlinear element with a variable transmission coefficient. Forecasting of external influences allows us to estimate the system response at a given exposure for some time ahead (forecasting step). On the basis of these assessments formed the corrective actions, allowing to minimize the error of the control system. When choosing a method of prediction functions in dynamic systems is required to obtain the best approximation with the minimum number of measured values. The main issues are the choice of the model type predicted functions and development of algorithms for forecast of dynamics of object of control. When the study considered two ways of constructing a predictive element based on factual methods using statistical analysis and processing of the obtained in the monitoring process: 1) forecasting using interpolation polynomials; 2) approximation of the data using the least squares method. The generalization of the results of modeling of algorithm of the predictive model, showed an advantage of forecasting with the use of the interpolation formula of Newton. The simulation results of the p parallel model prediction showed that the maximum absolute error of the adjustable values does not exceed 1% of the maximum value specifying the impact.

 

1. Kobzev A.A., Mishulin JU.E., Novikova N.A., Noskov E.V. Analiz sistem upravlenija s ne polnostju nabljudaemojj reguliruemojj koordinatojj // Izvestija instituta inzhenernojj fiziki. 2010. № 4. S. 16−21.
2. Novoselov B.V., Kobzev A.A., Mishulin JU.E., Batazov V.N. Sistema avtomaticheskogo upravlenija obektov vooruzhenija s prognoziruemojj modelju // Oboronnaja tekhnika. 2006. № 8. S. 29−32.
3. Alekseev E.R., CHesnokova O.V. Reshenie zadach vychislitelnojj matematiki v paketakh Mathcad12, MATLAB7, Mapl9. M.: NT Press. 2006.
4. Novoselov B.V., Kobzev A.A. Sistemy sledjashhikh privodov s samonastrojjkojj vnutrennikh parametrov i korrekciejj upravljajushhego vozdejjstvija // Voprosy oboronnojj tekhniki. 1998. Ser. 9. № 2 (222). S. 8−11.
5. Novikova N.A. Sposob nelinejjnojj korrekcii avtomaticheskikh sistem // Proektirovanie tekhnologija ehlektronnykh sredstv. 2003. № 3. S. 14−16.
6. Gudvin G.K., Grebe S.F., Salgado M.EH. Proektirovanie sistem upravlenija. M.: BINOM. Laboratorija znanijj. 2004.
7. Kobzev A.A., Mishulin JU.E., Novikova N.A. Algoritm raboty prognozirujushhego analizatora v sistemakh upravlenija dvizheniem // Izvestija VUZov. EHlektromekhanika. 2014. № 2. S. 68−73.