Asian Jounal of Information Technology’ © Medwell Joumals, 2006 5 (12): 1367-1372, 2006 Backstepping Control Design for Position and Speed Tracking of DC Motors Sound Chaouch and Mchamedl-Saiel Nait-Said Laboratory of Electromagnetic Induction and Propulsion Systems, Electrical Engineering Department; University of Batna, Algeria ‘Abstract: The study deals with the tracking control approach is based on the use of the non linearites in problem of DC motors using backstepping design. This the DC motor model with both electrical and mechanical dynamics and finding a direct relationship between the motor output and input quantities without affecting the speed regulation. The proposed control scheme is not only to stabilize the DC moter, but also to drive the speed tracking error to converge to zero asymptotically. System robustness and asymptotic position tracking performance are shown through simulation results, Key words: Backstepping control, DC motors, lvapunov stability, nonlinear sysiems INTRODUCTION DE drive systems are often used in industrial applications such as roboties where a wide range of speed tor position control is required. In other applications de motors are used to follow a predetermined speed or Position track under variable load. Variows nonlinear control and adaptive control methods for nonlinear system have been proposed. Among many others, those control methods incluce: input-output linearization with or without adaptive controf", feedback linearization and pseudo linearization”, adeptive contol In the past decade, research on backstepping control hhas increased") The backstepping control is a systematic and recursive design methodology for nonlinear feedbeck control. Applying these design methoxs, control objectives such as postion, velocity, can be achieved. Obviously, the control approach mentioned above leads to nonlinear controllers that are usually very complicated, and implementing this controller requires high performances microprocessors. Furthermore, this design method requires the access to the whole system dynamics, including both the ‘inner’ electrical subsystem and the "outer" mechanical subsystem, This study investigates the application of the bbackstepping control technique for tracking control of de motors. The objective i to make the motor output vit) track a desired trajectory yn; ‘This tracking _can be achieved through a backstepping algortinm by finding a direct relation between the motor output y(t) and its control input vit), The theoretical bases of the proposed control technique are derived in detail and some simulation results are provided to demonstrate the effectiveness ofthe proposed schemes, Backstepping theory: The control objective is to design a robust conttoller for the output y(t) of the system to tack the output yy of the reference model asymptotically Assume that not only yas but also its ist two derivatives Yi and Yq are all bounded functions of time. The diagram block of this backstepping control is shown in Fig, 1 ‘The backstepping design to achieve the position tracking objective is deseribed step-by-step as follows, Consider a drive system: o wt (X 85) Where uw: Conto! input 6 : System parameters ¥y Output stat x: Variable stare Step 1: For the position tracking objective, define the tracking error as 2 =Y~Yot a And its derivative as: 2 =Y~Yut 9 ‘The x, can be viewed as a virtual control in above equation. Define the following stabilizing function @ a= 0%, Corresponding Author: Scusd Chscuch, Laboratory of Elecvomagnetc Induction and Propulsion Systems, Flectical Enginezsing Department; University of Batna, Algeria 1367Asian J. Inform. Tech, 5 (12): 1367-1372, 2006 System ye “ave ‘melseping ht ‘onto we . oe ey + uk Fig. 1: Backatepping control aystem Where, ©, is a positive constant. So, the second regulated variable is chosen as 63 Oh — Vout 9 ‘The frst Lyapunor function is chosen a: oO Thon the derivative of, i Vean-e o Step 2: Hence, the derivative of the second regulated variable is calculated as ® ‘To design the controller, add terms concerning 2; t0 to form the following Lyapunov function ® ‘Using Eq, 7 and 9, the derivative of V, can be derived as Follows ao oa +a,(% +2) According to (10), the control law u is designed a8 follows: ay BoB NNO 8, Fy Fig. 2: DC Motor system dynamic model Where, ¢; is a positive constant, Substituting Eq. 11 into Bq. 10, the following equation can be obtained. ¥=-oz! -o3s0 a So, the backatepping control is asymptotically tabling De motor model: A dynamic model of the de motors used for designing the wacking conwoller, a thied-order de moter into state-space form which includes both the electrical and mechanical dynamics is given as ay UF. I cite Where x, is the motor states variables, [x; x, [0 oT.) i the input voltage 6 Rotor position (radians) «© Rotor speed rad/s) T, Electromagnetic torque (N.m) T, Load torque (Nm) Terminal voltage (V) R,L Armature resistance (Q), inductance (H) J Rotor inertia (Kg-m) F Coefficient of visoous trietion K, Torque constant (N.m/A). ‘A block diagram of the DC moter system is shown in Fig. 2A separately excited de motor with constant field ceuavert is considered inthis study, where K, is considered constant Backstepping design steps: The model developed in the preceding section is now wed to design the backstepping algorithm for rotor position and speed 1368Asian J. Inform. Tech, 5 (12): 1367-1372, 2006 tracking of a separately excited de motor, to achieve the stability and position racking objectives. Step 1: First ofall, we regard the velocity x, asthe control variables, The first regulated variable is the position tracking error signal a5 -¥a 0 Where y, isthe reference signal of angular velocity, the derivative of (14) is computed as RS XY as) If is the contro, the first Lyapunov candidate V, is chosen as a6, So the derivative of (16) is computed as Wane an = 4% Fo) At this point, the desired reference veloeity ean be selected by. as) om 3 ‘Where ois a postive design constant, therefore, (17) can be rewritten as ayy ~ozi <0 ‘The control y,, in (18) is asymptotically stabilizing hich inticates the desired velocity trajectory for position tracking, So the next step is the design to make the velocity behave according to (18). Step 2: The purpose of this control design is to achieve the reference torque, so the second regulated variable is chosen as, B= XsYa X48, ~Yy ea) With this definition, the dymmical equation of the error signal z ean be expressed as a a Hence, the derivative of (20) i calculated as: il tx +tn, - pe tph Fa a F ~azl+4 Again, if all the parameters are known, the second Lyapunov candidate is 2) ‘Toderivea stabilizing control certainly V, is obtained from V, with an added term of 2, Viana tats = ale om) In order to make the derivative of the complete Lyapunov function (24) be negative define, the torque reference is chosen as fallow 1 fe Holz, $92, -¥4 +E, 2 F Gz “Yat Xe yo =Wa-ebe, +06 +0,)2,-9,]-Pu-7, OD ‘Then, we get: 05) “ezi 323 <0 Step 3: Furthermore, the final error signal is defined as 2, on Ya ‘The derivative of the given error variable 2 is computed as: 2, Waren X14 0,)+5,(P1G +6)+ Fa) 66, Heal Ri (FUL+6, $6,841 Te +e +P) JT, —bu oy Tn onder to make negative the desivative of the complete Lyapunoy funetion (29), the torque eontrol input 4s chosen as: G0) 1369Asian J. Inform. Tech, 5 (12): 1367-1372, 2006 wohage (¥) ° 05 7 15 3 “Tine (es) Fig. 3: Input command for step of rotor speed reference 150 yada end scorer a Ba Bases) sec os sere rece eeecad ° a he a Tiwi eoecenaaicernitaarnaait Fig. 4: Motor rotor speed response for step of rotor speed reference ‘Where, ¢, > 0 a design constant, then, we get p= Tal = Oat $0 oy ‘Therefore, substituting (28) to (31), we are able to cobiain the control law as us Bie, caxtree)+ xP, +¢,)+F-a) Ho teniJ-KFlte rere, © Tye te tHe +62) Cleatty,V,.V,.V, in(19), (26) and (31)arenegatives, so it implies that the resulting closed loop system is asympéotcally stable and hence, all the erors variables 2, 2z. and z, will converge to zero asymptotically. As a result, from the definition of (14) and (20), the angular rotor é 3 {| po = : . £2 : e Tne : S is ee al position and the rotor spesd will converge to the reference speed, The above equation will be used to show that the closed loops system (13) is asymptotically stable andthe positive tracking objective i achieved. mulation results: To verily the stability and asymptotic sacking performance of the abave adaptive backstepping design, simulations are carried out in MatlabySizmulink ‘The effectiveness of the proposed tracking contraller was evaluated for position and speed tracking of a separately excited DC motor model The machine parameters according to the data acquired from an experiment platform taken from are: J = 0093 Kgm',F = 0.008, K,~0,55 Nm.A, R= 10, L ~ 0,046H. ‘The parameters ¢, used in the backstepping control are chosen as follows: ¢, = 1, e:= 1, ¢,= 3 to satisfy convergence conditions, The speed reference signal is chosen in the following two studies Case A: yne, which is o constant reference igure 3 shows the associated control input waveform of DC motor using backstepping, contol Figure 4 shows the desired speed trajectory and rotor speed response. It is noticed from these figures that speed trajectory have been satisfactorily obtained from nul eror given by 7, Front the speed tracking simulation results, we can see thatthe ratr position and the desired one present a zero err given by 7 presented in Fig. 5. The electromagnetic torque ad load torque are presented in Fig 6 its elear that tho electromagnetic torque is always 1370Asian J. Inform. Tech, forced to converge to the load torque with zero error, We note an improved stability of rotor ponition, rotor speed and electrical torque al ‘The ‘second rotor speed reference signal is a sinusoidal type. In this study, it indicates the tacking Case 2: yg which is periodic reference 30 oad tng response Motor tgs reponse 2 o 2 é Tastes) 1004 ° os 7 5 2 ° a Fig. 7: Input command for sinusoidal rotor speed reference 2» 2 dese pee 2a mcr ped pe ws Ts i 15 (22): 1367-1372, 2006 performance and control input by Fig. 7. The factual speed responses given by Fig 8 can also track the references speed in the beginning when the reference speed a sinusoidal signal. The error between the actual speed and the reference speed indicate that the proposed backstepping, control is effective os 8 ie . etn [a nana Tine (ne) T 7s 2 wo: ne ie) ws 7 is 2 Fig. 8: Motor rotor speed response fee sinwsoidal rotor speed reference 0s, T 15 2 Fig. 9: Motor rotor postion response for sinusoidal rotor speed reference 1371Asian J. Inform. Tech, 5 (12): 1367-1372, 2006 Tine) Cy T tT 2 Fig, 10: Motor load torque and electrical torque for sinusoidal rotor speed reference We can find also, that the rotor position and the torque errors given by Fig, 9 and 10 converge to zero rapidly. These references signals has sinusoidal forms, it ccan be seemed that the rotor position and the electrical torque can also track the references. From these simulation results, it is obvious that the proposed backstepping contoller is quite successful and presents an excellent performance. CONCLUSION ‘The stuly deseribes a nonlinear backstepping design, scheme forthe control of de motor system to achieve the desited speed tracking control objective. The resulting. closed-loop system i guaranteed to be asymptotically stable, and the speed tracking error and the position error are able to converge to zero asymptotically according, to Lyapunov stability theorem. Step by step control design and stability analysis are given and the effectiveness of this design is demonstrated through computer simulations, In addition, the simulation results have clearly illustrated that the proposed nonlinear hackstepping controllers are quite effective and efficient for de machine, The asymptotic tracking performance and the effectiveness of the control were achieved. REFERENCES: 1. Marino, R, S,Poresada and P. Valigi, 1993, Adaptive ‘input-output linearizing control of induetion motors, IEE Trans Automat Control, pp: 208-220. Slotine, J.and I. Weiping, «Applied nonlinear control Prentice-hall intemational Edition, 3. Douglas Lawrence, A, and WJ. Rugh, 1994, Inpot-output pseudolinearization for nonlinear systems. IEEE Trans Automat Control, pp: 2207-2217. 4 1372 IncJoung, Hand L. Sung-Joon, 1994. Input-outpat linearization with state equivalence and decoupling, IEEE Trans Automat Control, pp: 2269-2274 Von Raumer, T., JM. Dion and L. Dugard, 1994. Applied Nonlinear control of an induction motor ‘using digital signal processing, IEEE Trans Control Sys. Teoh: pp: 327-335, Ghanes, M, A. Glumineau and J. DeLeon, 2004. Backstepping observer validation for sensorless induction motor on lov frequencies Benchmark. IEEE Intl Conference on Industrial Tech, CIT), pp: 1368-1373, Tan, Hand J. Chang, 1999, Adaptive position ‘control of induction motor systems under mechanical ‘uncertainties. IEEE Intemational conference on power electronics and drive systems, PEDS'99, Hong Kong, pp: 597-602, Lin, FJ. and C.C. Lee, 2000. Adaptive backstepping ‘control for linear induction motor dive to track periodic references. IEE proceeding on electronic power application, pp: 449-458 Tan, Y,, J. Chang and H. Tan, 2008. Adaptive backstepping control and fiction compensation for AC Servo with inertia and lond uncertainties. IEEE, ‘Transactions on industrial electronics, pp: 944-950. Sheng, SK. and J.S. Lin, 2005, Sensorless speed tracking control with backstepping devign schome for permanent magnet synchronous motors. Proceedings of the 2005 IEEE conference on control applications, Toronto, Canada, pp: 487-492. ‘Tan, H. and J. Chang, 1999. Adaptive Backstepping contol of induction motor with uncertainties Proceedings ofthe American control conference, San Diego California, pp: 1-5 ‘Nehvit, MH. and. Fatehi, 1996, Tracking control of DC Motors via inpat-ontput linearization, Proceeding, ‘of Electric Machines and Power Sys, pp: 237-245.