Jtam-A4.dvi JOURNAL OF THEORETICAL AND APPLIED MECHANICS 54, 4, pp. 1067-1078, Warsaw 2016 DOI: 10.15632/jtam-pl.54.4.1067 SDRE APPLIED TO POSITION AND VIBRATION CONTROL OF A ROBOT MANIPULATOR WITH A FLEXIBLE LINK Jeferson José de Lima, Angelo Marcelo Tusset, Frederic Conrad Janzen, Vincius Piccirillo, Claudinor Bitencourt Nascimento UTFPR, Ponta Grossa, PR, Brazil; email: jefersonjl82@gmail.com; tusset@utfpr.edu.br; fcjanzen@utfpr.edu.br; piccirillo@utfpr.edu.br; claudinor@utfpr.edu.br José Manoel Balthazar UNESP, Bauru, SP, Brazil; email: jmbaltha@gmail.com Reyolando Manoel Lopes Rebello da Fonseca Brasil UFABC, Santo Andre, SP, Brazil; email: reyolando.brasil@ufabc.edu.br This paper presents position and vibration control of a flexible robot composed of two rigid andoneflexible links.Position is controlledby the currentappliedto theDCmotorarmature. To control vibrations of the flexible structure, Shape Memory Alloys (SMA) are used. Due to phase transformations, the SMA can change its stiffness through temperature variation, considering and taking advantage of this characteristic the vibration control is done.Control is achieved via the State Dependent Ricatti Equations (SDRE) technique, which uses sub- optimal control and system local stability search.The simulation results show the feasibility of the proposed control for the considered system. Keywords: SDRE control, flexible robotic arm, SMA, active vibration control 1. Introduction In the last decades, several researches on robotic manipulators with flexible links have come to attention due to their many advantages over rigid manipulators (Kalyoncu, 2008). Shawky et al. (2013) remarked that flexible manipulator systems offer several advantages in contrast to the traditional rigid manipulator. The advantages include faster response, lower energy consumption, relatively smaller actuators and less overall mass and, in general, smaller final cost. Due to its flexible nature, control of these flexible systems takes into account both rigid body and elastic degrees of freedom. It is important to recognize the flexible nature of the manipulator and construct a mathematical model for the system that accounts for the interactions between the actuators and payload. According to Banks et al. (1996) andGabbert and Schulz (1996), it is necessary to improve robotic manipulators including the use of smart materials to assist control by changing the structural behavior. There are several studies suggesting solutions to the problem of handlers with flexible ele- ments such as (Sawada et al., 2004), which have been exploited for the purposes of efficiency and flexibility of a handler in microgravity environment. In Inifene (2007), the dynamics of a flexible link and a comparative study of some implemented control is shown. (Heidari et al., 2013) proposes a nonlinear finite element modeling of flexibility of handlers. In Grandinetti et al. (2012), a three-degree-of-freedom cylindrical manipulator system with a flexible link on its tip is studied and experimentally implemented, and the active vibration control is analyzed. In Halim et al. (2014), the use of a decentralized vibration control scheme for suppressing vibration of a multi-link flexible robotic manipulator using embedded smart piezoelectric transducers is 1068 J.J. Lima et al. investigated. In Pereira et al. (2012), an adaptive control applied to single link-flexible manipu- lators, utilizing a feedback controller of the joint angle to guarantee trajectory tracking of the joint angle with an adaptive input shaper updated by algebraic non-asymptotic identification is proposed. Bottega et al. (2009) proposeda tracking controlmodel for aflexible link roboticmanipulator using simultaneouslymotor torques and piezoelectric actuators.Molter et al. (2010) used SDRE control technique design for flexible manipulators using piezoelectric actuators. In Shawky et al. (2013), the tip position of a one-link flexible manipulator was modeled and controlled. The control strategy was based on the SDRE control designmethod in the context of application to robotics.The experimental results revealed that thenonlinear SDREcontroller was near optimal and robust, and its performance compared favorably to the conventional PD controller strategy. This work proposes the control of a robotic manipulator for aerospace applications. It is composed of two rigid links modeled as proposed by Korayem et al. (2011), and one flexible linkmodeled as proposed by Janzen et al. (2014). The gravity effects are not considered, similar to the proposal presented by Sawada et al. (2004). In order to reduce vibrations of the flexible manipulator, it is proposed to use SMAs. ShapeMemoryAlloys (SMAs) are a group ofmetallic materials with the ability to return to a previously defined shape or size, after deformation, by applying a specific temperature. This effect occurs due to the shift in the materials crystalline structure between two different phases, called martensite and austenite (Piccirillo et al., 2008, 2009).WhenSMAsareheated, theypresent changes in crystal structure.These changes generate large forces that can be used in actuator systems (Ge et al., 2013). In this way, SMAs can be used in many application areas, like robotics, since they allow for a lighter weight manipulator and longer links, which are important for aerospace applications. The proposed feedback control system is designed to take the system into a desired coordi- nate. The feedback control is obtained using SDRE. The SDRE control choice is because the control computer algorithm is simple and highly effective for nonlinear feedback control. It has been successfully used in roboticmanipulators (Innocenti et al., 2000; Xin et al., 2001; Korayem et al., 2010; Nekoo, 2013; Korayem andNekoo, 2014, 2015a), and in roboticmanipulators with a flexible joint or/andwith flexible link (Korayem et al. 2011; Fenili andBalthazar, 2011; Shawky et al., 2013; Lima et al., 2014; Korayem and Nekoo, 2015b). The paper is organized is as follows. Section 2 provides the dynamic model for a two rigid and a flexible link manipulator. Section 3 presents the SDRE control project considering both control of justpositioning,withoutvibration control of theflexible link, as in thecase of vibration control of the flexible link by temperature variation (SMAs). Concluding remarks are given in Section 4. 2. Mathematical model The adopted manipulator model consists of two rigid links and the third flexible one. The DC motors and joints of the roboticmanipulator are presented inFig. 1, and the transmissionmotor torque to the manipulator is given by Eq. (2.6), The governing equations of motions are obtained through Lagrangian formalism. To derive the equations of motion of the manipulator, one needs the kinetic energy ℑ and the potential energy Γ in the Lagrangian function L L =ℑ−Γ (2.1) Next, onemust apply the equations of Euler-Lagrange (Lima et al., 2014) d dt (∂L ∂θ̇i ) − ∂L ∂θi = τi (2.2) SDRE applied to position and vibration control... 1069 Fig. 1. Schematic of the manipulator model, where θ1,2,3 are the joint angles of each link, θ4,5 are the correspondingmotor angles for the link, bs is the damping constant, k is the spring constant, and the variable z is the displacement of the beam end The right hand side of Eq. (2.2) are non-conservative forces. Thus, we obtain Eq. (2.3), the equations of motion of the link (Korayem et al., 2011; Lima et al., 2014) M(θ)θ̈=−C(θ, θ̇)−F(θ̇)+τ (2.3) whereM(θ) is the inertia matrix, θ̈ represents the acceleration of the i-th link,C(θ, θ̇) are the centrifugal and Coriolis forces, F(θ̇) describes friction between the joint and the link and τ is the torque effect on the rotors. The system matrices C(θ, θ̇) andF(θ̇) are presented by C(θ, θ̇)=    v11 v12 v13 v21 v22 v23 v31 v32 v33       θ̇1 θ̇2 θ̇M3    F(θ̇)=    kaθ̇1 kaθ̇2 bvθ̇M3    (2.4) where v3×3 are the variables of centrifugal/coriolis forces. The equations that define the dynamics of theDCmotor, are obtained by Eq. (2.1) andEq. (2.2), and represented by θ̈Mi = bvθ̇Mi −ktii+ τs J i̇i = −RMii−kbθ̇Mi+Vi Lm (2.5) where θ̈Mi is the acceleration and θ̇Mi the velocity of DC motor, ii is the current, Vi is the voltage applied in the DC motor armature, RM, kb and Lm are the resistance of the armature, the electromotive force constant and the inductance of the armature of the motor, respectively. For the coupling of the equations, we replace τi = τsi in Eqs. (2.3), (2.4) and (2.5)1, thus considering the characteristic of the proposed model flexible joints, according to (Lima et al., 2014) τs =    bs(θ̇M1 − θ̇1)+k(θM1−θ1) bs(θ̇M2 − θ̇2)+k(θM2 −θ2) ki3    (2.6) The dynamics of the system with a flexible beam coupled to the shaft of the DC motor is given by (Janzen et al., 2014) i̇ = −RMi−kbθ̇+V LM θ̇v = −bvθ̇+kti−C(θ, θ̇)+EIφ ′′ 0z Jv z̈ =−µż−w2z−αvθ̈+ θ̇ 2z (2.7) 1070 J.J. Lima et al. where E is Young’smodulus and I is themoment of inertia of the cross section of the beam, z is the displacement for the firstmode of vibration of the beam and φ′′0 is the constant of the beam vibrationmodes for the firstmode, obtained by applying the assumedmodemethod (Fenili and Balthazar, 2011). Thus, the proposed system is represented by Eq. (2.3) coupledwith Eq. (2.7), in the state space of the variables, represented by ẋ1 = x2 ẋ2 = p11(−kx1+kx3+ bsx4)+p12(−kx6+kx8+ bsx9)−p13(EIφ ′′ 0x14) +ϕ11x2+ϕ12x6+ϕ13x11 ẋ3 = x4 ẋ4 = k Jm x1+ bs Jm x2− k Jm x3− bs+ bv Jm x4+ kt Jm x5 ẋ5 =− kb Lm x4− Rm Lm x5 ẋ6 = x7 ẋ7 = p21(−kx1+kx3+ bsx4)+p22(−kx6+kx8+ bsx9)−p23(EIφ ′′ 0x14) +ϕ21x2+ϕ22x6+ϕ23x11 ẋ8 = x9 ẋ9 = k Jm x6+ bs Jm x7− k Jm x8− bs+ bv Jm x9+ kt Jm x10 ẋ10 =− kb Lm x9− Rm Lm x10 ẋ11 = x12 ẋ12 = p31(−kx1+kx3+ bsx4)+p32(−kx6+kx8+ bsx9)−p33(EIφ ′′ 0x14) +ϕ31x2+ϕ32x6+ϕ33x11 ẋ13 =− kb Lm x14− Rm Lm x15 ẋ14 = x15 ẋ15 =−αvϕ31x12−αv+(ω 2+αv(p33(EIφ ′′ 0 +x 2 12)))x14−µx15 (2.8) where x1 = θ1, x2 = θ̇1, x3 = θM1, x4 = θ̇M1, x5 = i1, x6 = θ2, x7 = θ̇2, x8 = θM2, x9 = θ̇M2, x10 = i2, x11 = θM3, x12 = θ̇M3, x13 = i3, x14 = z, x15 = ż and ϕnm are the nonlinear elements. The indexn indicates the equation of the system θ̈n it belongs to, and m is the corresponding coefficient of velocity θ̇m. We can see the elements of ϕnm below ϕ11 = p11(−ka−v11− bs)−p12v21−p13v31 ϕ12 =−p11v12+p12(−v22−ka−bs)−p13v32 ϕ13 =−p11v13−p12v23+p13(−v33− bv) ϕ21 = p21(−ka−v11− bs)−p22v21−p23v31 ϕ22 =−p21v12+p12(−v22−ka−bs)−p23v32 ϕ23 =−p21v13−p12v23+p23(−v33− bv) ϕ31 = p31(−ka−v11− bs)−p32v21−p33v31 ϕ32 =−p31v12+p32(−v22−ka−bs)−p33v32 ϕ33 =−p31v13−p32v23+p33(−v33− bv) (2.9) and p3×3 =M(θ) −1 (2.10) SDRE applied to position and vibration control... 1071 3. Proposed control Consider system (2.8) in the form ẋ=A(x)x+Bu (3.1) where x∈ℜn is the state vector,A(x)∈ℜn×n is the state-dependent matrix,B∈ℜn×m is the control matrix and u is a feedback control vector. The proposed controlu can be determined using SDRE control (Balthazar et al., 2014). The methodology used in SDRE control makes the synthesis of LQR control gains varying in time (Mracek and Cloutier, 1998). The quadratic cost function for the regulator problem is given by J = 1 2 ∞ ∫ t0 [eTQe+uTRu] dt (3.2) whereQ∈ℜn×n andR∈ℜm×m are positive definite matrices, and e= [ x1−x ∗ 1 0 0 0 0 x2−x ∗ 2 0 0 0 0 x3−x ∗ 3 0 0 0 0 ] T (3.3) where x are the system state variables, and x∗ is the desired state. Assuming full state feedback, the control law is given by u=−R−1BTPe (3.4) The State-Dependent-Riccati equation to obtain P is given by P(x)A(x)+A(x)TP(x)−P(x)BR−1BTP(x)+Q=0 (3.5) Innumerical simulations,we consider two situations. First,we consider control of thehandler viaDCmotors at the joints without control of vibrations of the flexible beam.Next, we consider control of the vibration of the flexible beam. It is assumed throughout this paper that the parameters considered are presented in Table 1. Table 1.Parameters for simulation L1 =0.5m L2 =0.5m L3 =1.2m M1 =1kg M2 =1kg M3 =1.318 ·10 −4kg kt1 =0.08Nm/A kt2 =0.08Nm/A kt3 =0.058Nm/A kb1 =0.008V/rpm kb2 =0.008V/rpm kb3 =0.008V/rpm bv1 =0.2Nms/rad bv2 =0.2Nms/rad bv3 =4.62Nms/rad Rm1 =1.7Ω Rm2 =1.7Ω Rm3 =1.914Ω Lm1 =0.003H Lm2 =0.003H Lm3 =3.4 ·10 −3H E =0.7 ·1011N/m2 Im =6.54 ·10 −5kg/m2 Ie =6.54 ·10 −7kg/m2 I =1.56 ·10−13m4 µ =0.1 φ′′0 =4.898m αv =0.821 w =11.3097rad/s The initial values for x are given by: x1(0) = π, x2(0)= 0, x3(0)= π, x4(0)= 0, x5(0)= 0, x6(0) = 0.6981, x7(0) = 0, x8(0) = 0.6981, x9(0) = 0, x10(0) = 0, x11(0) = 0.5236, x12(0) = 0, x13(0)= 0, x14(0)= 0 and x15(0)= 0. The desired state is: x∗1 =2.0944, x ∗ 2 =0, x ∗ 3 = x3, x ∗ 4 = x4, x ∗ 5 = x5, x ∗ 6 =−0.6981, x ∗ 7 =0, x∗8 = x8, x ∗ 9 = x9, x ∗ 10 = x10, x ∗ 11 = −0.5236, x ∗ 12 = 0, x ∗ 13 = x13, x13(0) = 0, x ∗ 14 = 0 and x∗15 =0. 1072 J.J. Lima et al. 3.1. Position control for DC motor (control I) Consider the case inwhich the controlu is applied only to position control of θ1, θ2 and θM3, and flexible link vibrations are not controlled directly. Thus, we defineQ,R=R1 and consider u= [u1 u2 u3] T, A= [A1 A2], B=B1, where the control signal (u1, u2 and u3) represents the voltage applied to the motor terminals (matrices Q, R1, A1, A2 and B1 in Appendix). In Figs. 2a,b,c, we can observe the position error of the manipulator (θ1, θ2 and θM3), considering only position control of the link by the motor current. In Fig. 2d, we can observe vibrations at the tip of the flexible link (z). Fig. 2. (a) Positioning θ1, (b) positioning θ2, (c) positioning motor θM3, (d) displacement of the beam z Although the system controlled by the motors can position the links at the desired points, we observe oscillations in the positioning (Fig. 2a,b,c) caused by vibrations of the flexible link (Fig. 2d). 3.2. Position control for DC motor and beam deflection control (control II) In the adopted second strategy we control θ1, θ2 and θM3 and control the flexible link vibrations using SMAs fixed to the link. SMAs capacity to change their physical properties with temperature variation allows for their use asmeans to vibration suppression of the flexible link. Falk (1980, 1993) proposed the SMAmathematical model used in this paper. It is based on the Devonshire theory that considers free energy in a polynomial form. This model does not consider internal variables and potential dissipation to describe the pseudo-elasticity and the shapememory effect. TheFalkmodel considers only the temperature T and the deformation (ε) (Janzen et al., 2014). The SMA polynomial model was chosen due to its capacity to represent qualitatively the pseudo-elasticity and shapememory effects of the SMA structure. TheHelmholtz free energy ψ SDRE applied to position and vibration control... 1073 waschosen sothereareminimumandmaximumpoints that represent the stability and instability of the SMA phases. The potential of free energy can be described by ρψ(ε,T) = 1 2 q(T −TM)ε 2− 1 4 bε4+ b2ε6 24q(TA−TM) (3.6) The variables TA and TM are temperatures in which the austenitic and martensitic phases are stable, q and b are positive constants. A constitutive equation can be written as (Savi et al., 2002) σ = q(T −TM)ε− 1 4 bε3+ b2ε5 4q(TA−TM) (3.7) where b =1.868·107MPa, q =523.29MPa/K,TM =364.3KandTM =288K for (Cu-Zn-Al-Ni). The SMA element can be coupled to the beam-like structure as a stiffness component which can be described by k(x,T) = q(T −TM)x− bx 3+ex5 (3.8) where q = qAr L b = bAr L3 e = eAr L5 Considering the case inwhich the objective is to position the beam-like link and reducing its oscillations, we defineQ,R=R2 and consideruf = [u1 u2 u3 u4] T,A= [A1 A3] andB=B2 (matrices Q, R2, A1, A3 and B2 in Appendix). And considering: q =1.56987, b =114367.348 and e =7232491.36 (Janzen et al., 2014). The control signal (u1, u2 and u3) represents the voltage applied to themotor terminals, and (u4) represents the temperature variation of the SMA given by u4 = ∆T = T −TM (3.9) In Figs. 3a,b and 3c, we can observe the manipulator position error (θ1, θ2 and θM3) consi- dering the positioning control of links through the motors, and the vibration control of flexible link through the use of SMA in Fig. 3d (z). As we can see in Fig. 3, the addition of SMA reduces vibrations of the flexible link, therefore reducing oscillations (θM3). Thus, we avoid inversion of the torques of themotors, which caused the oscillations seen in Fig. 2. In Fig 4a, we can observe the temperature variation of the SMA to control the vibration of the flexible link (∆T), and in Fig. 4b we present the comparative results between the two control strategies proposed. First, the structure is fully austenite, and it can be seen fromFig. 4 that no phase transfor- mation occurs in the alloy, namely, theSMAremains in the austenite phase (TA ­ 364.3K).This variation of the parameter ∆T acts on the beam rigidity (Eq. (3.8)) and reduces the amplitude of displacement of the flexible link as can be observed in Fig. 4b. It is also possible to observe that ∆T remained at a constant value of ∆T = 77K for t > 0.029s, due to the need of keeping the SMA in the austenite phase, enables control of the stiffness according to equation (3.8). By applying DC control motors and using flexible SMA one observes a smoother response for all angles θ1, θ2 and θM3 and a smaller amplitude of oscillation z. 1074 J.J. Lima et al. Fig. 3. (a) Positioning θ1, (b) positioning θ2, (c) positioning motor θM3, (d) displacement of the beam z Fig. 4. (a) Temperature variation in SMA (u4 = T −288= ∆T), (b) displacement of the beam z 4. Conclusion We can observe in the simulations results that vibration control of the flexible link allows for a more efficient control than the control provided by the motors only. This demonstrates the importance of controlling vibrations of the flexible link. By using the polynomial model of SMA it is possible to estimate the temperature variation in the SMA to attenuate vibrations of the flexible link. The temperature variation of the SMA causes changes in their physical characteristics, allowing for its use as a way to reduce vibrations of the flexible beam. The numerical results show that the proposed control strategy using SMA in vibration control of the flexible links is a good alternative to constructing lighter handlers with longer links, such as those applied in aerospace activities. SDRE applied to position and vibration control... 1075 Acknowledgements Theauthorswould like to acknowledgeConselhoNacionaldeDesenvolvimentoCientfico eTecnológico – CNPQ grant: 484729/2013-6 and grant: 447539/2014-0, for the financial support. Appendix A1 =                                0 1 0 0 0 0 0 0 0 0 −p11k ϕ11 p11k p11bs 0 −p12k ϕ12 p12k p12bs 0 0 0 0 1 0 0 0 0 0 0 k Jm bs Jm − k Jm −bs+bv Jm kt Jm 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 − kb Lm −Rm Lm 0 0 0 0 0 −p21k ϕ21 p21k p21bs 0 −p22k ϕ22 p22k p22bs 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 k Jm bs Jm − k Jm −bs+bv Jm kt Jm 0 0 0 0 0 0 0 0 − kb Lm −Rm Lm 0 0 0 0 0 0 0 0 0 0 −p31k ϕ31 p31k p31bs 0 −p32k ϕ32 p32k p32bs 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0                                BT1 =    0 0 0 1 LM 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 LM 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 LM 0 0    Q15×15 = {qij} qij =      0 if i 6= j 1 if i =2,3,4,5,7,8,9,10,12,13,15 100 if i =1,6,11,14 R1 =    0.1 0 0 0 0.1 0 0 0 0.1    R2 =      0.1 0 0 0 0 0.1 0 0 0 0 0.1 0 0 0 0 0.1      A2 =                               0 0 0 0 0 0 ϕ13 p13kt −p13EIφ ′′ 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 ϕ23 p23kt −p23EIφ ′′ 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 ϕ33 p33kt −p33EIφ ′′ 0 0 0 − kb LM −RM LM 0 0 0 0 0 0 1 0 −αvϕ33 αvp33kt w 2+αvp33EIφ ′′ 0 +x 2 12 µ                               1076 J.J. Lima et al. B2 =                               0 0 0 0 0 0 0 0 0 0 0 0 1 LM 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 LM 0 0 0 0 0 0 0 0 0 0 0 0 1 LM 0 0 0 0 0 0 0 0 −qx14                               A3 =                               0 0 0 0 0 0 ϕ13 p13kt −p13EIφ ′′ 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 ϕ23 p23kt −p23EIφ ′′ 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 ϕ33 p33kt −p33EIφ ′′ 0 0 0 − kb LM −RM LM 0 0 0 0 0 0 1 0 −αvϕ33 αvp33kt w 2+αvp33EIφ ′′ 0 +x 2 12−x 2 14(b−ex 2 14) µ                               References 1. Balthazar J.M., BassinelloD.G., TussetA.M., Bueno A.M., Pontes B.R., 2014,Nonli- near control in an electromechanical transducerwith chaotic behavior,Meccanica,49, 8, 1859-1867, doi: 10.1007/s11012-014-9910-4 2. BanksH.T., SmithR.C.,WangY., 1996,SmartMaterial Structures:Modeling, Estimation and Control, JohnWiley & Sons, Paris 3. BottegaV., PergherR., Fonseca J.S.O., 2009, Simultaneous control and piezoelectric insert optimization for manipulators with flexible link, Journal of the Brazilian Society of Mechanical Sciences and Engineering, 31, 2, 105-116, doi: 10.1590/S1678-58782009000200003 4. Falk F., 1980,Model free energy, mechanics, and thermodynamics of shape memory alloys,Acta Metallurgica, 28, 12, 1773-1780, doi: 10.1016/0001-6160(80)90030-9 5. Falk F., 1983, One dimensional model of shape memory alloys, Archives of Mechanics, 35, 1, 63-84. 6. Fenili A., Balthazar J.M., 2011, The rigid-flexible nonlinear robotic manipulator: modeling and control, Communications in Nonlinear Science and Numerical Simulation, 16, 5, 2332-2341, doi: 10.1016/j.cnsns.2010.04.057 SDRE applied to position and vibration control... 1077 7. Gabbert U., Schulz I., 1996, Optimal placement of piezoelectric actuators in vibration control of adaptative structures,Proceedings of ASME InternationalMechanical Engineering Congress and Exposition, 271-277 8. Ge S.S., Tee K.P., Vahhi I.E., Tay F.E.H., 2006, Tracking and vibration control of flexible robots using shape memory alloys, IEEE/ASME Transactions on Mechatronics, 11, 690-698, doi: 10.1109/TMECH.2006.886242 9. Grandinetti F.J., Soares A.M.S., Lamas W.Q., Goes L.C.S., 2012, Notes on vibration control of amicro/macromanipulatormountedonaflexible structure,Proceedings of the Institution of Mechanical Engineers, Part K: Journal of Multi-body Dynamics March, 226, 1, 72-82, doi: 10.1177/1464419311424090 10. HalimD., Luo X., Trivailo P.M., 2014,Decentralized vibration control of amulti-link-flexible robotic manipulator using smart piezoelectric transducers, Acta Astronautica, 104, 186-196, doi: 10.1016/j.actaastro.2014.07.016 11. HeidariH.R.,KorayemM.H.,HaghpanahiM., BatlleV.F., 2013,Optimal trajectoryplan- ning forflexible linkmanipulatorswith largedeflectionusinganewdisplacementsapproach,Journal of Intelligent and Robotic Systems, 72, 3/4, 287-300, doi: 10.1007/s10846-012-9807-1 12. InnocentiM., Baralli F., SalottiF., Caiti A., 2000,Manipulator path control using SDRE, American Control Conference, 5, 3348-3352, doi: 10.1109/ACC.2000.879186 13. JanzenF.C.,TussetA.M.,PiccirillioV., Balthazar J.M., SilveiraM., Pontes Junior B.R., 2014, Control of slewingmotions of flexible structures using shape memory alloy actuators, Proceedings of ASME International Mechanical Engineering Congress and Exposition, 1-7 14. Jnifene A., 2007, Active vibration control of flexible structures using delayed position feedback, Systems and Control Letters, 56, 3, 215-222, doi: 10.1016/j.sysconle.2006.10.005 15. Kalyoncu M., 2008,Mathematical modelling and dynamic response of a multistraight-line path tracing flexible robot manipulator with rotating-prismatic joint, Applied Mathematical Modelling, 32, 1087-1098, doi: 10.1016/j.apm.2012.10.042 16. Korayem M.H., Irani M., Nekoo S.R., 2010, Analysis of manipulators using SDRE: a closed loop nonlinear optimal control approach,Scientia Iranica, Transaction B:Mechanical Engineering, 17, 6, 456-467 17. Korayem M., Irani M., Nekoo S.R., 2011, Load maximization of flexible joint mecha- nical manipulator using nonlinear optimal controller, Acta Astronautica, 69, 7, 458-469, doi: 10.1016/j.actaastro.2011.05.023 18. Korayem M.H., Nekoo S.R., 2014, Nonlinear optimal control via finite time horizon state- dependent Riccati equation, International Conference on Robotics andMechatronics, 878-883, doi: 10.1109/ICRoM.2014.6991015 19. Korayem M.H., Nekoo S.R., 2015a, State-dependent differential Riccati equation to track con- trol of time-varying systems with state and control nonlinearities, ISA Transactions, 57, 117-135, doi: 10.1016/j.isatra.2015.02.008 20. Korayem M.H., Nekoo S.R., 2015b, Finite-time state-dependent Riccati equation for time- varying nonaffine systems: rigid and flexible jointmanipulator control, ISA Transactions, 54, 125- 144, doi:10.1016/j.isatra.2014.06.006 21. Lima J.J., TussetA.M., JanzenF.C., PiccirilloV.,NascimentoC.B., Balthazar J.M., Brasil M.R.L.F., 2014, Nonlinear state estimation and control applied to amanipulator robotic including drive motor,Mathematics in Engineering, Science and Aerospace: the Transdisciplinary International Journal, 5, 413-425 22. MracekP.C.,CloutierJ.R., 1998,Controldesigns for thenonlinearbenchmarkproblemvia the state-dependent Riccati equationmethod, International Journal of Robust and Nonlinear Control, 8, 401-433 1078 J.J. Lima et al. 23. MolterA., SilveiraO.A.A., Fonseca J.S.O., BottegaV., 2010, Simultaneous piezoelectric actuator and sensor placement optimization and control design of manipulators with flexible links using SDREmethod, Mathematical Problems in Engineering, 1-23, doi: 10.1155/2010/362437 24. Nekoo S.R., 2013, Nonlinear closed loop optimal control: a modified state-dependent Riccati equation, ISA Transactions, 52, 2, 285-290, doi:10.1016/j.isatra.2012.10.005 25. Pereira E., Trapero J.R., Dıaz I.M., Feliu V., 2012, Adaptive input shaping for single-link flexible manipulators using an algebraic identification,Control Engineering Practice, 20, 138-147, doi: 10.1016/j.conengprac.2011.10.004 26. Piccirillo V., Balthazar J.M., Pontes Jr. B.R., Felix J.L.P., 2008, On a nonlinear and chaotic non-ideal vibrating system with Shape Memory Alloy (SMA), Journal of Theoretical and Applied Mechanics (Poland), 46, 597-620 27. Piccirillo V., Balthazar J.M., Pontes Jr. B.R., Felix J.L.P., 2009, Chaos control of a nonlinear oscillator with shape memory alloy using an optimal linear control. Part II: Non-ideal energy source,Nonlinear Dynamics, 56, 3, 243-253, doi: 10.1007/s11071-008-9396-5 28. SaviM.A.,PachecoP.M.C.L.,BragaA.M.B., 2002,Chaos in a shapememory two-bars truss, International Journal of Non-Linear Mechanics, 37, 1387-1395 29. Sawada H., Ui K., Mori M., Yamamoto H., Hayashi R., Matunaga S., Ohkami Y., 2004, Micro-gravity experiment of a space robotic arm using parabolic flight, Advanced Robotics, 18, 3, 247-267, doi: 10.1163/156855304322972431 30. Shawky A., Zydek D., Elhalwagy Y.Z., Ordys A., 2013, Modeling and nonlinear control of a flexible-link manipulator, Applied Mathematical Modelling, 37, 9591-9602, doi: 10.1016/j.apm.2013.05.003 31. Xin M., Balakrishnan S.N., Huang Z., 2001, Robust state dependent Riccati equation ba- sed robot manipulator control, International Conference on Control Applications, 3, 369-374, doi: 10.1109/CCA.2001.973893 Manuscript received February 4, 2015; accepted for print December 7, 2015