Table of Contents Author Guidelines Submit a Manuscript
Mathematical Problems in Engineering
Volume 2013, Article ID 897579, 10 pages
http://dx.doi.org/10.1155/2013/897579
Research Article

Discrete Model Reference Adaptive Control for Gimbal Servosystem of Control Moment Gyro with Harmonic Drive

1Science and Technology on Inertial Laboratory, Beihang University, Beijing 100191, China
2Fundamental Science on Novel Inertial Instrument & Navigation System Technology Laboratory, Beihang University, Beijing 100191, China

Received 29 December 2012; Revised 14 March 2013; Accepted 30 March 2013

Academic Editor: Tadashi Yokoyama

Copyright © 2013 Bangcheng Han et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

The double-gimbal control moment gyro (DGCMG) demands that the gimbal servosystem should have fast response and small overshoot. But due to the low and nonlinear torsional stiffness of harmonic drive, the gimbal servo-system has poor dynamic performance with large overshoot and low bandwidth. In order to improve the dynamic performance of gimbal servo-system, a model reference adaptive control (MRAC) law is introduced in this paper. The model of DGCMG gimbal servo-system with harmonic drive is established, and the adaptive control law based on POPOV super stable theory is designed. The MATLAB simulation results are provided to verify the effectiveness of the proposed control algorithm. The experimental results indicate that the MRAC could increase the bandwidth of gimbal servo-system to 3 Hz and improve the dynamic performance with small overshoot.

1. Introduction

With the superior characteristics of large torque amplification and high precision, the control moment gyros (CMGs) are the key actuators for the attitude control system of space stations, satellites, and other crafts [1, 2]. CMG mainly consists of gimbal servosystem and high-speed rotor system. According to the number of the gimbal DOF (degree of freedom), the CMG can be divided into single-gimbal CMG (SGCMG) and double-gimbal CMG (DGCMG) [3, 4]. Depending on the supporting manner of high-speed rotor, CMG can be divided into mechanical bearing CMG and magnetically suspended CMG [5, 6].

The CMG outputs gyro moment through changing the direction of high-speed rotor angular momentum by driving the gimbal. The gyro moment equation is where is gyro moment, is gimbal angular velocity, and is high-speed rotor angular momentum. As shown in Figure 1, the angular velocity of inner and outer gimbal is, respectively, and . is the high-speed rotor angular momentum. and are gyro moments and should be overcome by the torque motor of gimbal servosystem. Considering space applications, there are strict requirements for the size and weight of CMG. In order to reduce the weight and volume of torque motor, the DGCMG gimbal servosystem in this paper employs harmonic drive as shown in Figure 2.

897579.fig.001
Figure 1: Structure diagram of DGCMG.
897579.fig.002
Figure 2: Harmonic drive.

Compared to other speed reducers, harmonic drives have many useful properties including near-zero backlash, high gear ratio, light weight, and compact size. They have high velocity reduction in a relatively small package permitting and high torque amplification with only small motors. Hence they are widely used in precision application including industrial and space robots, medical equipment, measuring instruments, and military defense equipment [7].

Because of the structure characteristics and assembly error, harmonic drives also have some nonlinear characters like nonlinear torsion-stiffness, nonlinear friction, kinematic error, and so forth, which degrade system transmission performance [8, 9]. Particularly, due to the thin-wall structure of flexspline, harmonic drives have a low and nonlinear torsion-stiffness [10]. The DGCMG gimbal servosystem is actually a two-mass system with flexible joint [11, 12]. The joint flexibility could bring poor dynamic performance. In [12], an analysis of control structure for electrical drive system with elastic joint is carried out. To solve this problem, many literatures have been reported. In [13], a PI controller with an analytic gain selection is proposed. The gains are given as functions of system parameters and desired dominant closed-loop poles. Due to the nonlinearity, the PI controller with fixed proportional and integral coefficients could not always meet the requirements. Several adaptive control schemes with torque feedback are introduced in [1416]. In [15], the adaptive control algorithm based on virtual decomposition could improve the tracking performance obviously compared to PID controller. Other control laws have been proposed for flexible joint. In [17], a model following adaptive control method for nonlinear flexible joint robots with time-varying parameters is introduced. An LQGR/LQG control for trajectory tracking control of flexible joint robotic system is presented in [18]. In [19], an   controller has a good step-response with the exception of larger overshot than PID controller.

The model reference adaptive control (MRAC) is a traditional adaptive control method which has been widely adopted in engineering practice, such as synchronization motor control [20, 21], power devices [22, 23]. Compared to conventional controllers, MRAC could achieve high precision performance under the condition of friction, load disturbance, parameter variation, and other nonlinear factors. In [20], an MRAC-based current control scheme of PMSM is presented to improve steady-state response degraded under the motor parameter variation. In [21], MRAC is designed for PMSM servosystem. The results indicate that MRAC has the features of smaller current fluctuation and faster regulated speed than the systems using PI controller.

In this paper, a model reference adaptive control strategy is introduced to improve the dynamic performance of DGCMG gimbal servosystem with harmonic drive. Because only the input, output, and error are used and the measurement of the torque is not required, the proposed method is simple. The reference model for MRAC is selected according to the requirement for dynamic performance with fast response and small overshoot. The discrete POPOV super stable theory is used to design the adaptive controller. Last the performance of MRAC is compared with that of a conventional PID controller.

This paper is organized as follows: Section 2 outlines the model of gimbal servosystem with harmonic drive; the selection of reference model is presented in Section 3; in Section 4, the discrete adaptive controller is designed based on POPOV super stable theory; in Section 5, the simulation analyses are presented; the experimental results in DGCMG are presented in Section 6.

2. The Model of Gimbal Servosystem with Harmonic Drive

As shown in Figures 1 and 3, DGCMG consists of inner and outer gimbal servosystems and a high-speed rotor system supported by inner gimbal (gyro-room). The mounting ring is fixed on a rigid base which is immobile in experiment.

897579.fig.003
Figure 3: View of DGCMG.

The inner and outer gimbal servosystems have the similar framework with same torque motor and harmonic drive. What is different is the rotational inertia of load. They also apply the same control method and could be controlled independently. So only the inner gimbal is studied as an example in this paper. The outer gimbal could adopt the same control method with different parameters. The structure diagram of the inner gimbal servosystem is shown in Figure 4. The two ends of the gimbal are, respectively, named as motor terminal and load terminal. The motor terminal mainly contains of torque motor and harmonic drive. The load terminal mainly consists of high-speed rotor, gyro-room, and photoelectric encoder. The harmonic drive consists of circular spline, flexspline, and wave generator.

897579.fig.004
Figure 4: Structure diagram of inner gimbal.

The electrical model of torque motor can be described as [24]

The gimbal servosystem with harmonic drive is essentially a two-mass system [11]. The dynamic model is given as

The gimbal control system has two control loops, speed loop and current loop. According to (3)-(4), the control diagram of the inner gimbal servosystem is shown in Figure 5. In Figure 5, is the speed loop controller.

897579.fig.005
Figure 5: Control diagram of inner gimbal servosystem.

3. Reference Model Selection

The schematic diagram of MRAC is shown in Figure 6. is the input of reference model and controlled object. The controlled object output is the actual response, and the reference output is the expected response of system. Define generalized error . The adaptive controller adapts the parameters of and based on generalized error . At last, reaches to 0. That is, controlled object output has kept up with reference model output. The design of model reference controller mainly contains reference model selection and adaptive controller design. The design of adaptive controller will be introduced in Section 4. The reference model is selected according to the requirement for dynamic performance and the structure characteristics of controlled object.

897579.fig.006
Figure 6: Schematic diagram of MRAC.
3.1. Controlled Object Model

In order to keep the fast dynamic response of current loop, the current loop and controlled object are treated as the generalized controlled object (dashed line frame of Figure 5). Only the speed loop is replaced with adaptive controller. The transfer function of the generalized controlled object is where

From (5), the controlled object is a fourth-order system. According real system parameters, the fourth-order system can be reduced to third-order system. The process of reducing order is as follows.

Define the electromagnetic time constant of torque motor by

Define the electromechanical time constant of torque motor by

According to Table 1, (7) and (8) could be calculated:  s,   s. Due to , the effect of to torque motor can be ignored, that is, . Because that damping coefficient is small and the gimbal angle velocity is also small (less than 0.1745 rad/s), the friction damping torque has little influence on gimbal system. So can be also ignored, that is, . Thence, (5) can be simplified as (9) which is a third-order system: where

tab1
Table 1: System physical parameters.

In order to observe the physical significance of (9) and select the reference model easily, (9) is needed to be further described.

Define the mechanical resonance-frequency by

Define the generalized electromechanical time constant by

Define the equivalent time constant by

According to (11)–(13), (9) could be rewritten as

Due to the small value of , so

Then (14) could be further described by where is damp ratio.

Table 1 shows the system physical parameters. According to the experimental results, the current-loop parameters are ,  .

According to (9) and the parameters in Table 1, the generalized controlled object is

Equation (17) is discredited by adding zero-order hold, with sampling time ,

3.2. Reference Model

According to (16), the system reference model is

The reference model contains a first-order inertial system and a second-order oscillation system :

According to the dynamic performance requirement and the structure characteristics of controlled object, the parameters of (19) are determined: ,  ,  and  .

The reference model is described as

Equation (21) is discredited by adding zero-order hold, with sampling time ,

4. Adaptive Controller Design Based on POPOV Super Stable Theory

4.1. POPOV Super Stable Theory

As shown in Figure 7, a nonlinear feedback control system consists of a forward block and a nonlinear time-variable block. The linear time-invariant block is The nonlinear time-variable block is Define the discrete POPOV integral inequality by

897579.fig.007
Figure 7: Nonlinear feedback control system.

Discrete POPOV super stable theory is described as follows. If the control system, shown in Figure 7, meets the following conditions (a) and (b), the control system is globally asymptotic stable.(a)The linear time-invariant block is strictly positive definite.(b)The nonlinear time-variable block satisfies the discrete POPOV integral inequality (25).

4.2. Adaptive Control Law Design

Figure 8 presents the structure of discrete MRAC system used in this paper. According to (18) and (22), the difference equation of controlled object can be described as

897579.fig.008
Figure 8: Structure of discrete MRAC system.

The difference equation of reference model can be described as

,  ,  , and in (26) and (27) can be described as

The values of ,  ,  , and are defined in (18) and (22).

From Figure 8, the following equation can be obtained: where and are, respectively, feedback and feedforward multinomial,

is linear-compensator multinomial

Error is defined as

Substituting (29) into (27) and then subtracting (26), error equation is given as where

Considering the linear-compensator multinomial , the generalized error equation is

As shown in Figure 7, is the linear time-invariant block and is the nonlinear time-variable block.

According to (18), the linear time-invariant block can be rewritten as

According to the POPOV super stable theory, the linear time-invariant block (36) should be strictly positive definite. So the coefficients of can be determined by When and are chosen as (38), the nonlinear time-variable block will satisfy the discrete POPOV integral inequality (25), which is proved at the end of this section: where ; , ; , .

As shown in Figure 8, the controller output can be described as

Now the proof of (38) meeting the discrete POPOV integral inequality is presented as follows.

The nonlinear time-variable block is given as

Equation (38) can be rewritten as

Substituting (41) into (40), is rewritten as

Substituting (42) into (25), the POPOV integral inequality is given as

In order to analyze (43), the following equation is introduced:

According to (44), (43) meets the flowing inequality

Due to , ,  and , (45) can be further described as

Equation (46) is the POPOV integral inequality. So (38) meeting the discrete POPOV integral inequality is proved.

5. Simulation Analyses

Simulation analyses are completed in Simulink of MATLAB 7.5. According to Figure 8, the simulation model is built. In the model, the adaptive controller is designed based on (22), (38)-(39), and the generalized controlled object is built based on Figure 5. The simulation parameters for adaptive controller are ,  ,  ,  , and  .

When the speed given to the system is square-wave, the output speed wave is shown in Figure 9, and the waves of the adaptive control parameters and are shown in Figure 10. As shown in Figures 9 and 10, after the second period, the controlled object output could keep up with reference model output, and the parameters and could also converge. The simulation results indicate that the adaptive control law proposed in this paper is feasible and effective.

897579.fig.009
Figure 9: System response with square-wave speed given.
fig10
Figure 10: The curve of adaptive control parameters.

6. Experimental Results

The DGCMG test system is originally developed by the Beijing University of Aeronautics and Astronautics in China. The experimental system is presented in Figure 11, and the schematic diagram is shown in Figure 12. The control system consists of power drive unit and signal processing unit. The proposed control algorithm is implemented in DSP TMS320C31 with a current-loop period of 200 s and a speed-loop period of 1 ms. The control system receives speed given from the monitor computer through CAN bus. In order to observe the experimental results, the speed signals are displayed on the oscilloscope. The resolution of photoelectric encoder is 1.24′′.

897579.fig.0011
Figure 11: View of DGCMG test system.
897579.fig.0012
Figure 12: Schematic diagram of DGCMG test system.

According to the simulation results in Section 5, the parameters for adaptive controller are ,  ,  ,  ,  ,  ,  and  .

In order to compare with MRAC, a PID controller is also adopted. As shown in Figure 5, the controllers of current and speed loops are all PID controllers. The current-loop parameters are and . The proportional and integral coefficients for speed loop are , . The speed feedforward coefficient is . The bandwidth and continuous-step tests are carried out to compare the performance of adaptive controller with that of PID controller.

6.1. Bandwidth Test

The speed given is , where = 10°/s. When the speed frequency is changed from 1 Hz to 3 Hz, the sine responses with PID controller and MRAC are shown in Figures 13 and 14.

897579.fig.0013
Figure 13: Sine response with PID controller, when the speed frequency is changed from 1 Hz to 3 Hz.
897579.fig.0014
Figure 14: Sine response with MARC, when the speed frequency is changed from 1 Hz to 3 Hz.

When the speed frequency is 1 Hz, the tracking effect of PID controller is the same as that of MRAC. But when the speed frequency increases to 3 Hz, the tracking results of MRAC are obviously better than that of PID controller which has large overshoot and phase lag. And as shown in Figure 14, when speed frequency changes from 1 Hz to 3 Hz suddenly, the adaptive controller could make the object-output track the new frequency quickly. According to the experimental results, the system bandwidth is 2 Hz with PID controller and 3 Hz with MRAC.

6.2. Continuous Step-Response Test

Because the control period of real attitude control system is far less than that of the gimbal servosystem, the speed given to the gimbal is not continuous. So the continuous step-response test is carried out to verify the performance of gimbal system. The speed is continuously given by 0.1°/s, 0.5°/s, 2°/s, 5°/s, 8°/s, and 10°/s, with the time interval of 500 ms. The experimental results are shown in Figures 15 and 16.

897579.fig.0015
Figure 15: Continuous step-response with PID controller.
897579.fig.0016
Figure 16: Continuous step-response with MRAC.

As shown in Figure 15, PID controller has poor dynamic performances with obvious noises and large overshoot which can reach to 40%. And Figure 16 shows that adaptive controller has small overshoot and smooth response.

The experimental results of continuous step-response show that MRAC could improve the dynamic performance with small overshoot and make the system work smoothly without obvious noises.

7. Conclusion

The DGCMG gimbal servosystem directly decides the performance of gyro moment. In order to improve the poor dynamic performance of gimbal servosystem caused by the low and nonlinear torsional stiffness of harmonic drive, a model reference adaptive control strategy is introduced in this paper. Because only the input, output, and error are used and the measurement of the torque is not required, the proposed method is simple. In order to keep the fast dynamic response of current loop, the current loop and controlled object are treated as the generalized controlled object. The reference model is selected according the dynamic performance with fast response and small overshoot. The discrete POPOV super stable theory is used to design the adaptive controller. The model of DGCMG gimbal servosystem with harmonic drive is established. The MATLAB simulation results are provided to verify the effectiveness of the proposed control algorithm. The experimental results indicate that the MRAC could increase the bandwidth of gimbal servosystem to 3 Hz and improve the dynamic performance with small overshoot. And this method also has good engineering value.

Symbols

R:Stator winding resistance of torque motor
: Stator winding inductance of torque motor
: Stator voltage of torque motor
: Back electromotive force (BEMF) coefficient of torque motor
: Torque coefficient of torque motor
:Rotational inertia of motor terminal including flexspline and toque motor rotor
:Rotational inertia of load terminal including high-speed rotor and gyro-room
:Damping coefficient of motor terminal including harmonic drive friction damping
:Damping coefficient of load terminal
:Angle position of motor terminal
:Angle position of load terminal
: Torsion angle of harmonic drive
:Reference input of current loop
:Stator current of torque motor
:Output torque of torque motor
:Output  torque of harmonic drive
:External  disturbing  torque including coupling gyro moment
: Torsional stiffness of harmonic drive
N: Reduction ratio of harmonic drive
: Amplification coefficient of current sampling circuit
: Amplification coefficient of power driver circuit
: Feedforward coefficient of speed loop
: Proportionality coefficient of current loop
: Feedback coefficient of current loop.

Acknowledgments

This work was supported in part by the National Outstanding Youth Fund of China under Grant 60825305 and in part by the Aviation Science Fund of China under Grant 2012ZB51019.

References

  1. S. P. Bhat and P. K. Tiwari, “Controllability of spacecraft attitude using control moment gyroscopes,” IEEE Transactions on Automatic Control, vol. 54, no. 3, pp. 585–590, 2009. View at Publisher · View at Google Scholar
  2. D. J. Krebs, “Model correlation of International Space Station Control Moment Gyroscope,” in Proceedings of the 39th AIAA/ASME/ASCE/AHS/ASC Structures, Structural Dynamics, and Materials Conference and Exhibit and AIAA/ASME/AHS Adaptive Structures Forum, pp. 813–818, April 1998, AIAA-98-1793. View at Scopus
  3. J. Ahmed, “Adaptive control of double-gimbal control moment gyro with unbalanced rotor,” Journal of Guidance, Control, and Dynamics, vol. 25, no. 1, pp. 105–115, 2002. View at Publisher · View at Google Scholar
  4. C. J. Heiberg, “A practical approach to modeling singal-gimbal control momentum gyroscope in agile spacecraft,” in Proceedings of AIAA Guidance, Navigation, and Control Conference and Exhibit, pp. 1–11, 2000, AIAA-2000-4545.
  5. J. C. Fang and Y. Ren, “High-precision control for a single-gimbal magnetically suspended control moment gyro based on inverse system method,” IEEE Transactions on Industrial Electronics, vol. 58, no. 9, pp. 4331–4341, 2001. View at Google Scholar
  6. J. Fang, S. Zheng, and B. Han, “AMB vibration control for structural resonance of double-gimbal control moment gyro with high-speed magnetically suspended rotor,” IEEE/ASME Transactions on Mechatronics, vol. 18, no. 1, pp. 32–43, 2013. View at Publisher · View at Google Scholar · View at Scopus
  7. T. D. Tuttle and W. P. Seering, “A nonlinear model of a harmonic drive gear transmission,” IEEE Transactions on Robotics and Automation, vol. 12, no. 3, pp. 368–374, 1996. View at Google Scholar · View at Scopus
  8. F. H. Ghorbel, P. S. Gandhi, and F. Alpeter, “On the kinematic error in harmonic drive gears,” Journal of Mechanical Design, vol. 123, no. 1, pp. 90–97, 2001. View at Google Scholar · View at Scopus
  9. P. S. Gandhi and F. H. Ghorbel, “Closed-loop compensation of kinematic error in harmonic drives for precision control applications,” IEEE Transactions on Control Systems Technology, vol. 10, no. 6, pp. 759–768, 2002. View at Publisher · View at Google Scholar · View at Scopus
  10. R. Dhaouadi, F. H. Ghorbel, and P. S. Gandhi, “A new dynamic model of hysteresis in harmonic drives,” IEEE Transactions on Industrial Electronics, vol. 50, no. 6, pp. 1165–1171, 2003. View at Publisher · View at Google Scholar · View at Scopus
  11. G. Zhang and J. Furusho, “Speed control of two-inertia system by PI/PID control,” IEEE Transactions on Industrial Electronics, vol. 47, no. 3, pp. 603–609, 2000. View at Publisher · View at Google Scholar · View at Scopus
  12. K. Szabat and T. Orlowska-Kowalska, “Vibration suppression in a two-mass drive system using PI speed controller and additional feedbacks—comparative study,” IEEE Transactions on Industrial Electronics, vol. 54, no. 2, pp. 1193–1206, 2007. View at Publisher · View at Google Scholar · View at Scopus
  13. S. E. Saarakkala, M. Hinkkanen, and K. Zenger, “Speed control of two-mass mechanical loads in electric drives,” in Proceedings of the IEEE Energy Conversion Congress and Exposition (ECCE '12), pp. 1246–1253, 2012.
  14. H. Chaoui and P. Sicard, “Adaptive neural network control of flexible-joint robotic manipulators with friction and disturbance,” in Proceedings of the 38th Annual Conference on IEEE Industrial Electronics Society (IECON '12), pp. 2644–22649, 2012.
  15. W. H. Zhu, E. Dupuis, M. Doyon, and J. C. Piedboeuf, “Adaptive control of harmonic drives based on virtual decomposition,” IEEE/ASME Transactions on Mechatronics, vol. 11, no. 5, pp. 604–614, 2006. View at Publisher · View at Google Scholar · View at Scopus
  16. L. Tian and A. A. Goldenberg, “Robust adaptive control of flexible joint robots with joint torque feedback,” in Proceedings of the IEEE International Conference on Robotics and Automation, pp. 1229–1234, May 1995. View at Scopus
  17. I. Hassanzadeh, H. Kharrati, and J. R. Bonab, “Model following adaptive control for a robot with flexible joints,” in Proceedings of the IEEE Canadian Conference on Electrical and Computer Engineering (CCECE '08), pp. 1467–1472, May 2008. View at Publisher · View at Google Scholar · View at Scopus
  18. Z. M. Doina, “LQG/LQR optimal control for flexible joint manipulator,” in Proceedings of the International Conference and Exposition on Electrical and Power Engineering (EPE '12), pp. 35–40, October 2012.
  19. C. Trautman and D. Wang, “Experimental H control of a single flexible link with a shoulder joint,” in Proceedings of the IEEE International Conference on Robotics and Automation, vol. 1, pp. 1235–1241, May 1995. View at Scopus
  20. K. H. Kim, “Model reference adaptive control-based adaptive current control scheme of a PM synchronous motor with an improved servo performance,” IET Electric Power Applications, vol. 3, no. 1, pp. 8–18, 2009. View at Publisher · View at Google Scholar · View at Scopus
  21. S. Zhang and Y. Pi, “Model reference adaptive control application study in PM synchronous motor servo system,” in Proceedings of the 2nd Conference on Power Electronics and Intelligent Transportation System (PEITS '09), pp. 428–432, December 2009. View at Publisher · View at Google Scholar · View at Scopus
  22. L. Yacoubi, K. Al-Haddad, L. A. Dessaint, and F. Fnaiech, “A DSP-based implementation of a nonlinear model reference adaptive control for a three-phase three-level NPC boost rectifier prototype,” IEEE Transactions on Power Electronics, vol. 20, no. 5, pp. 1084–1092, 2005. View at Publisher · View at Google Scholar · View at Scopus
  23. K. Y. Huang, H. C. Chin, and Y. C. Huang, “A model reference adaptive control strategy for interruptible load management,” IEEE Transactions on Power Systems, vol. 19, no. 1, pp. 683–689, 2004. View at Publisher · View at Google Scholar · View at Scopus
  24. D. H. Lee and J. W. Ahn, “A current ripple reduction of a high-speed miniature brushless direct current motor using instantaneous voltage control,” IET Electric Power Applications, vol. 3, no. 2, pp. 85–92, 2009. View at Publisher · View at Google Scholar · View at Scopus