Next Article in Journal
Semantic Segmentation with Transfer Learning for Off-Road Autonomous Driving
Next Article in Special Issue
Session Recommendation via Recurrent Neural Networks over Fisher Embedding Vectors
Previous Article in Journal
Ski Position during the Flight and Landing Preparation Phases in Ski Jumping Detected with Inertial Sensors
Previous Article in Special Issue
Contrast and Homogeneity Feature Analysis for Classifying Tremor Levels in Parkinson’s Disease Patients
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Neural Network Direct Control with Online Learning for Shape Memory Alloy Manipulators

by
Alfonso Gómez-Espinosa
1,*,
Roberto Castro Sundin
2,
Ion Loidi Eguren
3,
Enrique Cuan-Urquizo
1,* and
Cecilia D. Treviño-Quintanilla
1
1
Tecnologico de Monterrey, Escuela de Ingeniería y Ciencias, Ave. Epigmenio González 500, Fracc. San Pablo, Querétaro 76130, Mexico
2
KTH Royal Insitute of Technology, 114 28 Stockholm, Sweden
3
Escuela Politécnica Superior, Universidad Mondragón, 20500 País Vasco, Spain
*
Authors to whom correspondence should be addressed.
Sensors 2019, 19(11), 2576; https://doi.org/10.3390/s19112576
Submission received: 12 April 2019 / Revised: 2 June 2019 / Accepted: 4 June 2019 / Published: 6 June 2019
(This article belongs to the Special Issue Artificial Intelligence and Sensors)

Abstract

:
New actuators and materials are constantly incorporated into industrial processes, and additional challenges are posed by their complex behavior. Nonlinear hysteresis is commonly found in shape memory alloys, and the inclusion of a suitable hysteresis model in the control system allows the controller to achieve a better performance, although a major drawback is that each system responds in a unique way. In this work, a neural network direct control, with online learning, is developed for position control of shape memory alloy manipulators. Neural network weight coefficients are updated online by using the actuator position data while the controller is applied to the system, without previous training of the neural network weights, nor the inclusion of a hysteresis model. A real-time, low computational cost control system was implemented; experimental evaluation was performed on a 1-DOF manipulator system actuated by a shape memory alloy wire. Test results verified the effectiveness of the proposed control scheme to control the system angular position, compensating for the hysteretic behavior of the shape memory alloy actuator. Using a learning algorithm with a sine wave as reference signal, a maximum static error of 0.83° was achieved when validated against several set-points within the possible range.

1. Introduction

Nonlinear systems have been an active research area over the past few decades, partly fueled by the needs of modern industry. As new actuators and materials are incorporated into industrial processes, additional challenges can be posed by their sometimes complex behavior. An example of this is the nonlinear hysteresis commonly found in shape memory alloys (SMAs) [1], certain nanocomposite materials [2], and micro-electromechanichal systems (MEMS) [3].
In applications where hysteresis behavior is present, the inclusion of a suitable hysteresis model in the control system design allows the controller to have better tracking of the system and may result in a reduced error response of the controlled variable [4,5,6]. The methods that have been used to model these behaviors are many, and include artificial neural networks (ANNs) [4], elliptical approximations of hysteresis loops [5], the Preisach model [6], ideal order hexagonal arrays adjusted using the Monte Carlo method [7], a modified Prandtl–Ishlinskii Model [8], and the Semilinear Duhem Model [9].
Although the use of a suitable model allows a better control of a hysteretic system, a major drawback is that each system responds in a unique way, meaning that the controllers characteristics, such as effectiveness, speed, number of cycles, etc., may vary from one system to another, even if the control techniques are the same [10,11]. As an example, second-order models have been used to analyse and model output signal sensors. However, this process is time-consuming, and when completed, the resulting algorithm can only be applied successfully to the system it was specifically designed for [10]. Another previously used method for approximating and tracking the performance of a system, is to make use of ANNs. This has been done, for instance, for an inverted pendulum with an actuator displaying hysteretic behavior [11], and for an active magnetic bearing system [12]. Although ANNs are still under development in order to increase their accuracy in control system applications, they are frequently implemented due to their low computational cost [13].
Neural networks (NNs) have demonstrated their ability to identify, and compensate for, hysteresis in systems in which high precision is mandatory [14,15,16,17]; an adaptive wavelet NN controller with estimation of the friction force hysteresis was proposed for a piezo positioning mechanism [14], an ANN to identify, and compensate for, gear backlash hysteresis was tested for a precision position mechanism [15], a radial basis function neural network and a sliding control scheme was proposed for motion tracking control of piezoelectric actuators [16], and a dynamic recurrent NN with a proportional-derivative (PD) controller was implemented to identify and control a magneto restrictive actuator [17].
SMAs are recognized as promising smart materials for creating compact, high power-to-weight ratio actuators. However, position control of SMA actuators represents a major challenge for practical applications due to their nonlinear hysteretic behavior, producing steady-state errors when conventional controllers are used [18]. To overcome this problem, proportional-integral-derivate (PID) controllers with hysteresis models have been implemented [18,19,20]; for position control of SMA actuators using the generalized Prandtl–Ishlinskii inverse model [18], for micro-positioning control of SMA actuators by modeling the hysteresis using NNs [19], and for magnetic SMA actuators by using a radial basis function NN to obtain the Jacobian information of the system in order to adjust the controller parameters [20]. In addition, neural network controllers, previously trained to identify the system were proposed to control SMA actuators [21,22,23,24]; using the inverse of the ANN that replicate the dynamics of the SMA force actuator to implement the controller [21], implementing a model predictive controller based on a functional link ANN to control the linear memory metal actuator displacement [22], and realizing a recurrent neural model predictive, variable structure, controller designed to control a one degree of freedom (1-DOF) rotary manipulator actuated by an SMA wire [23,24]. The main disadvantages of these recurrent neural network controllers are the complexity of the control system, and the requirement of training the neural network, to identify the system, prior to the implementation of the control. During the identification of the system, a neural network captures the model of the plant. Later this plant model is utilized for the design of the controller. The control signal is only delivered to the plant, after these two stages are completed. Previous inputs/outputs data are gathered from the operation of the plant, and are utilized to train the neural network model, offline in batch mode [23]. Previous information from the controller outputs and from the plant outputs are the inputs utilized for the neural plant model, implementing a recurrent neural network structure [24].
In this work, a neural network direct control, with online learning, is developed for position control of shape memory alloy manipulators. The developed network consists of three inputs, four hidden layer perceptrons and one output, all using a sigmoidal activation function. In contrast to previous works, weight coefficients are updated online by using the actuator position data while the controller is applied to the system, without previous training of the neural network weights nor the inclusion of a hysteresis model. A real-time, low computational cost control system was implemented. Experimental evaluation was performed on a 1-DOF manipulator system actuated by a shape memory alloy wire and test results verify the effectiveness of the proposed control scheme to control the system angular position, compensating for the hysteretic behavior of the shape memory alloy actuator.
The remainder of the paper is organized as follows: Hysteresis nonlinearity is described in Section 2.1. Shape memory alloy nonlinear behavior is introduced in Section 2.2. Section 2.3 presents the NN direct controller with online learning using back-propagation. Section 3 describes the experimental setup and the results are presented in Section 4. Finally, in Section 5, concluding remarks are provided.

2. Materials and Methods

2.1. Hysteresis

Hysteresis is a strongly nonlinear phenomena, where strongly indicates that it cannot be linearized. The non-ability of being linearized is a consequence of the memory effect in hysteretic systems, which means that the output of the system is dependent not only on the current input, but also on the previous state of the system [6,9]. Hysteresis can be found in many different areas, including structural mechanics, aerodynamics, and electromagnetics [9]. Because of the memory effect, an input–output mapping for a system with hysteresis seldom is injective, since one input often may result in two distinct outputs depending on the system history (illustrated in Figure 1). This of course implicates that hysteresis cannot be modeled in the sense of an ordinary mathematical function, but requires a more sophisticated framework. One of these frameworks was proposed by Preisach as a means to model the hysteresis found in magnets, and was further improved upon by mathematician Krasnoselskii to become what is now referred to as the Preisach model [25].

The Classic Preisach Model

The model is most often represented by the Preisach operator Γ ^ , which can be expressed as a double integral of variables α , β
Γ ^ = α β μ ( α , β ) γ ^ α β d α d β ,
where μ ( α , β ) is a weight function called the Preisach function, and γ ^ α β is a fundamental hysteresis operator, often called hysteron. The hysteron is a mapping onto {−1,1} with memory properties as seen in Figure 2 [26]. Using the Preisach operator, systems with hysteresis, like the one in Figure 1, can now be modeled by finding the correct weight function μ ( α , β ) .

2.2. Shape Memory Alloys

Shape memory alloys, or SMAs, receive their name from their property of exerting a force to return to a certain memorized shape when heated, giving them the ability to convert heat into mechanical energy [23]. This effect is due to SMAs having two solid phases; a high temperature phase called austenite and a lower temperature phase called martensite, differentating themselves by having different crystal structures [24]. The austenite form is typically cubic and rigid, while the martensite usually is tetragonal, orthorhombic or monoclinic, giving it the possibility to change shape (e.g., stretch) upon being subjected to a force. Because of this, it is common to distinguish between the unstretched twinned martensite form and the stretched untwinned one. The relationship between the different phases can be seen in the stress–temperature-plane in Figure 3, for which it is also important to mention that the stress is sufficiently small to allow austenite form even under stress conditions when temperature is increased (seen in case (d)). For higher stress levels the austenite form could not be achieved even when increasing the temperature, but would rather stay in its detwinned martensite form (not shown in this figure) [27].
The transition between the untwinned martensite form and the austenite is highly hysteretic, as shown by the experimental results for a Flexinol SMA wire in Figure 4. The wire was 310 mm long, Ø 0.13 mm and had a 50 g weight suspended to it.

2.3. Neural Networks

2.3.1. Feedforward Neural Networks with Sliding Window

In accordance with [28], a multilayer recurrent neural network is an extension of the classic perceptron that allows for more possibilities when it comes to the desired behavior, being able to approximate dynamic systems. The increased functionality comes from incorporating a number of so-called hidden layers in between the input and output layers, from a learning process called back-propagation and from making use of previous input/output values (i.e., recurrent neural network). Recurrent neural networks can approximate dynamic systems, represented by differential equations, however, feed forward neural networks can only approximate algebraic functions. Conventional recurrent neural networks incorporate temporal information in the hidden layer, by using previous output values as inputs for this same layer. On the other hand, feedforward neural networks with sliding windows can be implemented using previous input/output values as inputs for the input layer.
A hidden layer consists of a certain amount of perceptrons h 1 , , h m , each one connected to all of the perceptrons (or inputs) h 1 , , h n in the preceding layer. Each of these connections are weighted by the coefficients w j i , where subscripts j , i indicate the connection between perceptrons/inputs h j and h i . A general illustration is shown in Figure 5, where the input layer uses present values as well as previous values z 1 , z 2 , z 3 for a feedforward neural network structure with a sliding window.
For a neural network with one hidden layer and a single output, the inputs can be defined as x i , the hidden layer perceptrons as h j and the weights as follows
w j i  
  the weight of the connection between input x i and hidden layer perceptron h j
v j  
  the weight of the connection between hidden layer perceptron h j and output perceptron.
it has been regarded as items, please confirm

2.3.2. Back-Propagation

Following the methodology of [28,29], the back-propagation algorithm can be used to train a multilayer neural network into reaching a desired behavior. This is done by iteratively changing the weights until the error is reduced to an acceptable level. A solution is thus any collection of weights w j i , v j that are able to achieve this. The back-propagation algorithm finds these weights by making use of the method of gradient descent. Using gradient descent the aim is to minimize the error function
E = 1 2 i | e i | 2 = 1 2 i | y r i y i | 2 .
The gradient E is only defined if E is continuous and differentiable, and thereby, the sigmoid s c : R ( 0 , 1 ) given by
s c ( x ) = 1 1 + e c x ,
will be used as activation function, where c can be chosen arbitrarily and decides the steepness of the curve.
As a means of increasing the flexibility of the activation function, a so-called bias, θ , can be used. The role of the bias is to shift the activation function s c ( x ) in the positive or negative direction of the x-axis. A simple way of adding the bias is to expand the input vector ( x 1 , , x n ) with 1, to get ( x 1 , , x n , 1 ) and the weight vector w = ( w 1 , , w n ) with θ , to get w = ( w 1 , , w n , θ ) . This permits the comfortable notation of a bias-shifted perceptron as
s c ( i w i · x i θ ) = 1 1 + exp [ c ( i w i · x i θ ) ] = s c ( w · x ) .

2.3.3. Calculation of the Gradient

For a neural network with one hidden layer and a single output, the gradient can be calculated in a straightforward manner using chain rule and the configuration of the control system, as seen in Figure 6 [28].
Since the weights w j i , v j are the ones to be adjusted, the gradient is calculated with respect to these. Therefore, the gradient is defined
E = E v j E w j i ,
with its partial derivatives
E v j = E e y e y e u e u u u r r v j = e y u ( 1 u ) e y e u h j
E w j i = E e y e y e u e u u u r r h j h j S j S j w j i = e y u ( 1 u ) v j h j ( 1 h j ) e y e u x i ,
where the relations
r = j v j h j
S j = i w j i x i ,
along with the definitions seen in Figure 5 and Figure 6 have been used, and where e u denotes the error between the current control signal and the control signal required to control the system. Note that, with the lack of a mathematical model of the system, the error e y cannot be expressed in analytical terms. Consequently, its partial derivative ∂ey/∂eu found in Equations (6) and (7), is unknown.
With the gradient explicitly calculated, the weights can now be adjusted iteratively as follows
v j ( t + 1 ) = v j ( t ) η E v j = v j ( t ) + η h j δ 1 e y e u
w j i ( t + 1 ) = w j i ( t ) η E w j i = w j i ( t ) + η x i δ j 2 e y e u
where η represents the learning rate and δ 1 = e y u ( 1 u ) , δ j 2 = δ 1 v j h j ( 1 h j ) .
Considering that
e y e u = sgn ( e y e u ) · | e y e u |
and letting η·|∂ey/∂eu|→η, Equations (10) and (11) simplify to
v j ( t + 1 ) = v j ( t ) + sgn ( e y e u ) η h j δ 1
w j i ( t + 1 ) = w j i ( t ) + sgn ( e y e u ) η x i δ j 2 ,
where sgn(∂ey/∂eu), in accordance with [29], can be found experimentally for the system.
The previously described neural networks suffice when we seek to approximate static functions. However, when the aim is to approximate a dynamic function, it is necessary to include the concept of time. This can be done by incorporating previous input/output values. For example, if u ( t ) is the output at a time t, then an arbitrarily large collection of previous output values u ( t T ) , u ( t 2 T ) , , u ( t k · T ) , where T is the discretization time step and k N , can be tracked. Using these previous output/input values as inputs, the neural network can obtain further understanding of the system dynamics and, for instance, know whether the output is increasing or decreasing, if the time derivate of the output is increasing or decreasing, etc.

3. Experimental Setup

A 1-DOF manipulator actuated by a Ø 0.13 mm Flexinol wire (DYNALLOY, Inc., 1562 Reynolds Ave., Irvine, CA, USA) was controlled using a pulse-width modulated current with a maximum of 230 mA that was fed by an Agilent e3631A power supply (5301 Stevens Creek Blvd., Santa Clara, CA, USA) set to an 8 V limit. The system controller consisted of a National Instruments myRIO-1900 (National Instruments Corporation, 11500 North Mopac Expy, Austin, TX, USA), whose 12-bit PWM output was connected to all eight amplifiers of a ULN2803 Darlington transistor array (Toshiba, 1-1, Shibaura 1-chome, Minato-ku, Tokyo, Japan) and a PDB181-K420K-103B potentiometer (Bourns, Inc., 1200 Columbia Ave., Riverside, CA, USA). Several controlled parameters of the setup, including the weight of the manipulator arm and the maximum current allowed, were set so as to allow for repeatability, ensuring that the Flexinol wire was deformed reversibly [27]. With the current setup it was verified that this meant keeping the maximum strain below 4%. In our case, no two-way shape memory effect (SME) was observed and it was thus necessary to apply stress to the Flexinol wire in order to achieve any deformation.
To ensure the correctness of the angular position of the 1-DOF manipulator, the system was calibrated using an OptiTrack–Motive (NaturalPoint, Inc., 3658 SW Deschutes Street, Corvallis, OR, USA) motion capture system in a 16-cameras setup at a 240 Hz sampling rate (Appendix A). Figure 7 shows a photograph of the experimental setup. The dimensions of the manipulator are included in Appendix B.
The output signal from the potentiometer was fed into a feedback loop through the myRIO-1900 analog input with 12 bits of resolution, as seen in Figure 6. The complete experimental setup circuit is illustrated in Figure 8.
The wire used as an actuator was made of Flexinol—a nickel-titanium alloy that exhibits SMA properties. Information from the manufacturer is seen in Table 1.
The software used to program the myRIO-1900 was NI LabVIEW 2017 (National Instruments Corporation, 11500 North Mopac Expy, Austin, TX, USA). A feedforward neural network with a three element sliding window (initialized to zero), consisting of 3 inputs, 4 hidden layer perceptrons and one output perceptron, was implemented and trained online using the back-propagation algorithm, without any prior training or knowledge of the system. The three utilized inputs, which were the present and two previous values of the output system error, were computed as the difference between the reference output value y r to the measured output value y. Weights were adjusted for each iteration, until an acceptable error was reached. When no further adjustment was required, the learning procedure could be stopped. Optimization of the neural network architecture was made by performing experiments starting with a low number of hidden layer perceptrons and then increasing them one by one until no improvement of performance could be found in comparison to earlier experiments. The learning algorithm can be found in its entirety in Appendix C and is illustrated in the flowchart in Figure 9. The sample time of the controller was set to 20 ms and the PWM frequency to 60 Hz.

4. Results

Prior to the experiments, a qualitative characterization of the system was performed. The results, confirming its hysteretic behaviour, can be seen in Figure 10.
Throughout every performed experiment, the Flexinol wire diamater as well as its original length and the ambient temperature remained constant. This is important since any change could affect the system behaviour [30].
The neural network coefficients were adjusted online by sending the manipulator to different set-points within the possible range, starting with the learning coefficient η = 2 , a value deemed appropriate by trial and error. The set-point was changed as soon as a low-error value was achieved and η reduced by a factor of 2 each time. The reduction factor 2 was also found by trial and error and corresponded to a value for which the weights were updated just enough to learn the new set-point without affecting the previously learnt ones. Finally, η was set to zero, thereby terminating the learning process. Online learning allows the controller to continue improving its performance for additional set-points and adapting to changes produced by system degradation and external disturbances. The learning process was terminated to evaluate the capability of the controller, limited to the knowledge acquired during the online learning, when being tested for set-points changes and load disturbances. It should, however, also be pointed out that the controller performed just as well—or even better—when η was kept bigger than zero, thus never terminating the learning. Results during a learning session, for the reference angular position y r , the measured angular position y and the output system error (marked beneath in red), are presented in Figure 11.
With η set to zero, the response of the system was verified for set-points −12°, −1°, 10°, and 20°, as seen in Figure 12. The maximum static error achieved was 1.28°.
A second learning procedure was implemented using a 0.01 Hz sine wave—ranging from −20° to 31°—as reference signal during learning. The learning coefficient η was primarily set to a value of 4 and gradually lowered to zero as the system started responding better to the reference signal. The learning process is illustrated in Figure 13. A validation done in the previous manner for set-points −12°, −1°, 10°, 20°, 31° can be seen in Figure 14 and resulted in a maximum static error of 0.83° for the set-point −1°. This error is reduced to 0.49° if we exclude the aforementioned set-point.
Further validation of the system performance after learning was made using triangular and sine wave forms as reference signals. The triangle wave form was set to a frequency of 0.04 Hz with an amplitude of 53° and the results are shown in Figure 15. The sinusoidals were set to an amplitude of 42° and to frequencies 0.005 Hz, 0.01 Hz and 0.02 Hz. Figure 16a shows the result for the learning frequency of 0.01 Hz and displays no significant phase lag. In Figure 16b the result for the 0.005 Hz sine wave is shown, neither displaying any significant signs of phase lag. Figure 17 shows the result for the 0.02 Hz sine wave, where a 0.3 s phase lag can be identified.
The effect of disturbances on the system was investigated by a sudden application of additional torque on the arm that was removed once the system had stabilized. This was done first by applying an extra 86% of torque (Figure 18a), and later an extra 143% (Figure 18b). In both cases, the perturbation was compensated for by the controller without additional learning to adjust the weights.

5. Conclusions

A real-time, low computational cost, neural network direct control, with online learning, was developed for position control of shape memory alloy manipulators. Neural network weight coefficients were updated online by using the sensor position data while the controller was applied to the system, without previous training of the neural network weights nor the inclusion of a hysteresis model. Experimental evaluation was performed on a 1-DOF manipulator system actuated by a shape memory alloy wire and test results verify the effectiveness of the proposed control scheme to control the system angular position, compensating for the hysteretic behavior of the shape memory alloy actuator. Using a learning algorithm based on sending the actuator to distinct set-points within the possible range, a maximum static error of 1.28° was achieved when validating it against a staircase reference signal. This was later improved upon using a 0.01 Hz sine wave reference signal for learning, resulting in a maximum error of 0.83° for the same type of validation, or 0.49° for set-points above −12°. The effectiveness of the controller was also tested against a 0.04 Hz ramp and sine waves of frequencies 0.005-, 0.01- and 0.02 Hz without showing any severe signs of phase lag, with the worst case (0.02 Hz) presenting a lag of 0.3 s. The controller also showed its ability to compensate for load disturbances, when an additional torque of 143% was applied.
For future work, a detailed comparison with results for other existing techniques could be performed, including changing the control system and neural network architecture, as well as the utilized learning algorithm. In addition, a study of the effect of periodic disturbances on the system could be performed, as well as an investigation into if the system can be controlled without the use of sensors.

Author Contributions

Conceptualization, A.G.-E.; methodology, A.G.-E.; software, I.L.E. and R.C.S.; validation, R.C.S. and I.L.E.; formal analysis, A.G.-E., R.C.S. and I.L.E.; investigation, R.C.S., A.G.-E. and I.L.E.; resources, A.G.-E.; data curation, R.C.S.; writing—original draft preparation, R.C.S. and A.G.-E.; writing—review and editing, A.G.-E., E.C.-U., C.D.T.-Q. and R.C.S.; visualization, R.C.S.; supervision, A.G.-E.; project administration, A.G.-E.; funding acquisition, A.G.-E.

Funding

The authors would like to acknowledge the financial support of Tecnologico de Monterrey, in the production of this work.

Acknowledgments

We thank Fabiola Díaz Nieto who provided insight and expertise that greatly assisted on National Instrument hardware and LabVIEW programming. The authorsalso wish to thank Alethia Jocelyn Delgado De la Paz for the 1-DOF manipulator 3D modeling, Guillermo Salvador Barrón Sánchez for the angular position calibration using the motion capture system, Arturo Gómez Sierra for the photography artwork, and Erika Nataly Santana Aguilar, Rocío Morfín Díaz, Ramon Ariel Ivan Muñoz Corona, and Omar Enrique Trejo Díaz for their assistance during the earlier stages of this project.

Conflicts of Interest

The authors declare no conflict of interest.

Abbreviations

The following abbreviations are used in this manuscript:
(A)NN(artificial) neural network
1-DOFone degree of freedom
MEMSmicro-electromechanical systems
MLPmultilayer perceptron
PIDproportional-integral-derivative
PWMpulse-width modulation
SMAshape memory alloy

Appendix A. Motion Capture System

The setup of the motion capture system is seen in Figure A1 and Figure A2 and resulted in the calibration equation
y = 53 x 128 ,
with a linear correlation coefficient R = 0.9997 , where y is the angular position and x the potentiometer voltage.
Figure A1. The setup of the motion capture system.
Figure A1. The setup of the motion capture system.
Sensors 19 02576 g0a1
Figure A2. Setup of the reflective markers.
Figure A2. Setup of the reflective markers.
Sensors 19 02576 g0a2

Appendix B. 1-DOF Manipulator

Figure A3. Schematic showing the dimensions in millimeters of the 1-DOF manipulator used in experiments: (a) Perspective view; (b) Front view; (c) Side view.
Figure A3. Schematic showing the dimensions in millimeters of the 1-DOF manipulator used in experiments: (a) Perspective view; (b) Front view; (c) Side view.
Sensors 19 02576 g0a3

Appendix C. Back-Propagation

Listing 1: The back-propagation algorithm in pseudocode.
Listing 1: The back-propagation algorithm in pseudocode.
  • # SETUP
  • define vector "inputs" = [x_1,...,x_n]
  • define vector "correct_outputs" = [d_1,...,d_n]
  • define boolean "DONE" = FALSE
  • initialize float "eta" > 0
  • initialize float "tolerance" > 0
  • initialize random weights "w_ij" and "v_j"
  • initialize float "t" > 0
  •  
  • # TRAINING
  • while not DONE:
  •  |  for x_k in inputs and d_k in correct_outputs:
  •  |   |  # Calculate hidden layer activation
  •  |   |  h_j = sigmoid( sum_over_i(x_i∗w_ji) )
  •  |   |
  •  |   |  # Calculate output layer activation
  •  |   |  u   = sigmoid( sum_over_j(h_j∗v_j) )
  •  |   |
  •  |   |  # Let the system respond
  •  |   |  # before calculating error
  •  |   |  wait t seconds
  •  |   |
  •  |   |  # Calculate error
  •  |   |  e_y_k = x_k-d_k
  •  |   |
  •  |   |  # Update weights
  •  |   |  w_ji += eta∗e_y_k∗u∗(1-u)∗h_j
  •  |   |  v_j  += eta∗e_y_k∗u∗(1-u)∗h_j∗(1-h_j)∗v_j∗x_i
  •  |  end for
  •  |
  •  |  if e_y_k < tolerance for all k:
  •  |      DONE = TRUE
  • end while

References

  1. Zhou, M.; Zhang, Q. Hysteresis model of magnetically controlled shape memory alloy based on a PID neural network. IEEE Trans. Magn. 2015, 51, 1–4. [Google Scholar]
  2. Chinni, F.; Spizzo, F.; Montoncello, F.; Mattarello, V.; Maurizio, C.; Mattei, G.; Bianco, L.D. Magnetic hysteresis in nanocomposite films consisting of a ferromagnetic AuCo alloy and ultrafine Co particles. Materials 2017, 10, 717. [Google Scholar] [CrossRef] [PubMed]
  3. Tan, Q.; Kang, H.; Xiong, J.; Qin, L.; Zhang, W.; Li, C.; Ding, L.; Zhang, X.; Yang, M. A wireless passive pressure microsensor fabricated in HTCC MEMS technology for harsh environments. Sensors 2013, 13, 9896–9908. [Google Scholar] [CrossRef] [PubMed]
  4. Tu, F.; Hu, S.; Zhuang, Y.; Lv, J.; Wang, Y.; Sun, Z. Hysteresis Curve Fitting Optimization of Magnetic Controlled Shape Memory Alloy Actuator. Actuators 2016, 5, 25. [Google Scholar] [CrossRef]
  5. Nasiri-Zarandi, R.; Mirsalim, M. Finite-element analysis of an axial flux hysteresis motor based on a complex permeability concept considering the saturation of the hysteresis loop. IEEE Trans. Ind. Appl. 2016, 52, 1390–1397. [Google Scholar]
  6. Li, Z.; Zhang, X.; Su, C.Y.; Chai, T. Nonlinear control of systems preceded by Preisach hysteresis description: A prescribed adaptive control approach. IEEE Trans. Control Syst. Technol. 2016, 24, 451–460. [Google Scholar] [CrossRef]
  7. Vázquez, M.; Nielsch, K.; Vargas, P.; Velázquez, J.; Navas, D.; Pirota, K.; Hernandez-Velez, M.; Vogel, E.; Cartes, J.; Wehrspohn, R.; et al. Modelling hysteresis of interacting nanowires arrays. Phys. B Condens. Matter 2004, 343, 395–402. [Google Scholar] [CrossRef]
  8. Dong, R.; Tan, Y. A modified Prandtl–Ishlinskii modeling method for hysteresis. Phys. B Condens. Matter 2009, 404, 1336–1342. [Google Scholar] [CrossRef]
  9. Oh, J.; Bernstein, D.S. Semilinear Duhem model for rate-independent and rate-dependent hysteresis. IEEE Trans. Autom. Control 2005, 50, 631–645. [Google Scholar] [Green Version]
  10. Almassri, A.; Wan Hasan, W.; Ahmad, S.; Shafie, S.; Wada, C.; Horio, K. Self-calibration algorithm for a pressure sensor with a real-time approach based on an artificial neural network. Sensors 2018, 18, 2561. [Google Scholar] [CrossRef]
  11. Liu, Z.; Lai, G.; Zhang, Y.; Chen, X.; Chen, C.L.P. Adaptive neural control for a class of nonlinear time-varying delay systems with unknown hysteresis. IEEE Trans. Neural Netw. Learn. Syst. 2014, 25, 2129–2140. [Google Scholar] [PubMed]
  12. Liu, Z.; Lai, G.; Zhang, Y.; Chen, C.L.P. Adaptive neural output feedback control of output-constrained nonlinear systems with unknown output nonlinearity. IEEE Trans. Neural Netw. Learn. Syst. 2015, 26, 1789–1802. [Google Scholar]
  13. Fulginei, F.R.; Salvini, A. Neural network approach for modelling hysteretic magnetic materials under distorted excitations. IEEE Trans. Magn. 2012, 48, 307–310. [Google Scholar] [CrossRef]
  14. Lin, F.J.; Shieh, H.J.; Huang, P.K. Adaptive wavelet neural network control with hysteresis estimation for piezo-positioning mechanism. IEEE Trans. Neural Netw. 2006, 17, 432–444. [Google Scholar] [CrossRef] [PubMed]
  15. Seidl, D.R.; Lam, S.L.; Putman, J.A.; Lorenz, R.D. Neural network compensation of gear backlash hysteresis in position-controlled mechanisms. IEEE Trans. Ind. Appl. 1995, 31, 1475–1483. [Google Scholar] [CrossRef]
  16. Liaw, H.C.; Shirinzadeh, B.; Smith, J. Robust neural network motion tracking control of piezoelectric actuation systems for micro/nanomanipulation. IEEE Trans. Neural Netw. 2009, 20, 356–367. [Google Scholar] [CrossRef] [PubMed]
  17. Cao, S.; Wang, B.; Zheng, J.; Huang, W.; Weng, L.; Yan, W. Hysteresis compensation for giant magnetostrictive actuators using dynamic recurrent neural network. IEEE Trans. Magn. 2006, 42, 1143–1146. [Google Scholar]
  18. Sayyaadi, H.; Zakerzadeh, M.R. Position control of shape memory alloy actuator based on the generalized Prandtl–Ishlinskii inverse model. Mechatronics 2012, 22, 945–957. [Google Scholar] [CrossRef]
  19. Asua, E.; Etxebarria, V.; García-Arribas, A. Neural network-based micropositioning control of smart shape memory alloy actuators. Eng. Appl. Artif. Intell. 2008, 21, 796–804. [Google Scholar] [CrossRef]
  20. Zhou, M.; Zhang, Q.; Wang, J. Feedforward-feedback hybrid control for magnetic shape memory alloy actuators based on the Krasnosel’skii-Pokrovskii model. PLoS ONE 2014, 9, e97086. [Google Scholar] [CrossRef]
  21. Ghasemi, Z.; Nadafi, R.; Kabganian, M.; Abiri, R. Identification and Control of Shape Memory Alloys. Meas. Control 2013, 46, 252–256. [Google Scholar] [CrossRef]
  22. Tai, N.T.; Ahn, K.K. A hysteresis functional link artificial neural network for identification and model predictive control of SMA actuator. J. Process Control 2012, 22, 766–777. [Google Scholar] [CrossRef]
  23. Nikdel, N.; Nikdel, P.; Badamchizadeh, M.A.; Hassanzadeh, I. Using neural network model predictive control for controlling shape memory alloy-based manipulator. IEEE Trans. Ind. Electron. 2014, 61, 1394–1401. [Google Scholar] [CrossRef]
  24. Nikdel, N.; Badamchizadeh, M.A. Design and implementation of neural controllers for shape memory alloy–actuated manipulator. J. Intell. Mater. Syst. Struct. 2015, 26, 20–28. [Google Scholar] [CrossRef]
  25. Janičić, V.; Ilić, V.; Pjevalica, N.; Nikolić, M. An approach to modeling the hysteresis in ferromagnetic by adaptation of Preisach model. In Proceedings of the 2014 22nd Telecommunications Forum Telfor (TELFOR), Belgrade, Serbia, 25–27 November 2014; pp. 761–764. [Google Scholar]
  26. Wang, X.; Sun, T.; Zhou, J. Identification of preisach model for a fast tool servo system using neural networks. In Proceedings of the 2008 IEEE International Symposium on Knowledge Acquisition and Modeling Workshop, Wuhan, China, 21–22 December 2008; pp. 232–234. [Google Scholar]
  27. Sun, L.; Huang, W.M.; Ding, Z.; Zhao, Y.; Wang, C.C.; Purnawali, H.; Tang, C. Stimulus-responsive shape memory materials: A review. Mater. Des. 2012, 33, 577–640. [Google Scholar] [CrossRef]
  28. Hernández-Alvarado, R.; García-Valdovinos, L.G.; Salgado-Jiménez, T.; Gómez-Espinosa, A.; Fonseca-Navarro, F. Neural network-based self-tuning PID control for underwater vehicles. Sensors 2016, 16, 1429. [Google Scholar] [CrossRef] [PubMed]
  29. Cui, X.; Shin, K.G. Direct control and coordination using neural networks. IEEE Trans. Syst. Man Cybern. 1993, 23, 686–697. [Google Scholar]
  30. An, L.; Huang, W.M.; Fu, Y.Q.; Guo, N. A note on size effect in actuating NiTi shape memory alloys by electrical current. Mater. Des. 2008, 29, 1432–1437. [Google Scholar] [CrossRef]
Figure 1. Typical input-output relation for a system with hysteresis.
Figure 1. Typical input-output relation for a system with hysteresis.
Sensors 19 02576 g001
Figure 2. Definition of the fundamental hysteresis operator γ ^ α β . When the input is in the range ( α , β ) , the output is positive if the input reached this range from above and negative if it was reached from below.
Figure 2. Definition of the fundamental hysteresis operator γ ^ α β . When the input is in the range ( α , β ) , the output is positive if the input reached this range from above and negative if it was reached from below.
Sensors 19 02576 g002
Figure 3. Stress–temperature plane showing schematic crystal structures of the different phases of a typical SMA.
Figure 3. Stress–temperature plane showing schematic crystal structures of the different phases of a typical SMA.
Sensors 19 02576 g003
Figure 4. Plot showing deformation versus current for a Flexinol wire. I + denotes that the current is monotonically increasing, while I denotes a monotonical decrease.
Figure 4. Plot showing deformation versus current for a Flexinol wire. I + denotes that the current is monotonically increasing, while I denotes a monotonical decrease.
Sensors 19 02576 g004
Figure 5. Schematic of a neural network with two hidden layers and a four element sliding window, where s c denotes the sigmoid activation function and w j i , w j i , v j are weight coefficients.
Figure 5. Schematic of a neural network with two hidden layers and a four element sliding window, where s c denotes the sigmoid activation function and w j i , w j i , v j are weight coefficients.
Sensors 19 02576 g005
Figure 6. A block scheme of the control system.
Figure 6. A block scheme of the control system.
Sensors 19 02576 g006
Figure 7. Photograph showing the experimental setup. The Flexinol wire has been marked in cyan.
Figure 7. Photograph showing the experimental setup. The Flexinol wire has been marked in cyan.
Sensors 19 02576 g007
Figure 8. The circuit layout.
Figure 8. The circuit layout.
Sensors 19 02576 g008
Figure 9. Flowchart illustrating the back-propagation algorithm.
Figure 9. Flowchart illustrating the back-propagation algorithm.
Sensors 19 02576 g009
Figure 10. Characterization of the system. I + denotes that the current is monotonically increasing, while I denotes a monotonical decrease
Figure 10. Characterization of the system. I + denotes that the current is monotonically increasing, while I denotes a monotonical decrease
Sensors 19 02576 g010
Figure 11. Learning session using set-points.
Figure 11. Learning session using set-points.
Sensors 19 02576 g011
Figure 12. Results after learning using set-points.
Figure 12. Results after learning using set-points.
Sensors 19 02576 g012
Figure 13. Learning session using a sine wave as reference signal.
Figure 13. Learning session using a sine wave as reference signal.
Sensors 19 02576 g013
Figure 14. Results after learning.
Figure 14. Results after learning.
Sensors 19 02576 g014
Figure 15. Results for a triangle wave form.
Figure 15. Results for a triangle wave form.
Sensors 19 02576 g015
Figure 16. (a) Results for a 0.01 Hz sine wave; (b) Results for a 0.005 Hz sine wave.
Figure 16. (a) Results for a 0.01 Hz sine wave; (b) Results for a 0.005 Hz sine wave.
Sensors 19 02576 g016
Figure 17. Result for a 0.05 Hz sine wave.
Figure 17. Result for a 0.05 Hz sine wave.
Sensors 19 02576 g017
Figure 18. Plots showing the results when applying an extra: (a) 85% torque; (b) 143% torque.
Figure 18. Plots showing the results when applying an extra: (a) 85% torque; (b) 143% torque.
Sensors 19 02576 g018
Table 1. Flexinol wire characteristics.
Table 1. Flexinol wire characteristics.
Model No.DiameterLengthOperational CurrentTransition TemperaturePull ForceResistance
STD-005-900.13 mm305 mm200 mA90°0.22 kg0.75 Ω/cm

Share and Cite

MDPI and ACS Style

Gómez-Espinosa, A.; Castro Sundin, R.; Loidi Eguren, I.; Cuan-Urquizo, E.; Treviño-Quintanilla, C.D. Neural Network Direct Control with Online Learning for Shape Memory Alloy Manipulators. Sensors 2019, 19, 2576. https://doi.org/10.3390/s19112576

AMA Style

Gómez-Espinosa A, Castro Sundin R, Loidi Eguren I, Cuan-Urquizo E, Treviño-Quintanilla CD. Neural Network Direct Control with Online Learning for Shape Memory Alloy Manipulators. Sensors. 2019; 19(11):2576. https://doi.org/10.3390/s19112576

Chicago/Turabian Style

Gómez-Espinosa, Alfonso, Roberto Castro Sundin, Ion Loidi Eguren, Enrique Cuan-Urquizo, and Cecilia D. Treviño-Quintanilla. 2019. "Neural Network Direct Control with Online Learning for Shape Memory Alloy Manipulators" Sensors 19, no. 11: 2576. https://doi.org/10.3390/s19112576

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop