A Nonlinear Disturbance Observer-Based Super-Twisting Sliding Mode Controller for a Knee-Assisted Exoskeleton Robot

Abstract

Exoskeleton knee-assistance (EKA) systems are wearable robotic technologies designed to rehabilitate and improve impaired mobility of the lower limbs. Clinical exercises are conducted on disabled patients based on physically demanding tasks which are prescribed by expert physicians. In order to carry out good tracking of the desired tasks, efficient controllers must be designed. In this study, a novel control framework is introduced to improve the robustness characteristics and tracking precision of EKA systems. The control approach integrates a super-twisting sliding mode controller (STSMC) with a nonlinear disturbance observer (NDO) to ensure robust and precise tracking of the knee joint trajectory. An evaluation of the proposed system is conducted through numerical simulations under the influence of external disturbances. The findings reveal considerable enhancements to trajectory tracking accuracy and disturbance rejection when compared to conventional STSMCs and sliding mode perturbation observer (SMPO)-based STSMCs.

1. Introduction

World Health Organization (WHO) reports that strokes affect around 15 million people worldwide each year, causing permanent disability for 5 million individuals [1]. Neurological injuries such as strokes, heart attacks, and spinal cord injuries can cause loss of mobility, sensory and cognitive impairments, and reduced independence. Rehabilitation and repetitive task-specific therapy have shown promising results in improving recovery. Traditional rehabilitation exercises require direct physician supervision, making them costly and time-consuming. Actuated bio-engineering exoskeletons offer a promising alternative by reducing therapist workload, enabling repetitive motion without fatigue, and providing precise measurements (e.g., angle, velocity, torque) for patient progress evaluation. Clinical trials have confirmed their effectiveness in rehabilitation [2,3]. For successful use, exoskeletons must be customized to match human joint motion and ensure ergonomic comfort, particularly in lower-limb orthoses. Typically, the knee exoskeleton comprises two links connected to the thigh and shank, actuated by a DC motor to provide the required torque. Controlling exoskeletons poses significant challenges owing to the presence of inherent uncertainties, such as unpredictable forces from complex human–exoskeleton interactions in addition to environmental factors such as terrain and collisions that cause nonlinearities and joint misalignment [4,5]. Research in this field is active because lower-limb exoskeletons must manage unpredictable disturbances and adapt to individual biomechanics. Modern studies focus on combining robust observers with advanced control techniques to improve tracking, stability, and disturbance rejection for exoskeleton knee systems, significantly enhancing rehabilitation outcomes and daily usability. In [6], M. Zhang et al. proposed a second-order super-twisting controller (STSMC) for motion control. This is a rehabilitation exoskeleton device which is actuated by an artificial pneumatic muscle (PAM). Compared to classical SMCs, the STSMC showed better robust characteristics against uncertainties and disturbances. In [7], S. Aole et al. introduced an improved active disturbance rejection (I-ADRC) controller, differentiator (TD), and nonlinear state error feedback (NLSEF) for use in rehabilitation robots. The simulation of the proposed controller shows reduced trajectory tracking error, the elimination of random, constant, and harmonic disturbances, and robustness against parameter variations and under the influence of noise. In [8], Ezhilarasi, D. et al. proposed an adaptive super-twisting sliding mode controller (ASTSMC) to evaluate the performance of a suggested 6-DOF lower limb exoskeleton model for tracking trajectory under external disturbances and parametric uncertainties. The controller performance was evaluated under wind disturbances with varying velocities and directions. The suggested controller achieved superior tracking accuracy when compared to a constant gain STSMC on both flat and uneven terrain. In [9], M. Mokhtari et al. presented a hybrid optimal adaptive high-order super-twisting sliding mode (AHOSTWSM) impedance controller to control the motion of the lower limbs. The optimization technique was based on the Harmony Search Algorithm (HSA) and was used to tune the controller parameters. The researchers developed an optimal sliding mode controller (SMC); the performance of the AHOSTWSM was compared with that of the SMC. The results of the simulation demonstrated the controller’s enhanced performance compared to that of the SMC. In [10], M. Rastegar and H. Kobravi proposed a hybrid method that merges Functional Electrical Stimulation (FES) with mechanical actuators for knee joint motion control in rehabilitation, employing adaptive and PD controllers. The results of the simulation showed that this algorithm enhanced tracking performance and disturbance rejection during rehabilitation tasks. In [11], Yang Yong et al. developed a control scheme for hydraulic rehabilitation exoskeleton knee joints that were capable of achieving accurate trajectory tracking in the presence of nonlinearities, modeling uncertainties and external disturbances. The suggested algorithm integrates a disturbance observer-based radial basis function neural network (RBF-NN) to estimate and compensate for modeling uncertainties. Comparative simulations established that the proposed control algorithm efficiently compensates for modeling uncertainties and ensures high-precision trajectory tracking for the exoskeleton knee joint. In [12], Jin. X. and Guo. J suggested a model predictive (MPC) strategy to achieve a specified tracking performance for lower limb rehabilitation exoskeletons. The researchers integrated an extended state observer (ESO) with the MPC framework to estimate total system disturbances. The suggested ESO-MPC controller considerably improved tracking accuracy for lower limb motion under disturbances exceeding classical MPC and fuzzy PID methods by more than 34%. In [13], AL-Dujaili et al. proposed an Active Disturbance Rejection Control (ADRC) scheme for the exoskeleton robot. Two observers, represented by a linear extended state observer (LESO) and a nonlinear extended state observer (NESO), were included to estimate the states and uncertainties. Particle Swarm Optimization (PSO) was utilized to optimally tune the observer’s parameters. In [14], He D. et al. proposed a model-free super-twisting terminal sliding mode controller (MF-STTSMC) integrated with a sliding mode disturbance observer for tracking control and rejection disturbance for n-DOF upper-limb rehabilitation exoskeletons subject to backlash and hysteresis nonlinearities. A sliding mode disturbance observer was incorporated to identify and compensate for unmolded disturbances and hysteresis effects. The results of the simulation confirmed that the proposed controller efficiently mitigated errors induced by backlash and hysteresis; moreover, it was found to achieve high-precision trajectory tracking and demonstrated strong robustness against external disturbances and model uncertainties. In [15], Alawad Nasir Ahmed et al. developed an active disturbance rejection controller (ADRC) for the coordinated motion of hip and knee joints in a rehabilitation exoskeleton device. The ADRC applies an extended state observer (ESO) to approximate and compensate for disturbances and unknown modeling errors online. The proposed scheme demonstrated superior performance in trajectory tracking disturbance rejection and robustness compared to the classical PID controller. In [16], Qin Lang et al. developed a super-twisting sliding mode control (ST-SMC) algorithm for exoskeleton joint motors that leveraged a load observer to improve robustness and control effectiveness under uncertain load conditions and disturbances. This methodology integrates the advantages of back-stepping control and sliding mode strategies, enabling the system to perform real-time estimation and compensation for disturbances. The outcomes of both the simulations and the experiments indicate that the proposed controller improved stability for exoskeleton motors compared to classical control methods. In [17], Kl Espinosa-Espejel et al. proposed a control strategy employing Artificial Neural Networks (ANNs) for walking trajectory generation in a lower limb exoskeleton for use during rehabilitation. The effectiveness of the ANN-based controller was verified through both numerical simulations and experimental validations. In [18], H Tiaiba et al. developed an online adaptive super-twisting sliding mode controller, which was implemented for an upper limb exoskeleton robot targeting arm rehabilitation. The controller incorporated an optimized Particle Swarm Optimization (PSO) algorithm for real-time parameter adjustment. The proposed controller outperformed the classic super-twisting algorithm in both performance and efficiency. In [19], Chen Nengdi et al. developed an enhanced sliding mode control strategy, employing a fractional-order nonlinear disturbance observer (FONDOB) in conjunction with an extreme learning machine (ELM), aiming to estimate the system uncertainty and to ensure the walking stability of the lower limb exoskeleton. The results of the experiment showed that the proposed fractional-order controller demonstrated superior control performance compared to the classical integer-order sliding mode control. In [20], Y. Wang et al. developed a proportional derivative iterative learning control strategy for a multi-joint soft lower limb assistive exoskeleton to support flexion and extension of the hip and knee joint throughout the gait cycle. A mean-based prediction approach was utilized to forecast the human walking cycle. Unlike the above studies, the observer-based scheme was presented in order to estimate disturbance and to compensate within a controlled loop.

In this study, an NDO-based STSMC is designed for an EKA system to handle nonlinear dynamics and disturbances during knee movement. The SMC is a nonlinear and robust control scheme, which is designed for solving various nonlinear control problems. The mechanism of the SMC acts to drive the states of nonlinear systems to a sliding surface and to enforce them staying there, irrespective of the presence of disturbances and uncertainties. One reported critical problem with SMC is the chattering effect due to sharp switching of control law components. One solution to this problem is the use of super-twisting sliding mode control (STSMC), which is the modified version of SMC [21,22,23,24,25,26,27,28,29].

The STSMC is a powerful and effective algorithm used for stabilizing nonlinear systems. Complementing this, the NDO offers simple architecture, rapid response, and high accuracy in disturbance estimation; therefore, the control system can prevent excessive joint torques that might cause discomfort or injury.

The main contributions of this study are summarized as follows:

• A novel control framework is designed which integrates both STSMC and NDO to approximate the external perturbations and improve the robustness of a controlled medical robot.
• Stability analysis of an NDO-based controller is conducted to ensure tracking convergence in the presence of uncertainty.
• Comparative analysis of the NDO-based STSMC and the SMPO-based STSMC algorithms is carried out under unified conditions, including external disturbance, to reveal the differences in robustness and control effort.

The structure of this paper is organized as follows: Section 2 outlines the methodology, which includes detailed mathematical modeling of an exoskeleton knee-assistive system, and the proposed control approach. The simulated results are outlined in Section 3. Finally, Section 4 illustrates the conclusions of this study.

2. Materials and Methods

The following sections explain the two main steps to achieving this study’s objectives. Section 2.1 details the mathematical modeling of an exoskeleton knee-assistive system. This section comprehensively explains the state space form of the EAK system. Section 2.2 presents a nonlinear control strategy based on a super-twisting sliding mode controller with a disturbance observer.

2.1. Mathematical Modeling of an Exoskeleton Knee-Assistive System

Figure 1 illustrates the geometric representation of an EKA system, which includes a fixed thigh link and a movable shin link actuated by a DC motor mounted at the knee joint. The motor generates the required torque for precise leg rotation, assisting impaired knees. The motion range is limited to between 0° (full extension) and 90° (resting posture). The system uses a fixed frame F and a rotating knee frame S, defined by angle $\theta$ and velocity $\dot{\theta }$. A coupled exoskeleton–leg model was derived via the Lagrangian method.

Initially, we will consider the exoskeleton and human leg as a coupled system and describe them with the following mathematical relationship [30]:

$$\mathcal{l}=E_k-E_g$$

(1)

where, $\mathcal{l}$ represents the Lagrangian operator. $E_g$ and $E_k$ denote the gravitational and kinetic energies, respectively.

$$E_k={\displaystyle \frac{1}{2}}J{\dot{\theta }}^2$$

(2)

$J$ represents the inertia of the combined exoskeleton–human leg system.

$$E_g={\tau }_g\left(1-sin\theta \right)$$

(3)

where ${\tau }_g$ represents the gravitational torque which can be represented as follows:

$${\tau }_g=mgl$$

(4)

where, $m$ denotes the mass of the coupled system, g the gravitational acceleration, and $l$ the distance between the knee joint and the center of gravity. The coupled system dynamics are then formulated based on the Euler–Lagrange equation as follows:

$$J\ddot{\theta }=-{\tau }_{g}cos\theta +{\tau }_{ext}$$

(5)

Here, ${\tau }_{ext}$ represents the total external torque on the system, defined as ${\tau }_{ext}={\tau }_f+\tau$, where $\tau$ is the motor-controlled input torque and ${\tau }_f$ is the friction torque expressed by

$${\tau }_f=-f_{s}sgin\dot{\theta }-f_v\dot{\theta }$$

(6)

$f_v$: viscous friction, $f_s$: solid friction. The coupled system dynamics are given as follows:

$$J\ddot{\theta }=-{\tau }_{g}cos\theta -f_{s}sgin\dot{\theta }-f_v\dot{\theta }+{\tau }_h+\tau$$

(7)

${\tau }_h$ denotes the load torque from system coupling. By renaming $\theta$ as $x_1$ and $\dot{\theta }$ as $x_2$, the state space form of the EAK system can be written as follows:

$${\dot{x}}_1=x_2$$

$${\dot{x}}_2={\displaystyle \frac{1}{J}}\left[-{\tau }_g\mathit{cos}\left(x_1\right)-f_{s}sign\left(x_2\right)-f_v{x}_2+{\tau }_h+\tau \right]$$

(8)

The nonlinear plant with internal/external disturbances is represented as follows:

$$\{\begin{array}{c}{\dot{x}}_1=x_2\\ {\dot{x}}_2=f\left(x,t\right)+b\tau +d\\ y=x_1\end{array}$$

(9)

where, $y$ is the output signal, $f\left(x,t\right)$ represents the nonlinear dynamics, $d={\tau }_h/J$ is the external disturbance, and $b$ is the control gain. The physical parameters of the EKA system are listed in

Table 1. Identification parameters.

Parameters	Value
Gravity Torque $\left({\tau }_g\right)$	$5 \mathrm{N}.\mathrm{m}$
Solid Friction Coefficient $\left(f_s\right)$	$0.6 \mathrm{N}.\mathrm{m}$
Viscous Friction Coefficient $\left(f_v\right)$	$1 \mathrm{N}.\mathrm{m}.\mathrm{s}/\mathrm{r}\mathrm{a}\mathrm{d}$
Inertia $\left(J\right)$	$0.4 \mathrm{k}\mathrm{g}.{\mathrm{m}}^2$

[30].

2.2. Control Methodology

To control the nonlinear dynamics of the exoskeleton knee system, the controller must ensure accurate tracking of the reference angle $\theta$ with minimal steady-state error, while avoiding overshoot and exceeding the maximum joint limit. To achieve this, a nonlinear control strategy based on a super-twisting sliding mode controller with a disturbance observer is proposed [31,32,33,34,35,36,37,38,39,40].

2.2.1. Nonlinear Disturbance Observer

For the purpose of designing the observer unit, the equation of state for the EAK system will first be rewritten as follows [41]:

$$\dot{x}=F+G\tau +D$$

(10)

where $\dot{x}={\left[{\dot{x}}_1,{\dot{x}}_2\right]}^T$, $F={\left[x_2,f\left(x,t\right)\right]}^T$, $G={\left[0,b\right]}^T$ and, $D={\left[0,d\right]}^T$. The proposed corresponding nonlinear disturbance observer will be defined as follows:

$$\widehat{D}=z+P$$

(11)

$$\dot{z}=-Lz-L(P+F+G\tau )$$

(12)

where $\widehat{D}$ represent the output generated by the proposed observer, $z$ represents the internal variable of the proposed observer, $L=diag[n;n]$ is defined as the gain matrix, and $n$ is a real number that must be greater than zero, while $P$ represents a design function, which can be defined by

$$\dot{P}=L\dot{x}$$

(13)

The estimation error of the disturbance observer and its corresponding time derivative are defined as follows:

$$\tilde{D}=D-\widehat{D}$$

(14)

The disturbance $D$ is considered time-varying and can be represented as follows:

$$\left|\dot{D}\right|\le \alpha$$

(15)

Then,

$$\dot{\tilde{D}}=\dot{D}-\dot{\widehat{D}}=\dot{D}-L\tilde{D}$$

(16)

Rewriting Equation (16), we obtain the following:

$$\dot{\tilde{D}}+L\tilde{D}=\dot{D}$$

(17)

Multiplying both sides by the integrating factor $\left(e^{Lt}\right)$ produces the following:

$$\dot{\tilde{D}}{e}^{Lt}+L\tilde{D}{e}^{Lt}=\dot{D}{e}^{Lt}$$

$${\displaystyle \frac{d}{dt}}\left(e^{Lt} \tilde{D}\left(t\right)\right)=\dot{D}{e}^{Lt}$$

(18)

Integrating Equation (18) during the period [0,t] yields the following:

$$\tilde{D}\left(t\right)=\tilde{D}\left(0\right)e^{-Lt}+e^{-Lt}{\int }_0^t\dot{D}{e}^{L\tau } d\tau$$

(19)

Since $\left|\dot{D}\right|\le \alpha$, we then have

$$\tilde{D}\left(t\right)\le \tilde{D}\left(0\right)e^{-Lt}+e^{-Lt}\alpha {\int }_0^t{e}^{L\tau } d\tau$$

(20)

Thus,

$$\tilde{D}\left(t\right)\le \left(\tilde{D}\left(0\right)-{\displaystyle \frac{\alpha }{L}}\right)e^{-Lt}+{\displaystyle \frac{\alpha }{L}}$$

(21)

As t → ∞, we then have

$$\tilde{D}\left(\infty \right)\le {\displaystyle \frac{\alpha }{L}}$$

(22)

Therefore, the solution of Equation (17) is uniformly ultimate bounded, and the observer error state ultimately enters a small residual set around origin with a radius depends on the ratio $(\alpha /L)$. Thus, the disturbance estimate $\widehat{D}$ converges to zero provided that Equation (22) is satisfied to ensure asymptotic stability.

2.2.2. Super-Twisting Sliding Mode Controller Scheme Design

Let $e$ define the error between the angle $x_1$ and the desired angle ${\theta }_d$

$$e=x_1-{\theta }_d$$

(23)

The first derivative of the error equation is given by

$$\dot{e}=x_2-{\dot{\theta }}_d$$

(24)

In the STSMC, the desired dynamic behavior of the system can be expressed as a mathematical condition called the sliding surface, which is typically formulated as a function of the system error and its time derivatives, defined as follows:

$$s=\dot{e}+\lambda e$$

(25)

Now, by taking the time derivative of Equation (25), the following can be obtained:

$$\dot{s}=\ddot{e}+\lambda \dot{e}={\dot{x}}_2-{\ddot{\theta }}_d+\lambda \dot{e}=f\left(x,t\right)+b\tau +d-{\ddot{\theta }}_d+\lambda \dot{e}$$

(26)

In the design theory of STSMC, the control law is formulated with two parts: equivalent $\left({\tau }_{eq}\right)$ and switching $\left({\tau }_{sw}\right)$, respectively.

$$\tau ={\displaystyle \frac{1}{b}}\left({\tau }_{eq}+{\tau }_{sw}\right)$$

(27)

The ${\tau }_{eq}$ part can be derived by setting $\dot{s}=0$. Therefore, based on Equation (26), ${\tau }_{eq}$ can be determined as follows:

$${\tau }_{eq}=-f\left(x,t\right)-\widehat{d}+{\ddot{\theta }}_d-\lambda \dot{e}$$

(28)

Based on the super-twisting algorithm [42], the ${\tau }_{sw}$ part is proposed to be

$${\tau }_{sw}=-c_1\sqrt{\left|s\right|}sgn\left(s\right)-c_2\int sgn\left(s\right)dt$$

(29)

Then, the control law can now be written as

$$\tau ={\displaystyle \frac{1}{b}}\left(-f\left(x,t\right)-\widehat{d}+{\ddot{\theta }}_d-\lambda \dot{e}-c_1\sqrt{\left|s\right|}sgn\left(s\right)-c_2\int sgn\left(s\right)dt\right)$$

(30)

Accordingly, Equation (26) becomes

$$\dot{s}=\left(d-\widehat{d}\right)-c_1\sqrt{\left|s\right|}sgn\left(s\right)-c_2\int sgn\left(s\right)dt$$

$$\dot{s}=\check{d}-c_1\sqrt{\left|s\right|}sgn\left(s\right)-c_2\int sgn\left(s\right)dt$$

(31)

To guarantee asymptotic stability of the EAK system under the STSMC, the Lyapunov candidate function is selected as follows:

$$V={\displaystyle \frac{1}{2}}{s}^2$$

(32)

Taking the time derivative of Equation (32) and using Equation (30), the following can be obtained:

$$\dot{V}=s\left(\check{d}-c_1\sqrt{\left|s\right|}sgn\left(s\right)-c_2\int sgn\left(s\right)dt\right)\le 0$$

(33)

Thus, $t>0$, $(\check{d}\le {\displaystyle \frac{\alpha }{L}})$, confirming that the NDO-based control system is asymptotically stable. Figure 2 illustrates the schematic diagram of the NDO-based STSMC applied to the EAK system.

3. Results and Discussion

This section validates the effectiveness of the proposed controller through numerical simulations, employing the “Ode45” solver as the numerical integration method. In order to verify the efficiency of the proposed control framework in improving the performance of the system in terms of durability of performance in the presence of external disturbances and reducing the control effort, we conducted a performance comparison with the SMPO-based STSMC, as detailed in [43], and the classical STSMC. The gains of the proposed control frameworks were tuned via the Particle Swarm Optimization (PSO) algorithm [44]; the PSO-based gains are listed in

Table 2. The PSO-based gains of the proposed control frameworks.

STSMC		STSMC-Based SMPO		STSMC-Based NDO
Parameter	Value	Parameter	Value	Parameter	Value
$\lambda$	8.3015	$\lambda$	8.3015	$\lambda$	8.3015
$c_1$	4.9998	$c_1$	4.9998	$c_1$	4.9998
$c_2$	3.3176	$c_2$	3.3176	$c_2$	3.3176
------	------	$k$	10.8517	$n$	150
------	------	$\beta$	37.6018	------	------
------	------	$\delta$	0.0072	------	------

.

The tracked response of the EAK system under ideal conditions, in the absence of external disturbances and uncertainty based on classical STSMC, is depicted in Figure 3. Specifically, Figure 3a illustrates the knee angle trajectory, Figure 3b depicts the angular velocity response, Figure 3c shows the error signal, and Figure 3d provides the corresponding control action generated by the proposed controller. The results are utilized to highlight the capability of the STSMC in accurately tracking the desired knee angle.

To assess the robustness of proposed observe-based controller, an external disturbance is introduced by exerting torque on the human leg during training,. Figure 4 presents the dynamic responses of the controlled system under turbulent conditions. Specifically, Figure 4a illustrates the knee angle trajectory, Figure 4b depicts the angular velocity response, Figure 4c shows the tracking error, and Figure 4d presents the corresponding control action generated by the proposed control frameworks. In addition, Figure 5 compares the performance of the NDO with that of the SMPO.

To quantitatively evaluate the performance of the proposed control framework, a number of evaluation functions were used. The root mean square error (RMSE) and root mean square disturbance error (RMSDE) were used to measure the average tracking deviation over time, providing a reliable indicator of tracking accuracy. Additionally, time domain performance indices, integral of the time-weighted absolute error (ITAE), were used to evaluate the proposed control framework’s ability to minimize the steady-state error.

Table 3. Evaluation of the controlled exoskeleton knee-assist system.

Controller	ITAE	RMSE	RMSDE
STSMC	1.2874	0.0467	-----
STSMC based on NDO	0.0468	0.0455	0.0686
STSMC based on SMO	0.0529	0.0462	0.0878

summarizes the overall efficiency of the EAK system under the proposed control frameworks.

$$RMSE=\sqrt{{\displaystyle \frac{ \sum _1^m{e}^2}{m}}}$$

(34)

$$RMSDE=\sqrt{{\displaystyle \frac{\sum _1^m{(d-\widehat{d})}^2}{m}}}$$

(35)

$$ITAE={\int }_0^{\infty }t\left|e\right| dt$$

(36)

The results of the simulation show that the NDO-based STSMC successfully achieved high-precision tracking performance even under external disturbances, which ensures that the movements required during the rehabilitation sessions are performed completely smoothly. The controller achieved lower RMSE, RMSDE, and ITAE values compared to classic STSMCs and SMPO-based STSMC methods, indicating strong robustness and effective disturbance rejection by the NDO.

4. Conclusions

In this work, the NDO-based STSMC is designed for trajectory tracking by estimating the external disturbance and to reject it within controlled EAK system. The proposed approach integrates the strengths of high-order STSMC with a disturbance observer to enhance the system’s robustness against uncertainties and external disturbance. The performance evaluation of the NDO-based STSMC algorithm showed better performance and robustness characteristics compared to the SMPO-based STSMC and classical STSMC algorithms. The superiority of its performance arises from the NDO’s capability to accurately estimate the external disturbances to be later compensated in the closed loop. In a numeric sense, the NDO-based STSMC exhibited a 22% lower root mean square disturbance error (RMSDE) compared to the SMPO-based STSMC.

In spite of its improved performance due to the proposed control framework, there are some challenging limitations to be solved in the future extension of this study. The noise and saturation of the actuator are the main limitations which have to be taken into account in the design of the proposed controller.

For future work, other control strategies can be suggested and compared to the proposed controller for the same application, such as Adaptive Neural Network Control, adaptive sliding mode control, and Predefined-Time Reliable Control [45,46,47]. Further extension of this study can be made by implementing the proposed control frame-work in a real-time environment. The embedded system design based on the FPGA and graphical system design (based on LabVIEW) can be developed to implement a real-world controlled medical system.