International Journal of Power Electronics and Drive Systems (IJPEDS)

Vol. 16, No. 1, March 2025, pp. 418~430

ISSN: 2088-8694, DOI: 10.11591/ijpeds.v16.i1.pp418-430  418

Enhanced reaching law for improved response in sliding mode

control of PMSM motors with fuzzy logic integration

Khanh Quoc Truong, Son Huynh, Dung Hoang Vo, Minh Duc Pham
Power Electronics Research Laboratory, Faculty of Electrical and Electronics Engineering,
Ho Chi Minh City University of Technology (HCMUT), Ho Chi Minh City, Vietnam
Vietnam National University, Linh Trung Ward, Thu Duc City, Ho Chi Minh City, Vietnam

Article Info ABSTRACT

Article history: The rising demand for high-performance permanent magnet synchronous
motors (PMSMs) is responsible for the development of PMSM speed control.
Received May 5, 2024 Although the proportional-integral controller is often used in field-oriented
Revised Sep 11, 2024 control (FOC) for motor speed regulation, it has drawbacks like slow response
Accepted Sep 19, 2024 and instability. This paper proposed an enhanced sliding mode controller with
a modified sliding surface to achieve better speed control performance. In
comparison to proportional-integral or PI controller, fuzzy logic controller,
Keywords: conventional sliding mode controller, the proposed control approach uses a
reaching law that incorporates a fuzzy logic controller. A smoother and faster
Enhanced reaching law response time is targeted by the proposed approach compared to conventional
FOC sliding mode control. Practical small-scale PMSM experiments certify the
Fuzzy logic algorithm effectiveness of our proposed enhanced sliding mode control.
Motor speed regulation
PMSM This is an open access article under the CC BY-SA license.

Corresponding Author:
Minh Duc Pham
Power Electronics Research Laboratory, Faculty of Electrical and Electronics Engineering
Ho Chi Minh City University of Technology (HCMUT)
268 Ly Thuong Kiet Street, District 10, Ho Chi Minh City, Vietnam
Email: [email protected]

1. INTRODUCTION
Permanent magnet synchronous motors (PMSMs) are now being widely used in many applications
due to their high-power density, efficiency, and torque-to-current ratio. Their speed performance is significant,
and they are critical in applications such as robotics, electric vehicles, and machine tools [1], [2]. In an
excavator, for instance, suitable speed regulation enables the machine to move precisely and effectively
perform excavation operations. It ensures that lifting and lowering heavy loads are done smoothly, increasing
operational safety when used by cranes and elevators. Controlling the speed of a PMSM is important in electric
vehicles (EVs) for acceleration, maintaining constant velocity, and effectively operating regenerative braking. In
contrast, torque control is common in heavy industry applications such as excavators and product conveyor belts.
The importance of speed control versus torque control varies significantly depending on application. In this
research, we are particularly focused on designing a speed controller for PMSM motor in EV acceleration systems.
PMSMs are characterized by nonlinear and multi-variable dynamics influenced by flux linkage and
direct-quadrature or DQ-axes inductances, which can vary during motor operation due to magnetic saturation.
It is also difficult to produce both high accuracy and speed, as PMSMs exhibit strong nonlinearity due to the
strong coupling between torque and speed in a mechanical model. To address the complex issues arising from
the mechanical equation of PMSMs, the field-oriented control (FOC) technique emerged [3]-[5]. In the FOC
technique, the motor control system is transformed into a synchronous reference frame, simplifying the control
task by decoupling the torque and flux components.

Journal homepage: http://ijpeds.iaescore.com

Int J Pow Elec & Dri Syst ISSN: 2088-8694  419

The outer control loop is the speed control loop, responsible for generating the q-axis reference
current. The inner control loop regulates the motor current using a proportional-integral (PI) controller within
the DQ frame. This PI control loop effectively tracks the d and q components of the motor current references,
ensuring precise control over torque and flux. To control the PMSM motor speed, the linear PI controller is
employed to adjust the q current reference of the inner control loop for tracking motor speed [6]-[8]. Although
PI speed control offers adequate control precision and robustness, it often falls short in terms of slow response
and stability when confronted with disturbances such as load changes and variations in speed [9], [10].
Conventional studies on PMSM speed control using the PI controller specifically highlight significant
challenges related to speed response and stability [11]-[13]. That is because speed regulation tends to be a
linear calculation, but PMSM dynamics are non-linear. For this reason, more improved control methods are
needed to improve the slow response problem and system stability.
Nonlinear control techniques have emerged as potential alternatives that can address the limitations
of linear controllers [14]-[16]. Nonlinear control techniques such as adaptive control, predictive control, and
sliding mode control have demonstrated the potential to improve the stability, precision, and robustness of
PMSM motor control. Adaptive control can handle uncertainties by adjusting control parameters in real-time,
but it needs accurate system parameters and may cause system instability [17]-[19]. However, on the other
hand, predictive control involves predicting future system behavior which is able to achieve satisfactory
performance but at high computational costs and sensitivity due to system mismatches [20]-[22]. Sliding mode
control provides robustness in the presence of uncertainties and disturbances using discontinuous controls;
however, it may introduce chattering problems [23]-[25]. The use of sliding mode control is important for
stability reasons in the presence of disturbances and uncertainties.
Given the drawbacks of previous works, it requires an improved control law and reaching conditions
for chattering attenuation along with speed response improvements. An enhanced sliding mode control is
presented in this paper to improve control response and stability with reduced chattering. A sliding surface
with the improved reaching law is added in this proposed control approach, which also includes fuzzy logic
control integration. These modifications will allow the system state to arrive at the sliding surface as fast as
possible and reduce chattering of control actions thanks to the adjustable sliding gain. Furthermore, the system
is always stable with this defined reaching law. This control approach not only provides smoother dynamics
but also obtain a shorter response time than the conventional sliding mode control. The control theory is verified
by means of experimental results with a small-scale PMSM system.

2. FOC AND MATHEMATICAL MODEL OF PMSM

2.1. Principle of FOC algorithm
In permanent magnet synchronous motors, the stator current affects both the magnetic flux and the
generated torque. Hence, field-oriented control was proposed to decouple flux and torque can be achieved by
analyzing the instantaneous current into two components: One component aligned with the rotor magnetic
field, called the direct axis (d), and the other perpendicular to the rotor magnetic field, called the quadrature
axis (q) [26]. Field-oriented control is the basic idea of synthesizing sinusoidal inverter voltage to provide
progressive speed control with optimal torque output. Maximum torque is reached when the angle between the
magnetic fields of the stator and rotor becomes 90 degrees. The stator magnetic flux is aligned orthogonally to
the rotor by forcing the d-axis current to 0 and properly adjusting the q-axis.

2.2. Coordination between speed controller and FOC

The block diagram of a conventional speed controller with FOC with speed controller is shown in
Figure 1. In the figure, the control process begins with measuring the motor three-phase currents, which are
then transformed through Clark and Park transformations to obtain currents along the independent d and q
∗
axes. Then, the current rotor speed is compared with the reference speed (𝜔𝑚 ) to generate a speed error signal
∗
(𝑥1 = (𝜔𝑚 − 𝜔𝑚 )), which is fed into an outer speed controller to produce a reference current along the q-axis
(𝑖𝑞∗ ). In the inner control loop, two PI controllers are utilized to regulate the currents along the d and q axes,
generating reference voltage signals along these axes independently (𝑢𝑞 , 𝑢𝑑 ). Following this, inverse Clark and
Park transformations are applied to convert the voltage signals from the dq axes to three-phase voltage signals
for the space vector pulse width modulation (SVPWM) block. The SVPWM is a modulation technique used to
generate PWM signals for controlling the switching devices in an inverter [27]-[29].

2.3. Mathematical model of PMSM

To implement the FOC algorithm on a digital processor, a mathematical model of the motor is
essential. Through the Clark and Park transformation, the mathematical model of the PMSM is transformed
into the following mathematical representation in the DQ frame. Where 𝑈𝑑 , 𝑈𝑞 are d-axis and q-axis voltages,

Enhanced reaching law for improved response in sliding mode control of PMSM … (Khanh Quoc Truong)
420  ISSN: 2088-8694

𝑖𝑑 , 𝑖𝑞 are d-axis and q-axis currents, 𝐿𝑑 , 𝐿𝑞 are d-axis and q-axis inductance, 𝑅𝑠 is coil resistance, 𝜔𝑒 is the
electrical speed of the rotor, 𝜔𝑚 is the mechanical angular speed of the rotor, and 𝜓𝑓 is permanent magnet flux
linkage. For the steady state, the differential terms can be disregarded since there are no variations in currents,
the (1) can be simplified as (2).

𝑑𝑖𝑑
𝑈𝑑 = 𝑅𝑠 𝑖𝑑 − 𝜔𝑒 𝐿𝑞 𝑖𝑞 + 𝐿𝑑
𝑑𝑡
{ 𝑑𝑖𝑞 (1)
𝑈𝑞 = 𝑅𝑠 𝑖𝑞 + 𝜔𝑒 (𝐿𝑑 𝑖𝑑 + 𝜓𝑓 ) + 𝐿𝑞
𝑑𝑡

𝑈𝑑 = 𝑅𝑠 𝑖𝑑 − 𝜔𝑒 𝐿𝑞 𝑖𝑞
{ (2)
𝑈𝑞 = 𝑅𝑠 𝑖𝑞 + 𝜔𝑒 (𝐿𝑑 𝑖𝑑 + 𝜓𝑓 )

The relationship between electromagnetic torque and the d and q-axis currents in the DQ reference
frame is given by the (3):

3
𝑇𝑒 = 𝑝[𝜓𝑓 + (𝐿𝑑 − 𝐿𝑞 )𝑖𝑑 ]𝑖𝑞
2
{ 𝐽 𝑑𝜔𝑚 (3)
𝑇𝑒 − 𝑇𝐿 =
𝑝 𝑑𝑡

where 𝑇𝑒 is electromagnetic torque, 𝑇𝐿 is load torque, 𝑝 is number of the pole pairs, 𝐽 is inertia moment. In the
FOC algorithm, assuming the angle 𝜃 is accurately measured from the motor, to achieve the optimal torque,
controlling the current 𝑖𝑑 = 0 is necessary. Then, the dynamic equation of the PMSM motor is simplified as (4).

3
𝑇𝑒 = 𝑝𝜓𝑓 𝑖𝑞
2
{3 𝐽 𝑑𝜔𝑚 (4)
𝑝𝜓𝑓 𝑖𝑞 − 𝑇𝐿 =
2 𝑝 𝑑𝑡

In (4), the relationship between the mechanical angular speed of the rotor 𝜔𝑚 and the controlled current on the
quadrature axis (𝑖𝑞 ) is determined. This relationship forms the fundamental equation for controlling the
mechanical angular speed to generate the reference current 𝑖𝑞∗ .

Figure 1. Block diagram of FOC with speed controller

3. STATE VARIABLES OF THE PMSM SPEED CONTROLLER AND CONVENTIONAL

SLIDING MODE CONTROL
Sliding mode control ensures stability in uncertain and noisy conditions through discontinuous control
laws. However, selecting an inappropriate sliding surface can lead to chattering problems, which are common
issues in sliding control systems. Thus, designing a sliding mode control system with reduced chattering is
crucial. This drives the research to develop an enhanced approach based on the challenges arising from
conventional approaches.

Int J Pow Elec & Dri Syst, Vol. 16, No. 1, March 2025: 418-430
Int J Pow Elec & Dri Syst ISSN: 2088-8694  421

3.1. Formulating the state variables of the PMSM speed controller

∗
The system state variables of the speed controller are defined as (5). Where 𝜔𝑚 and 𝜔𝑚 are speed
references and the instant speed of the PMSM motor. Derived from (5), 𝑥̇ 2 is derived as (6).
∗
𝑥1 = 𝑒 = 𝜔𝑚 − 𝜔𝑚
{ 𝑑𝑥1 𝑑𝜔𝑚 𝑝 3 (5)
𝑥2 = 𝑒̇ = =− = − ( 𝑝𝜓𝑓 𝑖𝑞 − 𝑇𝐿 )
𝑑𝑡 𝑑𝑡 𝐽 2

𝑑 2 𝜔𝑚 3𝑝2 𝑑𝑖𝑞
𝑥̇ 2 = 𝑒̈ = − =− 𝜓𝑓 (6)
𝑑𝑡 2 2𝐽 𝑑𝑡

3𝑝2 𝑑𝑖𝑞
For simplicity, we assign 𝜓𝑓 and in (6) as (7).
2𝐽 𝑑𝑡

3𝑝2 𝑑𝑖𝑞
𝐾= 𝜓𝑓 ; 𝑈 = (7)
2𝐽 𝑑𝑡

From (5) and (6), the state equation of the system can be expressed as (8).

𝑥̇ 0 1 𝑥1 0
[ 1] = [ ][ ] + [ ]𝑈 (8)
𝑥̇ 2 0 0 𝑥2 −𝐾

Once the system state variables are identified, the subsequent step involves selecting a representative
sliding surface, typically as in (9).

𝑠 = 𝑥2 + 𝑐𝑥1 (9)

With the expression of the sliding surface, the value of the coefficient 𝑐 is correlated with the asymptotic
stability of the sliding mode and the convergence rate towards the sliding surface. Given the defined sliding
surface, the reference current along the q-axis (𝑖𝑞∗ ) can be determined from its expression (9) as in (10).
∗
𝑑𝑖𝑞 1
𝑠̇ = 𝑥̇ 2 + 𝑐𝑥̇1 = 𝑥̇ 2 + 𝑐𝑥2 → 𝑥̇ 2 = 𝑠̇ − 𝑐𝑥2 = −𝐾 → 𝑑𝑖𝑞∗ = (−𝑠̇ + 𝑐𝑥2 )𝑑𝑡 (10)
𝑑𝑡 𝐾

Taking the integral of both sides of (10), we obtain (11).

1
𝑖𝑞∗ = ∫(−𝑠̇ + 𝑐𝑥2 )𝑑𝑡 (11)
𝐾

The expression for the reference current 𝑖𝑞∗ in (11) forms the basis for developing the control algorithm
for the speed controller. Through this expression, the dependency of the sliding surface convergence rate 𝑠̇ on
generating the reference current 𝑖𝑞∗ can be observed, and designing the equation for 𝑠̇ is crucial and directly
impacts control performance. Assuming that the PMSM angle is accurately measured from the motor encoder,
the motor speed 𝜔𝑚 is calculated, and this value is used to determine 𝑥1 and 𝑥2 . Then, these state variables are
used to determine the reference current 𝑖𝑞∗ , as shown in (11).

3.2. The conventional reaching law method

The conventional reaching law method aims to ensure fast convergence of the system to the sliding
surface while mitigating chattering and overshoot. The linear reaching law offers simplicity and ease of
implementation. The conventional reaching law is design is given by (12). In the conventional reaching law
method, the reaching time 𝑡1 is defined as the time required for the states to approach the sliding surface, and
it can be used to evaluate the controller performance. The reaching time 𝑡1 can be calculated according to (12)
as in (13).

𝑠̇ = −𝜀 𝑠𝑔𝑛( 𝑠) (12)

𝑑𝑠 𝑡 0 1 𝑆(0)
𝑠̇ = = −𝜀 𝑠𝑔𝑛( 𝑠) → ∫0 1 𝑑𝑡 = ∫𝑆(0) 𝑑𝑠 → 𝑡1 = (13)
𝑑𝑡 −𝜀 𝑠𝑔𝑛(𝑠) 𝜀

Based on the above expression, it can be seen that the reaching time 𝑡1 depends on the constant 𝜀, and
the initial state position of 𝑆(0). To improve the time to reach the sliding surface 𝑡1 , the constant 𝑞 is

Enhanced reaching law for improved response in sliding mode control of PMSM … (Khanh Quoc Truong)
422  ISSN: 2088-8694

introduced. The linear reaching law in the conventional method in [24] offers improvement in reaching time 𝑡1 ,
making it a popular choice in many control applications. The conventional reaching law is designed as in (14):

𝑠̇ = −𝜀 𝑠𝑔𝑛( 𝑠) − 𝑞𝑠 (14)

where 𝜀, 𝑞 are positive constants. Substituting (14) into (11), we have the following current reference as in (15).
1
𝑖𝑞∗ = ∫(𝜀 𝑠𝑔𝑛( 𝑠) + 𝑞𝑠 + 𝑐𝑥2 ) 𝑑𝑡 (15)
𝐾

From (15), it is necessary to consider the influences of constants 𝜀 and 𝑞 in controlling the speed of
PMSM. In particular, a larger value of constants 𝜀 and 𝑞 results in a shorter approach time. However, this also
leads to a sudden increase in motor speed, potentially causing it to slide out of the 𝑠 surface and resulting in
chattering. Furthermore, if the initial position 𝑆(0) is further from the sliding surface, the approach time will
significantly increase.
In sliding mode control, the chattering issue often occurs when the control system is operating near
the boundary of the sliding surface, where small perturbations or uncertainties can cause the control signal to
fluctuate rapidly between different values. This issue may lead to instability, degraded performance, and
increased energy consumption. To analyze the chattering issue, 𝑠̇ in the continuous domain is converted to the
discrete domain. As real-time control is indispensable in practical control system implementations, control
operations must be carried out within a discrete domain. From (14), the discrete expression of the conventional
reaching law method as 𝑠 approaches 0 is represented as in (16):

𝑠(𝑛 + 1) − 𝑠(𝑛) = −𝜀𝑇 𝑠𝑔𝑛(𝑠(𝑛)) (16)

where 𝑇 is the sampling time. Under the assumption that the system trajectory reaches the sliding-mode surface
within a finite step, which implies that 𝑠(𝑛) = 0+ and 𝑠(𝑛) = 0− , the equation for the subsequent period can
be derived with 𝑠(𝑛) = 0+ and 𝑠(𝑛) = 0− as in (17).

𝑠(𝑛 + 1) = −𝜀𝑇(𝑤ℎ𝑒𝑛𝑠(𝑛) = 0+ ); 𝑠(𝑛 + 1) = 𝜀𝑇(𝑤ℎ𝑒𝑛𝑠(𝑛) = 0− ) (17)

The width of the discrete sliding-mode band 𝛿1 is calculated between the boundary 𝑠(𝑛) = 0+ and 𝑠(𝑛) = 0− .
Figure 2 shows the state trajectory of the conventional reaching law. It can be seen in Figure 2 that the width
of the discrete sliding-mode band 𝛿1 in (18) causes the system to fail to reach the equilibrium point O and the
state trajectory oscillating around the point O. This result leads to a significant chattering problem when
increasing parameter 𝜀 to reduce the reaching time to the sliding surface.

𝛿1 = 2𝜀𝑇 (18)

Figure 2. State trajectory of the conventional reaching law

4. PROPOSED ENHANCED REACHING LAW DESIGN WITH FUZZY LOGIC CONTROLLER

4.1. The basic idea of enhanced reaching law integrated with fuzzy logic controller
To effectively address the chattering issue and improve the system response, it is essential to establish
two coefficients 𝑞 and 𝜀 as variables that can change according to the state variable 𝑥1 and the sliding surface

Int J Pow Elec & Dri Syst, Vol. 16, No. 1, March 2025: 418-430
Int J Pow Elec & Dri Syst ISSN: 2088-8694  423

position 𝑠. This forms the basis for proposing enhanced reaching law design with fuzzy logic controller for
adaptive adjusting of coefficients 𝑞 and 𝜀. Additionally, the gain 𝑓𝑢𝑧𝑧𝑦(𝑞) is adaptively adjusted using the
proposed fuzzy logic controller. Hence the proposed reaching law is given by the (19):

𝜀|𝑥1 |𝑎
𝑠̇ = − 𝑠𝑔𝑛( 𝑠) − 𝑓𝑢𝑧𝑧𝑦(𝑞)𝑠 (19)
𝑚|𝑥1 |𝑎 +𝑘𝑒 −𝜁|𝑠|

where 𝜀, 𝑚, 𝑎, 𝑘, 𝜁 are positive constants and 0 < 𝑚 < 1, 𝑓𝑢𝑧𝑧𝑦(𝑞) is the output of the fuzzy logic controller
according to the input 𝑠 and 𝑠̇ . The proposed reaching law is analyzed in two states: reaching the sliding mode
surface and operating near the boundary of the sliding surface. For simplicity, the system is analyzed when the
system state variable is positive.
In the reaching sliding mode surface state, the system state variable is located far away from the
sliding surface, so that 𝑘𝑒 −𝜁|𝑠| ≈ 0. Hence the (19) will converge to the (20).
𝜀
𝑠̇ ≈ − 𝑠𝑔𝑛( 𝑠) − 𝑓𝑢𝑧𝑧𝑦(𝑞)𝑠 (20)
𝑚

Compared to the conventional reaching law in (14), which the coefficient 𝑞 is a fixed number, the proposed
reaching law has the adaptive reaching gain 𝑓𝑢𝑧𝑧𝑦(𝑞). This fuzzy output is adaptively adjusted according to
the input 𝑠 and 𝑠̇ . Thanks to the adaptive adjustment of fuzzy output, the reaching time is improved with a high
gain 𝑓𝑢𝑧𝑧𝑦(𝑞), and this variable is reduced when approaching the sliding mode surface.
In the boundary operation of the sliding surface state, the following condition is satisfied 𝑠 → 0, and
the (19) will converge to the (21). The (21) depend on 𝑥1 , and its value is gradually reduced to 0 as 𝑥1 moves
to the origin. The proposed control law contributes to a smooth response when the system state 𝑥1 near the
sliding surface, which contribute to eliminate chattering issues. The combination of adaptive gains 𝑓𝑢𝑧𝑧𝑦(𝑞)
𝜀|𝑥1 |𝑎
in the reaching sliding mode surface state and in the boundary operation of the sliding surface, the
𝑚|𝑥1 |𝑎 +𝑘
state contributes to the improved system response and chattering reduction. Figure 3 shows the block diagram
of the proposed enhanced sliding mode control.

𝜀|𝑥1 |𝑎
𝑠̇ ≈ − 𝑠𝑔𝑛( 𝑠) (21)
𝑚|𝑥1 |𝑎 +𝑘

Figure 3. Block diagram of the proposed enhanced sliding mode control

4.2. The design of fuzzy(q) in fuzzy logic controller

The control variable 𝑓𝑢𝑧𝑧𝑦(𝑞) is the output of the fuzzy logic controller, allowing an adaptive tune
according to the state variables 𝑥1 and 𝑥2 . Based on the experience, the input coefficient of the state variables
𝑥1 and 𝑥2 range from [-10; 10], and the output 𝑓𝑢𝑧𝑧𝑦(𝑞) with gain ranges from 0 to 2000. The fuzzy set is
defined as follows: 𝑠 = {𝑁𝐵, 𝑁𝑀, 𝑁𝑆, 𝑍𝐸, 𝑃𝑆, 𝑃𝑀, 𝑃𝐵}; 𝑠̇ = {𝑁𝐵, 𝑁𝑀, 𝑁𝑆, 𝑍𝐸, 𝑃𝑆, 𝑃𝑀, 𝑃𝐵}, where negative
big (NB), negative medium (NM), negative small (NS), zero (ZE), positive small (PS), positive medium (PM)
, and positive big (PB) represent different levels of input variables. The output 𝑓𝑢𝑧𝑧𝑦(𝑞) = {𝑆, 𝑀𝑆, 𝑀, 𝑀𝐵, 𝐵},
where small (S), medium small (MS), medium (M), medium big (MB), and big (B) represent different levels
of output. The fuzzy rule table is provided in Table 1 and the relationship between 𝑠, 𝑠̇ , and the output 𝑓𝑢𝑧𝑧𝑦(𝑞)
is shown in Figures 4(a)-4(c). By using the membership functions and fuzzy rule table, the coefficient 𝑞 is
adaptively adjusted according to the value of 𝑠 and its derivatives. The 3D plot of 𝑞, 𝑠, 𝑑𝑠/𝑑𝑡, and their
relationship is shown in Figure 5. When 𝑠 and its derivative have high values, which indicates the state

Enhanced reaching law for improved response in sliding mode control of PMSM … (Khanh Quoc Truong)
424  ISSN: 2088-8694

trajactory is located far from the sliding surface, the coefficient 𝑞 is increased to enhance the reaching speed
to the sliding surface. On the other hand, when the value 𝑠 and its derivative are small, when means the state
trajactory is near the sliding surface, the coefficient 𝑞 is minimized to ensure that the state trajectory reaches
the sliding surface without overshoot.

Table 1. The rule base of the fuzzy controller

𝑠 𝑠̇
NB NM NS ZE PS PM PB
NB B MB M MS M MB B
NM MB M M M M M MB
NS MB M MS S MS M MB
ZE M MS S S S MS M
PS MB M MS S MS M MB
PM MB M M M M M MB
PB B MB M MS M MB B

(a) (b)

(c)

Figure 4. The membership function of the fuzzy controller for input 𝑠, 𝑠̇ , and 𝑓𝑢𝑧𝑧𝑦(𝑞):
(a) input 𝑠, (b) input 𝑠̇ , and (c) output 𝑓𝑢𝑧𝑧𝑦(𝑞)

Figure 5. The fuzzy logic control surface

Substituting (19) into (11), reference current 𝑖𝑞∗ is obtained as (22).

1 𝜀|𝑥1 |𝑎
𝑖𝑞∗ = ∫ ( 𝑎 −𝜁|𝑠| 𝑠𝑔𝑛( 𝑠) + 𝑓𝑢𝑧𝑧𝑦(𝑞)𝑠 + 𝑐𝑥2 ) 𝑑𝑡 (22)
𝐾 𝑚|𝑥 1 | +𝑘𝑒

In the boundary operation of the sliding surface state, the sliding mode surface s approaches 0, and the
proposed enhanced reaching law in (19) can be converted to the discrete-time domain as (23):

Int J Pow Elec & Dri Syst, Vol. 16, No. 1, March 2025: 418-430
Int J Pow Elec & Dri Syst ISSN: 2088-8694  425

𝜀𝑇|𝑥1 |𝑎
𝑠(𝑛 + 1) − 𝑠(𝑛) = − 𝑠𝑔𝑛(𝑠(𝑛)) (23)
𝑚|𝑥1 |𝑎 +𝑘

where 𝑇 represents the sampling period. Under the assumption that the system's trajectory reaches the sliding-
mode surface within a finite step, which implies that 𝑠(𝑛) = 0+ ,𝑠(𝑛) = 0− , the equation for the subsequent
period can be derived with 𝑠(𝑛) = 0+ and 𝑠(𝑛) = 0− as in (24).

𝜀𝑇|𝑥1 |𝑎 𝜀𝑇|𝑥1 |𝑎
𝑠(𝑛 + 1) = − (𝑤ℎ𝑒𝑛𝑠(𝑛) = 0+ ); 𝑠(𝑛 + 1) = (𝑤ℎ𝑒𝑛𝑠(𝑛) = 0− ) (24)
𝑚|𝑥1 |𝑎 +𝑘 𝑚|𝑥1 |𝑎 +𝑘

The width of the discrete sliding-mode band 𝛿2 is calculated between the boundary 𝑠(𝑛) = 0+ and 𝑠(𝑛) = 0− .
Compared to the conventional reaching law, in which the width of the discrete sliding-mode band 𝛿1 in (18) is
a constant value, the proposed method has an adaptive width of the discrete sliding-mode band 𝛿2 in (25),
which reduces the chattering issue. To illustrate the effectiveness of the proposed method, the sliding-mode
band is shown in Figure 6. In Figure 6, the sliding surface 𝑠 trajectory gradually approaches 0, 𝑥1 and 𝑥2 also
approaches 0. It can be seen that discrete sliding-mode band 𝛿2 reduces to 0, rather than oscillating around the
origin 0 of the conventional method in Figure 2.

2𝜀𝑇|𝑥1 |𝑎
𝛿2 = (25)
𝑚|𝑥1 |𝑎 +𝑘

Figure 6. State trajectory of the proposed enhanced reaching law

4.3. Stability analysis

To assess the stability of the PMSM systems, the Lyapunov stability criterion and stability analysis
are employed [30]. In the Lyapunov theory, the condition for the state variables to reach the sliding surface
𝑠 = 𝑥2 + 𝑐𝑥1 is that the condition 𝑉̇ < 0 must be satisfied. In this study, we choose the Lyapunov function as
a quadratic function as in (26). Taking the derivative of the (26), we obtain (27).
1
𝑉 = 𝑠2 (26)
2

𝑉̇ = 𝑠𝑠̇ (27)

Substituting (19) to (27), we get (28):

𝜀|𝑥1 |𝑎
𝑉̇ = 𝑠 [− 𝑠𝑔𝑛( 𝑠) − 𝑓𝑢𝑧𝑧𝑦(𝑞)𝑠] →
𝑚|𝑥1 |𝑎 + 𝑘𝑒 −𝜁|𝑠|
𝜀|𝑥1 |𝑎
𝑉̇ = − |𝑠| − 𝑓𝑢𝑧𝑧𝑦(𝑞)𝑠 2 < 0 (28)
𝑚|𝑥1 |𝑎 +𝑘𝑒 −𝜁|𝑠|

where 0 < 𝑚 < 1, 𝜀 > 0, 𝑚 > 0, 𝑘 > 0, and 𝑓𝑢𝑧𝑧𝑦(𝑞)𝑠 2 > 0. The (28) guarantees the Lyapunov stability
criteria 𝑉̇ < 0. Therefore, the proposed reaching law is well-suited for the PMSM motor speed control system,
ensuring both stability and enhanced response.

Enhanced reaching law for improved response in sliding mode control of PMSM … (Khanh Quoc Truong)
426  ISSN: 2088-8694

5. EXPERIMENTAL RESULTS
To verify the effectiveness of the proposed control scheme, an experimental hardware platform for
the PMSM drive system is set up. Texas Instruments microcontroller TMS280S28379D is adapted to apply the
control algorithm and generate the PWM signals for controlling the PMSM motor. Figure 7 shows the
experimental hardware platform, and the PMSM motor parameters are listed in Table 2. The control parameters
of the conventional and proposed reaching law are listed in Table 3. For comparison, the conventional PI
controller and fuzzy logic controller, which focuses on computational efficiency and real-time control, are
adopted [31], [32].

Figure 7. Experimental hardware platform for PMSM drive system

Table 2. Parameters of the PMSM motor

Symbol Parameter Value
𝑝 Number of pole pairs 4
𝐿𝑑 , 𝐿𝑞 d-axis and q-axis inductance 0.59 𝑚𝐻
𝑅𝑠 Coil resistance 1.02 Ω
𝐽 Inertia moment 0.28 𝑘𝑔. 𝑚𝑚2
𝜓𝑓 Permanent magnet flux linkage 0.042 𝑊𝑏
𝑛𝑁 Rated speed 3000 𝑟𝑝𝑚
𝑉𝑁 Rated voltage 24 𝑉𝑑𝑐
𝐼𝑁 Rated current 4𝐴

Table 3. Parameters of the conventional and proposed reaching law

Symbol Parameter Conventional reaching law
N/A Conventional reaching law 𝑠̇ = −𝜀 𝑠𝑔𝑛( 𝑠) − 𝑞𝑠
𝜀|𝑥1 |𝑎
N/A Proposed reaching law 𝑠̇ = − 𝑠𝑔𝑛( 𝑠) − 𝑓𝑢𝑧𝑧𝑦(𝑞)𝑠
𝑚|𝑥1 |𝑎 + 𝑘𝑒 −𝜁|𝑠|
𝑞 Reaching law coefficient 500
𝜀 Constant rate term 300
𝑘 Error gain 0.1
𝑎 Exponential error term 2
𝑚 Input control gain 0.05
𝜁 Error exponential coefficient 100
𝑐 Sliding mode gain 100
𝐾𝑃 Proportional gain 0.5
𝐾𝐼 Integral gain 5

Figures 8(a)-8(d) illustrate the observed speed response as the PMSM speed increased from 500 rpm
to 1000 rpm using a conventional PI controller, fuzzy logic controller, conventional sliding mode, and proposed
sliding mode controller. The conventional PI controller (𝐾𝑃 = 0.008, 𝐾𝐼 = 0.0005), as shown in Figure 8(a),
had a response time of 1.1 seconds. Although the PI controller reached the desired speed, it exhibited a
significant speed overshoot of 100 rpm. This level of overshoot indicates that while the PI controller can handle
the speed change, it cannot completely resolve the nonlinearity effect of the PMSM motor. As seen in
Figure 8(b), the fuzzy logic controller outperformed the PI controller, with a rise time of less than 0.7 seconds
and a reduced speed overshoot of about 44 rpm. The faster response and smaller overshoot demonstrate the
fuzzy logic controller better adaptability to the inherent nonlinearity of the PMSM motor.

Int J Pow Elec & Dri Syst, Vol. 16, No. 1, March 2025: 418-430
Int J Pow Elec & Dri Syst ISSN: 2088-8694  427

In Figure 8(c), the conventional sliding mode control has a response time of 0.9 seconds, which is
faster than the PI controller but slower than the fuzzy logic controller. Furthermore, the overshoot of
conventional sliding mode control is smaller than that of the PI controller. Figure 8(d) shows the response of
proposed sliding mode controller. The response time in Figure 8(d) was around 0.4 seconds, which is lower
than PI controller fuzzy logic and also conventional sliding mode control. Also, the speed overshoot was small
at only 30 rpm which is way lower than all other controllers. The results obtained reveal that the sliding mode
control approach is capable of achieving high speed response with lessed overshoot in comparison of
conventional techniques. To evaluate the control performance under different speed conditions, the PMSM
speed reference is increased suddenly from 1000 rpm to 1500 rpm, and the results are shown in Figure 9. In
Figure 9(a), the response time of the PI controller comes up to be around 1.1 seconds, which is lower compared
to all other controllers, and it has a slightly higher speed overshoot of around 100 rpm. The fuzzy logic
controller again performed better than the PI mode in Figure 9(b) with a response time of 0.7 seconds and an
overshoot of 44 rpm. While this was faster and more stable compared to the PI controller, the overshoot still
exist which may be caused by an inadequate membership function.
The response time and overshoot for the conventional sliding mode control is 0.9 s and 85 rpm shown
in Figure 9(c). It is better than the PI controller but not alike to fuzzy logic and sliding mode control. As shown
in Figure 9(d), the new designed sliding mode controller also had satisfactory tracking performance with a
response time of only 0.4 seconds and slight overshooting speed up to about 30 rpm. Table 4 provides a
comprehensive comparison between conventional PI, fuzzy logic controller, conventional sliding mode
controller, and the proposed sliding mode controller. The experimental results clearly show that the proposed
controller has better performance than conventional PI, fuzzy controller, and conventional sliding mode control
methods in terms of response and overshoot. These results validate the proposed controller effectiveness across
different speed ranges, maintaining a consistent control performance over the conventional methods. The
proposed control approach provides notable advantages in regulating motor speed dynamics, surpassing the
conventional technique in different speed ranges.

Figure 8. Experimental speed response comparison of the PI controller, fuzzy logic controller, conventional
sliding mode, and proposed sliding mode controller as the PMSM speed increases from 500 rpm to
1000 rpm: (a) PI controller, (b) fuzzy logic controller, (c) conventional reaching law, and
(d) proposed sliding mode controller

Enhanced reaching law for improved response in sliding mode control of PMSM … (Khanh Quoc Truong)
428  ISSN: 2088-8694

Figure 9. Experimental speed response comparison of the PI controller, fuzzy logic controller, conventional
sliding mode, and proposed sliding mode controller as the PMSM speed increases from 1000 rpm to
1500 rpm: (a) PI controller, (b) fuzzy logic controller, (c) conventional reaching law, and
(d) proposed sliding mode controller

Table 4. Comparison of PMSM speed control methods

Methods Criterion
Complexity Overshoot Response Stability Realtime implementation
in microcontroller
Conventional PI controller Simple High Slow Depend Yes
Fuzzy logic controller Medium Small Medium Depend Yes
Conventional sliding mode control Medium Medium Medium Good Yes
Proposed enhanced sliding mode control Medium Small High Good Yes

6. CONCLUSION
This study presents an enhanced sliding mode reaching law with fuzzy logic that can simultaneously
reduce the chattering effect and improve system response time. The main contributions of this paper consist in
establishing a mathematical model of the PMSM and analyzing conventional reaching law limitation.
Additionally, the paper presents a modified reaching law model through a fuzzy controller for better speed
control performance. By adopting the proposed reaching law, the sliding gains are allowed for adjustable
sliding gain, which improves the system trajectory to reach the sliding surface quickly. When the state
trajectory is located near the sliding surface, the sliding gain is minimized to ensure that the state trajectory
approach the sliding surface without overshoot. As a result, the integration of fuzzy and sliding mode
significantly reduces the chattering, resulting in smoother and more stable motor speed control. In addition, the
reachability condition ensures system stability, which is proved by Lyapunov stability theory. A small-scale
PMSM experimental system has confirmed the effectiveness of this method. Also, experimental validations
are conducted in various speed ranges, which further support the above results.

ACKNOWLEDGEMENTS
We acknowledge Ho Chi Minh City University of Technology (HCMUT), VNU-HCM for supporting
this study.

Int J Pow Elec & Dri Syst, Vol. 16, No. 1, March 2025: 418-430
Int J Pow Elec & Dri Syst ISSN: 2088-8694  429

REFERENCES
[1] X. Sun, Z. Shi, G. Lei, Y. Guo, and J. Zhu, “Analysis and design optimization of a permanent magnet synchronous motor for a
campus patrol electric vehicle,” IEEE Transactions on Vehicular Technology, vol. 68, no. 11, pp. 10535–10544, Nov. 2019, doi:
10.1109/TVT.2019.2939794.
[2] H. Chen, A. M. El-Refaie, and N. A. O. Demerdash, “Flux-Switching Permanent Magnet Machines: A Review of Opportunities
and Challenges - Part I: Fundamentals and Topologies,” IEEE Transactions on Energy Conversion, vol. 35, no. 2, pp. 684–698,
Jun. 2020, doi: 10.1109/TEC.2019.2956600.
[3] P. Ramesh, M. Umavathi, C. Bharatiraja, G. Ramanathan, and S. Athikkal, “Development of a PMSM motor field-oriented control
algorithm for electrical vehicles,” Materials Today: Proceedings, vol. 65, pp. 176–187, 2022, doi: 10.1016/j.matpr.2022.06.080.
[4] Y. Lu, Z. Jiang, C. Chen, and Y. Zhuang, “Energy efficiency optimization of field-oriented control for PMSM in all electric system,”
Sustainable Energy Technologies and Assessments, vol. 48, p. 101575, Dec. 2021, doi: 10.1016/j.seta.2021.101575.
[5] M. Zhao, Q. An, C. Chen, F. Cao, and S. Li, “Observer based improved position estimation in field-oriented controlled PMSM with
Misplaced Hall-Effect Sensors,” Energies (Basel), vol. 15, no. 16, p. 5985, Aug. 2022, doi: 10.3390/en15165985.
[6] Q. Chen, S. Kang, L. Zeng, Q. Xiao, C. Zhou, and M. Wu, “PMSM control for electric vehicle based on fuzzy PI,” International
Journal of Electric and Hybrid Vehicles, vol. 12, no. 1, p. 75, 2020, doi: 10.1504/IJEHV.2020.104251.
[7] R. Cai, R. Zheng, M. Liu, and M. Li, “Optimal selection of PI parameters of FOC for PMSM using structured H∞-synthesis,” in
IECON 2017 - 43rd Annual Conference of the IEEE Industrial Electronics Society, IEEE, Oct. 2017, pp. 8602–8607. doi:
10.1109/IECON.2017.8217511.
[8] P. Q. Khanh and H. P. H. Anh, “Hybrid optimal fuzzy Jaya technique for advanced PMSM driving control,” Electrical Engineering,
vol. 105, no. 6, pp. 3629–3646, Dec. 2023, doi: 10.1007/s00202-023-01911-6.
[9] R. S. Widagdo, B. Hariadi, and I. A. Wardah, “Simulation of speed control on a PMSM using a PI controller,” Jambura Journal of
Electrical and Electronics Engineering, vol. 6, no. 1, pp. 63–69, Jan. 2024, doi: 10.37905/jjeee.v6i1.22287.
[10] J. M. Lazi, Z. Ibrahim, M. H. N. Talib, A. Alias, A. Nur, and M. Azri, “Speed and position estimator of for sensorless PMSM drives
using adaptive controller,” International Journal of Power Electronics and Drive Systems (IJPEDS), vol. 10, no. 1, pp. 128–136,
2019, doi: 10.11591/ijpeds.v10.i1.pp128-136.
[11] Z. Li, W. Zhang, G. Liu, B. Wang, and Y. Zhang, “A novel integral-proportional (I-P) speed controller in PMSM motor drive,” in
Proceeding of the 11th World Congress on Intelligent Control and Automation, IEEE, Jun. 2014, pp. 4236–4241. doi:
10.1109/WCICA.2014.7053425.
[12] L. Wang, M. Tian, and Y. Gao, “Fuzzy self-adapting PID control of PMSM servo system,” in 2007 IEEE International Electric
Machines & Drives Conference, IEEE, May 2007, pp. 860–863. doi: 10.1109/IEMDC.2007.382781.
[13] J.-W. Jung, V. Q. Leu, T. D. Do, E.-K. Kim, and H. H. Choi, “Adaptive PID speed control design for permanent magnet synchronous
motor drives,” IEEE Transactions on Power Electronics, vol. 30, no. 2, pp. 900–908, Feb. 2015, doi: 10.1109/TPEL.2014.2311462.
[14] C. Youssef, Z. Said, and D. Abdelkarim, “A review of non-linear control methods for permanent magnet synchronous machines
(PMSMs),” in The International Conference on Artificial Intelligence and Smart Environment. Cham: Springer Nature Switzerland,
2024, pp. 446–452. doi: 10.1007/978-3-031-48465-0_59.
[15] Y. Chaou, S. Ziani, H. Ben Achour, and A. Daoudia, “Nonlinear control of the permanent magnet synchronous motor PMSM using
Backstepping method,” WSEAS TRANSACTIONS on SYSTEMS and CONTROL, vol. 17, pp. 56–61, Jan. 2022, doi:
10.37394/23203.2022.17.7.
[16] X. Liu, H. Yu, J. Yu, and L. Zhao, “Combined speed and current terminal sliding mode control with nonlinear disturbance observer
for PMSM drive,” IEEE Access, vol. 6, pp. 29594–29601, 2018, doi: 10.1109/ACCESS.2018.2840521.
[17] P. Mani, R. Rajan, L. Shanmugam, and Y. H. Joo, “Adaptive fractional fuzzy integral sliding mode control for PMSM model,”
IEEE Transactions on Fuzzy Systems, vol. 27, no. 8, pp. 1674–1686, Aug. 2019, doi: 10.1109/TFUZZ.2018.2886169.
[18] H. Qiu, H. Zhang, L. Min, T. Ma, and Z. Zhang, “Adaptive control method of sensorless permanent magnet synchronous motor
based on super-twisting sliding mode algorithm,” Electronics (Basel), vol. 11, no. 19, p. 3046, Sep. 2022, doi:
10.3390/electronics11193046.
[19] J. Liu, S. Cao, Y. Yi, and Y. Fang, “Adaptive control for permanent magnet synchronous motor based on disturbance observer,” in
2017 36th Chinese Control Conference (CCC), IEEE, Jul. 2017, pp. 3533–3537. doi: 10.23919/ChiCC.2017.8027905.
[20] X. Zhang, L. Zhang, and Y. Zhang, “Model predictive current control for PMSM drives with parameter robustness improvement,”
IEEE Trans Power Electron, vol. 34, no. 2, pp. 1645–1657, Feb. 2019, doi: 10.1109/TPEL.2018.2835835.
[21] S. Niu, Y. Luo, W. Fu, and X. Zhang, “Robust model predictive control for a three-phase PMSM motor with improved control
precision,” IEEE Transactions on Industrial Electronics, vol. 68, no. 1, pp. 838–849, Jan. 2021, doi: 10.1109/TIE.2020.3013753.
[22] Z. Wang, A. Yu, X. Li, G. Zhang, and C. Xia, “A novel current predictive control based on fuzzy algorithm for PMSM,” IEEE
Journal of Emerging and Selected Topics in Power Electronics, vol. 7, no. 2, pp. 990–1001, Jun. 2019, doi:
10.1109/JESTPE.2019.2902634.
[23] A. Wang and S. Wei, “Sliding mode control for permanent magnet synchronous motor drive based on an improved exponential
reaching law,” IEEE Access, vol. 7, pp. 146866–146875, 2019, doi: 10.1109/ACCESS.2019.2946349.
[24] Y. Wang, Y. Feng, X. Zhang, and J. Liang, “A new reaching law for antidisturbance sliding-mode control of PMSM speed regulation
system,” IEEE Trans Power Electron, vol. 35, no. 4, pp. 4117–4126, Apr. 2020, doi: 10.1109/TPEL.2019.2933613.
[25] K. M. Naikawadi, S. M. Patil, K. Kalantri, and M. R. Dhanvijay, “Comparative analysis of features of online numerical methods
used for parameter estimation of PMSM,” International Journal of Power Electronics and Drive Systems (IJPEDS), vol. 13, no. 4,
pp. 2172–2180, 2022.
[26] Mohd. Marufuzzaman, M. B. I. Reaz, and M. A. Mohd. Ali, “FPGA implementation of an intelligent current dq PI controller for
FOC PMSM drive,” in 2010 International Conference on Computer Applications and Industrial Electronics, IEEE, Dec. 2010, pp.
602–605. doi: 10.1109/ICCAIE.2010.5735005.
[27] A. Kumar and D. Chatterjee, “A survey on space vector pulse width modulation technique for a two-level inverter,” in 2017 National
Power Electronics Conference, NPEC 2017, vol. 2018-January, pp. 78–83, Jul. 2017, doi: 10.1109/NPEC.2017.8310438.
[28] D. Venkata Ramana, P. Chandrasekar, R. Sreedhar, K. Karunanithi, K. Aruna Rani, and A. Vijayakumar, “Performance analysis of
three-phase Cuk inverter with PWM and SPWM modulation based on power quality,” in Sixth International Conference on Intelligent
Computing and Applications: Proceedings of ICICA 2020, Springer, 2021, pp. 91–99, doi: 10.1007/978-981-16-1335-7_9.
[29] R. Sreedhar, K. Karunanithi, P. Chandrasekar, and R. B. Teja, “Nearest space-vector control strategy for high-resolution multilevel
inverters,” in 2021 6th International Conference for Convergence in Technology (I2CT), IEEE, 2021, pp. 1–6, doi:
10.1109/I2CT51068.2021.9417882.
[30] Y. Li, J. Zhang, and Q. Wu, Adaptive Sliding Mode Neural Network Control for Nonlinear Systems. Academic Press 2018.

Enhanced reaching law for improved response in sliding mode control of PMSM … (Khanh Quoc Truong)
430  ISSN: 2088-8694

[31] B. Adhavan, A. Kuppuswamy, G. Jayabaskaran, and V. Jagannathan, “Field oriented control of permanent magnet synchronous
motor (PMSM) using fuzzy logic controller,” in 2011 IEEE Recent Advances in Intelligent Computational Systems, IEEE, Sep.
2011, pp. 587–592. doi: 10.1109/RAICS.2011.6069379.
[32] H. Celik and T. Yigit, “Field-oriented control of the PMSM with 2-DOF PI controller tuned by using PSO,” in 2018 International
Conference on Artificial Intelligence and Data Processing (IDAP), IEEE, Sep. 2018, pp. 1–4. doi: 10.1109/IDAP.2018.8620902.

BIOGRAPHIES OF AUTHORS

Khanh Quoc Truong is a researcher at the Faculty of Electrical and Electronic

Engineering, Ho Chi Minh City University of Technology, Vietnam. His research interests
include control theory, PMSM, embedded systems, and power electronics. He can be contacted
at email: [email protected].

Son Huynh is a researcher at the Faculty of Electrical and Electronic Engineering,

Ho Chi Minh City University of Technology, Vietnam. His research interests include control
theory, PMSM, embedded systems, and power electronics. He can be contacted at email:
[email protected].

Dung Hoang Vo is a researcher at the Faculty of Electrical and Electronic

Engineering, Ho Chi Minh City University of Technology, Vietnam. His research interests
include control theory, PMSM, embedded systems, and power electronics. He can be contacted
at email: [email protected].

Minh Duc Pham received the master and Ph.D. degrees in Electrical Engineering
from Ulsan University, South Korea. He is currently a full-time lecturer at Ho Chi Minh City
University of Technology, Vietnam. His research interests include hybrid AC-DC microgrids,
PMSM, low-cost inverters, and renewable energy. He can be contacted at email:
[email protected].

Int J Pow Elec & Dri Syst, Vol. 16, No. 1, March 2025: 418-430