Hindawi

Journal of Control Science and Engineering

Volume 2019, Article ID 3612634, 13 pages
https://doi.org/10.1155/2019/3612634

Research Article
Explicit GPC Control Applied to an Approximated Linearized
Crane System

Daniel Guerra Vale da Fonseca ,1,2 André Felipe O. de A. Dantas ,3

Carlos Eduardo Trabuco Dórea ,1 and André Laurindo Maitelli 1

1
Department of Automation and Computer Engineering, CT, Federal University of Rio Grande do Norte, 59078-970 Natal, RN, Brazil
2
Federal Institute of Education, Science and Technology of Rio Grande do Norte, Natal 59112-490, Brazil
3
Master of Process Engineering, Potiguar University, Natal 59054-180, Brazil

Correspondence should be addressed to Daniel Guerra Vale da Fonseca; dangvf@gmail.com

Received 5 October 2018; Revised 10 December 2018; Accepted 18 December 2018; Published 3 February 2019

Guest Editor: Umer H. Shah

Copyright © 2019 Daniel Guerra Vale da Fonseca et al. This is an open access article distributed under the Creative Commons
Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is
properly cited.

This paper proposes a MIMO Explicit Generalized Predictive Control (EGPC) for minimizing payload oscillation of a Gantry Crane
System subject to input and output constraints. In order to control the crane system efficiently, the traditional GPC formulation,
based on online Quadratic Programming (QP), is rewritten as a multiparametric quadratic programming problem (mp-QP). An
explicit Piecewise Affine (PWA) control law is obtained and holds the same performance as online QP. To test effectiveness, the
proposed method is compared with two GPC formulations: one that handle constraints (CGPC) and another that does not handle
constraints (UGPC). Results show that both EGPC and CGPC have better performance, reducing the payload swing when compared
to UGPC. Also both EGPC and CGPC are able to control the system without constraint violation. When comparing EGPC to
CGPC, the first is able to calculate (during time step) the control action faster than the second. The simulations prove that the
overall performance of EGPC is superior to the other used formulations.

1. Introduction Among crane applications, Gantry Crane Systems (GCS)

For decades cranes have been playing an essential role as a
transportation tool for objects in various areas [1]. Industry
and academia have studied its application in construction,
steelworks, shipyards, and offshore container crane [2], due
to its operational flexibility and high payload capacity [3].
However, the crane trolley rapid movement causes an exces-
sive payload oscillation, which is an obstacle that limits safe
and fast transportation. This swing happens during or even
at the end of trolley movement due to crane dynamics and
external disturbances [4]. Hence, the main issue involving
crane system rests in the ability to quickly move the trolley,
in order to perform load-carrying tasks in a short period;
at the same time undesirable fluctuations of the payload
are decreased, using an efficient control system [5–7]. This
is a challenge task, since crane control systems usually are
underactuated [8].
are highlighted as the most preferred application in industries
[9, 10]. This type of crane is mobile and is supported by a
pair of rigid steel legs, being frequently used in factories,
shipyards, and warehouses [11]. In GCS, the trolley mass,
payload mass, and cable length directly influence payload
swing and trolley displacement [12].
To control the GCS and other crane types, several strate-
gies may be applied. Reference [11] brings a very deep and
complete review about these techniques over the last two
decades. They are categorized into Open Loop, Closed Loop,
and Hybrid (combines Open and Closed Loop) strategies, as
depicted through Figure 1. About open loop schemes, they are
easy to implement and less costly, since there is no need for
sway angle or cart position sensors. Among them, the input
shaping was widely used by researchers, being present in
several works [13–15]. Despite being an efficient feed-forward
control, it has as main drawback the sensitive to external
2 Journal of Control Science and Engineering

Input Shaping

Open Loop Filter

Command
PID
Smoothing

Linear Quadratic
Linear Control
Regulation

State Feedback
Controller
Crane Control
Fuzzy Logic
Intelligent Control
Neural Network

Model Predictive
Closed Loop
Control

Combined Open Linear Quadratic

Loop and Closed Optimal Control
Gaussian
Loop (Hybrid)
Generalized
Adaptive Control
Predictive Control

Sliding Mode Control Gain Scheduling

Other control H-Infinity

schemes

-synthesis

Figure 1: Crane control strategies [11].

disturbance. Because of this, it may be combined with Closed
Loop ones, as seen in [16, 17].
The other types of schemes used into crane operations
are the feedback ones. This type of technique uses the desired
output response to adjust their performance; hence they are
less sensitive to disturbances. However, they are also slow, due
to the input delay in the feedback loop.
Many control loop strategies were used to improve a GCS
performance. Some examples are the state feedback controller
(SFB) using the Ackerman formula, the PID controller
[18], sliding mode control (SMC), adaptive controllers [19],
Fuzzy Control System [20], and Linear Quadratic Regulator
(LQR) [21]. However, the aforementioned controllers are not
fully suitable, since most of them do not consider system
constraints, and they do not have their parameters obtained
from an optimization problem. In this context, the MPC is a
reasonable choice, because the constraints are included into
its formulation. Moreover, the gantry crane can be considered
a Multiple Input Multiple Output (MIMO) process [10],
which is easily dealt by MPC controllers. It should also
be mentioned that high-performance controllers, based on
solutions of an online optimization problem, like MPCs, are
often computationally costly, mainly when applied to MIMO
processes.
Therefore, this work proposes to obtain an explicit
formulation for the Multivariable Generalized Predictive
Control (GPC) based into a multiparametric programming
and applying it to a Gantry Crane System (GCS) subject to
input and output constraints. The used crane system model is
based on [9, 22]. The proposed controller is compared with
two conventional GPC control strategies: one that does not
handle constraints but uses an analytical solution from cost
function; and another one that uses Quadratic Programming
(QP) to calculate the control action under constraints.
This paper proposal differs from other MPC crane appli-
cations because it is able to calculate the control actions
(at runtime), more quickly than when obtained from online
optimization. Also, a multiparametric GPC formulation is an
under-explored topic in literature, with few examples [23, 24].
The remainder of this paper is organized as follows: in
Section 2, a GCS concept is deepened, where we explain
the used crane model and its linearized form. After that,
Journal of Control Science and Engineering 3

x The gantry crane motion equation, considering the DC

motor dynamic, is given by
Ｇ2
F
𝑅𝑟𝑝 𝐾𝑒 𝑟 𝑑𝑅𝑟𝑝
𝑉= (𝑚1 + 𝑚2 ) 𝑥̈ + ( + ) 𝑥̇
𝐾𝑡 𝑟 𝑟𝑝 𝐾𝑡 𝑟
(3)
𝑚1 𝑙𝑅𝑟𝑝
 + (𝜃̈ cos 𝜃 − 𝜃2̇ sin 𝜃)
l 𝐾𝑡 𝑟

where 𝑉, 𝑅, 𝐾𝑡 , 𝐾𝑒 , 𝑟, and 𝑟𝑝 are the input voltage, resistance,

torque constant, electric constant, gear ratio, and pulley
Ｇ1
radius, respectively.
Equation (3) can be rewritten as follows:
Figure 2: Schematic diagram of GCS.
𝑉 = 𝐾1 𝑥̈ + 𝐾2 𝑥̇ + 𝐾3 (𝜃̈ cos 𝜃 − 𝜃2̇ sin 𝜃) (4)

we address the traditional GPC controller, with and without where

constraints, in Sections 3 and 4. In order to design the explicit
𝑅𝑟𝑝
controller, in Section 5, the algorithm to obtain the Piecewise 𝐾1 = (𝑚1 + 𝑚2 ) ,
Affine control law parameters is described. In sequence, 𝐾𝑡 𝑟
a numerical example of a crane is used to highlight the
proposed controller behavior and compare it to the other 𝐾𝑒 𝑟 𝑑𝑅𝑟𝑝
𝐾2 = ( + ) (5)
GPC strategies (Section 6). Finally, the conclusions of the 𝑟𝑝 𝐾𝑡 𝑟
paper are presented.
𝑚1 𝑙𝑅𝑟𝑝
and 𝐾3 =
2. Gantry Crane Dynamics 𝐾𝑡 𝑟

Crane dynamic model has been an important research area
over the years for controller design [4, 13, 22, 25]. Because
the crane controller needs to quickly move the trolley at
the same time it decreases undesirable fluctuations of the
payload, a model that describes crane dynamics is very
important to understand how the relationship between the
trolley movement and the payload swing works.
Among crane systems, the Gantry Crane System (GCS)
is one of the most preferred in industries [9, 10]. In this
application, the trolley mass, payload mass, and cable length
influence payload swing and trolley displacement [12] as
shown in Figure 2.
Figure 2 shows, along one axis, a GCS scheme where 𝑥 is
the trolley horizontal displacement with mass 𝑚2 and 𝜃 is the
payload oscillation angle. The payload with mass 𝑚1 (subject
to gravity action 𝑔) is connected to the cart with a cable of
As depicted, the deflection angle can be considered
sufficiently small during the crane system control. This allows
the model linearization since 𝜃 near zero imply sin 𝜃 ≈ 𝜃,
cos 𝜃 ≈ 1 and 𝜃 ̇ = 0. Therefore, (2) and (4) become
𝑉 = 𝐴𝑥̈ + 𝐵𝑥̇ + 𝐶𝜃 ̈ (6)
𝑙𝜃 ̈ + 𝑥̈ + 𝑔𝜃 = 0 (7)
These ODEs represent crane dynamic while the model is
linearized at 𝜃 = 0. Thus, a single input two outputs system
can be obtained from (6) and (7), with the following transfer
functions:
𝑋 (𝑠) 𝑙𝑠2 + 𝑔
= (8)
𝑉 (𝑠) (𝐴𝑙 − 𝐶) 𝑠 + 𝐵𝑙𝑠3 + 𝐴𝑔𝑠2 + 𝐵𝑔𝑠
4

length 𝑙. While the cart is in motion, the trolley is subject to 𝜃 (𝑠) −𝑠

damping friction 𝑑. Note that, in this system, the number of
inputs is lower than the number of outputs which means this
is an underactuated system.
The GCS can be represented by a nonlinear model as
shown in (1) and (2):
(𝑚1 + 𝑚2 ) 𝑥̈ + 𝑚1 𝑙𝜃 ̈ cos 𝜃 − 𝑚1 𝑙𝜃2̇ sin 𝜃 + 𝑑𝑥̇ = 𝐹
𝑚1 𝑙2 𝜃̈ + 𝑚1 𝑙𝑥̈ cos 𝜃 + 𝑚1 𝑔𝑙 sin 𝜃 = 0
The driving force 𝐹 that moves the trolley is provided by
a DC motor, where the input voltage is limited. In this paper,
the system used by [22] will be taken into consideration,
where the DC motor is limited by ±5 volts and the crane
horizontal track length for trolley movement is about 0.8m.
= (9)
𝑉 (𝑠) (𝐴𝑙 − 𝐶) 𝑠3 + 𝐵𝑙𝑠2 + 𝐴𝑔𝑠 + 𝐵𝑔
In order to apply a predictive controller to the Gantry
Crane System, some assumptions shall be considered, as
follows:
(1) (1) the payload is considered a point mass and works like
a pendulum;
(2)
(2) the mass and stiffness of the hoisting rope are
neglected;
(3) the effects of wind disturbances are not considered.
The information shown into Table 1 summarize all the
parameter values used for the system based on [9, 22].
4 Journal of Control Science and Engineering

Table 1: Gantry Crane model parameters.

Parameter Value Parameter Value

Payload mass (𝑚1 ) 2.5 Kg Resistance (𝑅) 2.6Ω
Trolley mass (𝑚2 ) 2.0 Kg Torque constant (𝐾𝑡 ) 0.007 Nm/A
Cable length (𝑙) 0.6 m Electric constant (𝐾𝑒 ) 0.007 Vs/rad
Gravitational Acceleration (𝑔) 9.81m/s2 Radius of pulley (𝑟𝑝 ) 0.02 m
Damping coefficient (𝑑) 0.001 Ns/m Gear ratio (𝑟) 0.15
DC Motor voltage limits ±5𝑉 Horizontal track length ≈ 0.8m

Since proposed GPC and the other used techniques The following GPC algorithm consists of applying a
are formulated in discrete-time, the system model also is control sequence capable of minimizing the following cost
discretized. In this case, the sample time 𝑇𝑠 is chosen function:
as 0.1 seconds. Note that MPC controllers may need a 𝑁𝑝
high computational time. Therefore a conservative choice of 󵄩 󵄩2
𝐽 = ∑ 󵄩󵄩󵄩𝑦̂ (𝑘 + 𝑗 | 𝑘) − 𝑤 (𝑘 + 𝑗)󵄩󵄩󵄩𝑅
sample time, that do not compromise the performance, is 𝑗=1
made. (11)
𝑁𝑢
󵄩 󵄩2
+ ∑ 󵄩󵄩󵄩Δ𝑢 (𝑘 + 𝑗 − 1)󵄩󵄩󵄩𝑄
3. Multivariable GPC Formulation 𝑗=1

In this section, the Generalized Predictive Control (GPC) ̂

formulation is discussed. This controller is one of the
most famous Model Predictive Control (MPC) methods. Its
fundamental idea is to calculate a control sequence in a
control horizon (𝑁𝑢 ) in order to minimize a cost function
defined over a prediction horizon (𝑁𝑝 ). Some MPC are
also able to handle input, output, and states constraints.
Despite being similar to other predictive controllers as seen
in [26, 27], the GPC calculates an analytical solution for a
minimization problem without constraints. Additionally, this
controller is capable of dealing with MIMO systems, as well
as controlling both unstable and nonminimum phase system
[28].
Recently, MPC has still been relevant in different crane
control approaches as [29–33]. Concerning GPC, [34, 35]
used this strategy to control an overhead crane system, while
[36] applied a GPC to control an offshore crane operation.
Hence, in this section, the GPC described in [28] is adapted
and rewritten to control crane systems.
Consider the CARIMA model for a multivariable process
with 𝑚 inputs and 𝑛 outputs described as follows:
1 Δ
A (𝑧−1 ) 𝑦 (𝑘) = B (𝑧−1 ) 𝑢 (𝑘 − 1) + C (𝑧−1 ) 𝜐 (𝑘)
Δ (10)
A (𝑧−1 ) Δ𝑦 (𝑘) = B (𝑧−1 ) Δ𝑢 (𝑘 − 1) + C (𝑧−1 ) 𝜐 (𝑘)
where 𝑦(𝑘+𝑗 | 𝑘) is the j-th system optimal output prediction;
𝑁𝑝 is the prediction horizon while 𝑁𝑢 is the control horizon;
𝑤(𝑘 + 𝑗) is the future system set point; Δ𝑢 is the control
action increment; finally, 𝑅 and 𝑄 are weighting matrices
(positive definite) for tracking error and control increments,
respectively.
Considering the white noise case, where 𝐶(𝑧−1 ) is equal
to identity matrices of order 𝑛 × 𝑛 (𝐶(𝑧−1 ) = I), the optimal
prediction for output vector can be generated as follows.
The Diophantine Equation is given as
̃ (𝑧−1 ) + 𝑧−𝑗 F (𝑧−1 )
I = E𝑗 (𝑧−1 ) A (12)
̃ −1 ) = A(𝑧−1 )Δ, E𝑗 (𝑧−1 ) and F𝑗 (𝑧−1 ) are polynomial
where A(𝑧
matrices whose orders are 𝑗 − 1 and 𝑛𝑎 (degree of 𝐴(𝑧−1 )),
respectively.
From this point, the argument (.)(𝑧−1 ) will be omitted for
readability when its use is implicit.
If (10) is multiplied by ΔE𝑗 𝑧𝑗 , then we have
(ΔE𝑗 𝑧𝑗 ) 𝐴𝑦 (𝑘) = (ΔE𝑗 𝑧𝑗 ) B𝑢 (𝑘 − 1)
1
+ (ΔE𝑗 𝑧𝑗 ) I𝜐 (𝑘) (13)
̃ (𝑘 + 𝑗) = E𝑗 𝐵Δ𝑢 (𝑘 + 𝑗 − 1) + E𝑗 𝜐 (𝑘 + 𝑗)
E𝑗 A𝑦
Using (12), we obtain

−1 −1
(I − 𝑧−𝑗 F) 𝑦 (𝑘 + 𝑗) = E𝑗 𝐵Δ𝑢 (𝑘 + 𝑗 − 1)
where A(𝑧 ) and C(𝑧 ) are 𝑛 × 𝑛 monic polynomial
matrices and B(𝑧−1 ) is a 𝑛 × 𝑚 polynomial matrix. The + E𝑗 𝜐 (𝑘 + 𝑗)
backward shift operator, 𝑧−1 , is used to define Δ = 1 − (14)
𝑧−1 . The variables 𝑦(𝑘), 𝑢(𝑘), and 𝜐(𝑘) are the output vector, 𝑦 (𝑘 + 𝑗) = F𝑗 𝑦 (𝑘) + E𝑗 𝐵Δ𝑢 (𝑘 + 𝑗 − 1)
input vector, and noise vector signals, at time instant 𝑘,
respectively. + E𝑗 𝜐 (𝑘 + 𝑗)
Journal of Control Science and Engineering 5

Because E𝑗 degree is 𝑗 − 1, all the noise terms are in future. more precisely as
Considering E[𝜐(𝑘)] = 0, the expected value of 𝑦(𝑘 + 𝑗) is
1 𝑇
𝑦̂ (𝑘 + 𝑗 | 𝑘) = E [𝑦 (𝑡 + 𝑗)] 𝐽= u Hu + b𝑇 𝑢 + 𝑓0 (22)
2
(15)
= F𝑗 𝑦 (𝑘) + E𝑗 BΔ𝑢 (𝑘 + 𝑗 − 1) where

Considering E𝑗 B = G𝑗 + 𝑧−𝑗 G𝑗𝑝 the prediction of 𝑦 can be H = 2 (G𝑇 𝑅G + 𝑄)

written as
b𝑇 = 2 (f − w)𝑇 𝑅G (23)
𝑦̂ (𝑘 + 𝑗 | 𝑘) = G𝑗 Δ𝑢 (𝑘 + 𝑗 − 1) + G𝑗𝑝 Δ𝑢 (𝑘 − 1)
(16) 𝑓0 = (f − w)𝑇 𝑅 (f − w)
+ F𝑗 𝑦 (𝑘)
Note that the last two right-hand side terms depend only If there are no constraints, the optimal solution is given
on past values, which correspond to process free response if by
the control action is maintained constant. Additionally, the −1
first term depends only on the future values and is known as u = −H−1 b = (G𝑇 𝑅G + 𝑄) G𝑇 𝑅 (w − f) (24)
the forced response.
Finally, (16) can be generalized as 4. Constrained GPC Formulation
𝑦̂ (𝑘 + 𝑗 | 𝑘) = G𝑗 Δ𝑢 (𝑘 + 𝑗 − 1) + f𝑗 (17) So far, given formulations do not consider any limit to the
where process. This approach is not so realistic since in practice all
processes are susceptible to some constraint. In this sense,
𝑦̂ (𝑘 + 1 | 𝑘) the crane system described in Section 2 shows boundaries
[ ]
[ 𝑦̂ (𝑘 + 2 | 𝑘) ] in both DC motor voltage and cart path size (see Table 1).
[ ]
[ .. ] Hence, to keep the machine safe, physical limitations must be
[ ]
[ . ] respected. To solve this problem, the GPC controller can use
[ ] optimization algorithms in order to calculate control action
[𝑦̂ (𝑘 + 𝑁 𝑝 | 𝑘) ] that respect imposed constraints.
The control action increment calculated from (24) is
𝐺0 0 ⋅⋅⋅ 0𝑛×𝑚 Δ𝑢 (𝑘) an optimal solution for the optimization problem shown
[ ][
[ 𝐺1
[ 𝐺0 ⋅⋅⋅ 0𝑛×𝑚 ] ][ Δ𝑢 (𝑘 + 1) ] ] in (22). However, if this solution violates some constraints,
[ ] [ ] (18)
=[ . .. ] [ .. ] the control effort would be directly saturated either by the
[ .. . ] [ ]
[ d d ][ . ] control algorithm or by the actuator itself. Thus, this is not the
best control action, since it does not guarantee the optimum
[𝐺𝑁𝑝 −1 𝐺𝑁𝑝 −2 ⋅ ⋅ ⋅ 𝐺𝑁𝑝 −𝑁𝑢 ] [Δ𝑢 (𝑘 + 𝑁𝑢 − 1)]
solution for the cost function. This situation can be solved
when constraints are taken into account.
f1
[ ] According to [28], process constraints usually impose
[ f2 ] amplitude limits in the control actions (in crane case this will
[ ]
+[ ]
[ .. ] be the voltage applied to DC motor and its variation) and in
[ . ] the outputs (maximum crane trolley track length and payload
[ ]
oscillation). This can be described by
[f𝑁𝑝 ]
𝑗−1 Δ𝑢𝑚𝑖𝑛 ≤ 𝑢 (𝑘) ≤ Δ𝑢𝑚𝑎𝑥 ∀𝑘
with f𝑗 = G𝑗𝑝 Δ𝑢(𝑘 − 1) + F𝑗 𝑦(𝑘) and G𝑗 (𝑧−1 ) = ∑𝑖=𝑗−𝑁𝑢 𝐺𝑖 𝑧−𝑖
and 𝐺𝑖 = 0𝑛×𝑚 for 𝑖 < 0. 𝑢𝑚𝑖𝑛 ≤ 𝑢 (𝑘) − 𝑢 (𝑘 − 1) ≤ 𝑢𝑚𝑎𝑥 ∀𝑘 (25)
The predictions can be expressed in a simplified form as
𝑦𝑚𝑖𝑛 ≤ 𝑦 (𝑘) ≤ 𝑦𝑚𝑎𝑥 ∀𝑘
y = Gu + f (19)
Considering a process with 𝑚 inputs and 𝑛 outputs,
Therefore, (11) can be rewritten as subject to a prediction horizon 𝑁𝑝 and a control horizon 𝑁𝑢 ,
(25) can be expressed by
𝐽 = (Gu + f − w)𝑇 𝑅 (Gu + f − w) + u𝑇 𝑄u (20)

where 𝑅 and 𝑄 are diagonal matrices with 𝑅 and 𝑄 as entries
of main diagonal, respectively.
Equation (11) can assume a quadratic form, as follows:
𝐽 = u𝑇 (G𝑇 𝑅G + 𝑄) u + (f − w)𝑇 𝑅Gu
(21)
+ (f − w)𝑇 𝑅 (f − w)
1𝑢 Δ𝑢𝑚𝑖𝑛 ≤ u ≤ 1𝑢 Δ𝑢𝑚𝑎𝑥
1𝑢 𝑢𝑚𝑖𝑛 ≤ 𝑇𝑟𝑖𝑎𝑛𝑔 u + 𝑢 (𝑘 − 1) 1𝑢 ≤ 1𝑢 𝑢𝑚𝑎𝑥 (26)
1𝑦 𝑦𝑚𝑖𝑛 ≤ Gu + f ≤ 1𝑦 𝑦𝑚𝑎𝑥
where 1𝑢 is a matrix of dimension (𝑚𝑁𝑢 × 𝑚) formed by
𝑁𝑢 (𝑚 × 𝑚) identity matrices; 1𝑦 is a (𝑛𝑁𝑝 × 𝑛) matrix formed
6 Journal of Control Science and Engineering

by 𝑁𝑝 (𝑛 × 𝑛) identity matrices. Finally, T𝑟𝑖𝑎𝑛𝑔 is a 𝑁𝑢 × 𝑁𝑢 1𝑁

𝑢 Δ𝑢𝑚𝑎𝑥
𝑢𝑐

[ ]
lower triangular block matrix whose nonnull block entries are [ −1𝑁 ]
𝑢 Δ𝑢𝑚𝑖𝑛
𝑢𝑐
𝑚 × 𝑚 identity matrices. These constraints can be expressed [ ]
[ 𝑁 ]
[ 1 𝑢𝑐 𝑢 𝑁𝑢𝑐 ]
in a simplified form as [ 𝑢 𝑚𝑎𝑥 − 1𝑢 𝑢 (𝑘 − 1) ]
c𝑁𝑐 [
= [ 𝑁𝑢𝑐 ]
𝑁𝑢𝑐 ]
[−1𝑢 𝑢𝑚𝑖𝑛 + 1𝑢 𝑢 (𝑘 − 1)]
R𝑐 u ≤ c [ ]
(27) [ 𝑁 ]
[ 1𝑦 𝑦𝑐 𝑦𝑚𝑎𝑥 − f 𝑁𝑦𝑐 ]
[ ]
𝑁𝑦𝑐 𝑁𝑦𝑐
where [ 1 𝑦 𝑦𝑚𝑖𝑛 + f ]
(30)
I𝑚𝑁𝑢 ×𝑚𝑁𝑢 𝑁
𝑁
[ ]
[−I𝑚𝑁𝑢 ×𝑚𝑁𝑢 ] where the terms 1𝑁 𝑢 , 𝑇𝑟𝑖𝑎𝑛𝑔 , 1𝑦 , and 𝐺
𝑢𝑐 𝑢𝑐 𝑦𝑐 𝑁𝑦𝑐
are truncated
[ ] forms of 1𝑢 , 𝑇𝑟𝑖𝑎𝑛𝑔 , 1𝑦 , and 𝐺, respectively.
[ ]
[ T𝑟𝑖𝑎𝑛𝑔 ] The objective function 𝐽 is then rewritten as
Rc = [
[ −T
]
]
[ 𝑟𝑖𝑎𝑛𝑔 ] 1
[ ] min 𝐽 = u𝑇 Hu + b𝑇 u + 𝑓0
[ ] u 2
[ G ] (31)
[ G ] s.t.: R𝑁𝑐 u ≤ c𝑁𝑐
(28)
1𝑢 Δ𝑢𝑚𝑎𝑥 By solving this programming problem, it is possible to
[ ] keep the system under constraints ensuring optimal perfor-
[ −1𝑢 Δ𝑢𝑚𝑖𝑛 ]
[ ] mance.
[ ]
[ 1𝑢 𝑢𝑚𝑎𝑥 − 1𝑢 𝑢 (𝑘 − 1) ]
c=[
[−1 𝑢 + 1 𝑢 (𝑘 − 1)]
]
[ 𝑢 𝑚𝑖𝑛 𝑢 ] 5. Explicit GPC
[ ]
[ 1𝑦 𝑦𝑚𝑎𝑥 − f ]
[ ] In Sections 3 and 4, the GPC strategies with and without
[ 1 𝑦
𝑦 𝑚𝑖𝑛 + f ] constraints were introduced. The main problem with the
Unconstrained GPC (UGPC) is that it does not calculate
the optimum control action taking constraints into account,
Therefore, the minimization of objective function 𝐽 can which can lead to loss of performance or limits violation.
be formulated as a QP problem in the following form: On the other hand, the Constrained GPC (CGPC), despite
regarding constraints, computes the optimal solution at each
1 𝑇 iteration of online control. This means that at each time step a
min 𝐽= u Hu + b𝑇 u + 𝑓0 large number of calculations must be performed, which imply
u 2 (29)
in high computational cost.
s.t.: Rc u ≤ c One solution to reduce the online QP computational cost
is to calculate an explicit control law using multiparametric
There are several optimization techniques to solve this prob- programming (mp). Controller designs have been developed
lem. In this case, the QP must be solved at each iteration with mp, focusing on linear systems control subject to
of control algorithm. In other words, at each time step, the constraints [38]. The explicit/multiparametric MPC [39] is
control action is calculated and sent to the process. When one of them, and it has been applied in several practical
applied to a MIMO process, the amount of memory and applications [40, 41], including crane systems [30]. Many of
CPU time required to solve the optimization problem may these are industrial applications where the available hardware
grow, according to chosen prediction and control horizons. and control software are limited to Advanced Process Control
To reduce this, the simplified form presented by (28) can be (APC) [42].
expressed considering a Constraint Horizon [37] for input The multiparametric programming is capable of subdi-
constraints (𝑁𝑢𝑐 ≤ 𝑁𝑢 ) and another for output constraints viding the parameter space into characteristic regions (CR),
(𝑁𝑦𝑐 ≤ 𝑁𝑝 ). where the optimal value and the optimizer are expressed
as explicit functions of the parameters [43]. This offline
controller design approach produces an optimal solution,
I𝑚𝑁𝑢𝑐 ×𝑚𝑁𝑢 which is defined as a Piecewise Affine (PWA) control law
[ ] over polyhedral regions. Therefore, this strategy can find an
[−I𝑚𝑁𝑢𝑐 ×𝑚𝑁𝑢 ]
[ ] explicit control law for the Constrained GPC, similar to the
[ ]
[ T𝑁 𝑢𝑐 ] Unconstrained GPC control law, as follows
[ 𝑟𝑖𝑎𝑛𝑔 ]
R𝑁𝑐 =[ ]
[ −T𝑁𝑢𝑐 ] (1) The convex polyhedral set 𝐶𝑃 is partitioned into 𝑁𝐶𝑃
[ 𝑟𝑖𝑎𝑛𝑔 ]
[ ] polyhedral regions:
[ G𝑁𝑦𝑐 ]
[ ]
𝑁𝑦𝑐 𝐶𝑃𝑖 = {𝑥 ∈ R𝑝 | 𝑃𝑖 𝑥 ≤ 𝑑𝑖 } 𝑖 = 1, ⋅ ⋅ ⋅ , 𝑁𝐶𝑃 (32)
[ G ]
Journal of Control Science and Engineering 7

(2) The optimal solution 𝑢∗ (𝑥(𝑘)) : 𝐶𝑃 󳨀→ R𝑞 is an PWA

function in the form Input: Model and Constraints parameters.
Output: 𝑆𝑗 , 𝑞𝑗 , 𝐶𝑃 (Characteristics Regions).
𝑢∗ (𝑥 (𝑘)) = 𝑆𝑖 𝑥 (𝑘) + 𝑞𝑖 𝑖𝑓 𝑥 (𝑘) ∈ 𝐶𝑃𝑖 (33) (1) Assign parameters (see equation (34))
(2) Solve the mp-QP of equation (34)
The QP problem shown in (11) can be rewritten as a
multiparametric quadratic programming problem (mp-QP). Algorithm 1: GPC design.
Thus, consider the following mp-QP formulation available in
the MPT toolbox for MATLAB [44]:
1 𝑇 The presented algorithm summarizes the steps required
min u Hu + (x𝑇 F + 𝐶𝑓 ) u + x𝑇 Yx + 𝐶𝑥 x + 𝐶𝑐 to obtain an explicit law. This law is obtained by including
u 2 (34) the necessary variables and solving the problem. As output, it
s.t.: R𝑁𝑐
𝑐 u ≤ W + Ex returns both the explicit law parameters (𝑆𝑗 and 𝑞𝑗 ) and 𝐶𝑃
as space parameter divided into characteristic regions 𝐶𝑃𝑗 .
The mp-QP parameters expressed in (34) are defined based After designing the Explicit GPC, the next part is to
on GPC formulation presented in Sections 3 and 4, as follows: control the constrained system with the generated explicit
law. Algorithm 2 details the required sequence to perform the
w−f online control.
[ 𝑁 ] As seen, the loop is described in three main steps. At first,
[
x = [1𝑢 𝑢 (𝑘 − 1)]
𝑢𝑐
], it is necessary to calculate the parameters f 𝑁𝑦𝑐 and w − f. Note
𝑁𝑦𝑐
[ f ] that at first iteration 𝑢(𝑘−1) = 0. These parameters determine
the vector 𝑥. In sequence, the obtained values of 𝑥 are used
−2G𝑅 to find the associated polyhedral region. Finally, (33) is used
F=[ ], to calculate the control action for the current time instant.
0(𝑛𝑁𝑦𝑐+𝑚𝑁𝑢𝑐 )×𝑚𝑁𝑢
The loop continues until the maximum number of iterations
Δ𝑢𝑚𝑎𝑥 is reached or the execution of the algorithm is forced to finish.
[ ]
[ Δ𝑢𝑚𝑖𝑛 ]
[ ] 6. Numerical Example
[ ]
[ 𝑈𝑚𝑎𝑥 ]
[
W=[ ], (35)
] In this section, the linearized model presented in Section 2
[ 𝑈𝑚𝑖𝑛 ]
[ ] is controlled, assuming input and output constraints to illus-
[𝑌 ]
[ 𝑚𝑎𝑥 ] trate the proposed approach discussed in previous sections.
[ 𝑌𝑚𝑖𝑛 ] Three situations are explored. In the first one, the crane
system is controlled by an Unconstrained GPC (UGPC), in
02𝑚𝑁𝑢𝑐 ×[𝑛(𝑁𝑝 +𝑁𝑦𝑐 )+𝑚𝑁𝑢𝑐 )] the second, the system is controlled by a Constrained GPC
[ ] (CGPC), using QP subject to constraints, and, in the third,
[ 0𝑚𝑁𝑢𝑐 ×𝑛𝑁𝑝 −I𝑚𝑁𝑢𝑐 ×𝑚𝑁𝑢𝑐 0𝑚𝑁𝑢𝑐 ×𝑛𝑁𝑦𝑐 ]
[ ] the system is controlled by the Explicit GPC (EGPC). For
[ ]
E=[ 0 I 0
[ 𝑚𝑁𝑢𝑐 ×𝑛𝑁𝑝 𝑚𝑁𝑢𝑐 ×𝑚𝑁𝑢𝑐 𝑚𝑁𝑢𝑐 ×𝑛𝑁𝑦𝑐
]
] comparison purposes, the same parameters are used in all
[ ] situations.
[ 0𝑛𝑁𝑝𝑐 ×𝑛𝑁𝑝 0𝑛𝑁𝑦𝑐×𝑚𝑁𝑢𝑐 −I𝑛𝑁𝑦𝑐 ×𝑛𝑁𝑦𝑐 ]
[ ] The first step to control the crane system is to obtain the
[ 0𝑛𝑁𝑝𝑐 ×𝑛𝑁𝑝 0𝑛𝑁𝑦𝑐 ×𝑚𝑁𝑢𝑐 I𝑛𝑁𝑦𝑐 ×𝑛𝑁𝑦𝑐 ] discretized model and, then, rewrite it in CARIMA form. By
using the parameters shown in Table 1, it is possible to obtain
with
𝑋 (𝑧)
u = [Δu (𝑘) Δu (𝑘 + 1) . . . Δu (𝑘 + 𝑁𝑢 − 1)] 𝑉 (𝑧)
(37)
Y = 0[𝑛(𝑁𝑝 +𝑁𝑦𝑐 )+𝑚𝑁𝑢𝑐 ]×[𝑛(𝑁𝑝 +𝑁𝑦𝑐 )+𝑚𝑁𝑢𝑐 ] 4.9630 × 10−5 (𝑧3 − 0.8388𝑧2 − 0.8387𝑧 + 1)
=
𝐶𝑓 = 0 (36) 𝑧4 − 3.6432𝑧3 + 5.2863𝑧2 − 3.6430𝑧 + 1

𝐶𝑥 = 0 After system discretization, GPC is designed to obtain

𝐶𝑐 = 0
and H and RN c
c as the same of(23) and (30), respectively.
Once the programming problem is settled, it is necessary
to design the Explicit GPC control and use it to control the
crane system. To better exemplify the controller design step
see Algorithm 1.
(23) and (24). In sequence, tuning parameters must be found
in order to maximize the performance of the controller
when applied to the GCS. In this case, the parameters of
the Table 2 were chosen for all three study cases previously
listed. It should be mentioned that the MPC parameters were
empirically adjusted in the numerical simulations. It was also
considered a maximum tolerance of ±0.0029rad for payload
swing and the track length of 0.8m.
8 Journal of Control Science and Engineering

Input: 𝑆𝑗 , 𝑞𝑗 , 𝐶𝑃 (Caracteristics Regions).

Output: Control action.
(1) Calculate 𝑥 values from equation (34)
(2) Verify to which polyhedral region 𝑥 belongs
(3) Calculate the control action from equation (33) associated to the 𝑗 − 𝑡ℎ polyhedral region

Algorithm 2: Online control loop.

Table 2: GPC parameters. Crane Control - Unconstrained GPC

0.01
Parameters Value 0.008
0.006
𝑁𝑝 45

Oscillation (rad)
Prediction horizon
0.004
𝑁𝑢 1 Control horizon 0.002
𝑁𝑦𝑐 13 Constraint output horizon 0
𝑁𝑢𝑐 1 Constraint input horizon −0.002
−0.004
𝑇𝑠 0.1s Sample time
−0.006
0.6 0 Weighting coefficient for −0.008
𝑅 [ ]
0 150 tracking error −0.01
[ ] 0 10 20 30 40 50 60 70 80
𝑄 0.0005𝐼𝑚×𝑚 Weighting coefficient for control Time (s)
increments
Δ𝑢𝑚𝑎𝑥 +0.7V Figure 3: Payload oscillation.
Input increment upper limit
Δ𝑢𝑚𝑖𝑛 −0.7V Input increment lower limit
𝑈𝑚𝑎𝑥 +5V Input upper limit
to calculate the Goodhart index 𝐼𝑔ℎ [45], which is determined
𝑈𝑚𝑖𝑛 −5V Input lower limit as follows:
+0.8m
𝑌𝑚𝑎𝑥 [ ] Output upper limit 𝐼𝑔ℎ = 𝛼1 𝜀1 + 𝛼2 𝜀2 + 𝛼3 𝜀3 (38)
[+0.0029rad]
0m where
𝑌𝑚𝑖𝑛 [ ] Output lower limit
[−0.0029rad] ∑ 𝑢 (𝑘)
𝜀1 =
𝑁
2
∑ [𝑢 (𝑘) − 𝜀1 ]
In all cases, the main objective is to achieve zero payload 𝜀2 = (39)
𝑁
swing while putting the trolley at the desired position. During
󵄨󵄨 󵄨
the simulation, the set point for the trolley position is 0.77m. ∑ 󵄨𝑤 (𝑘) − 𝑦 (𝑘)󵄨󵄨󵄨
The cable length, cart mass, and payload mass are those 𝜀3 = 󵄨
𝑁
shown in Table 1.
The performance of the controllers is analyzed based with 𝑁 being the number of samples. The parameters 𝛼1 ,
on minimization of payload oscillation and transient and 𝛼2 , and 𝛼3 will weight what is more relevant to the control
residual swing angle when the cart is in movement. The Mean performance: spent control energy, variability imposed to
Square Error (MSE) can be used as a performance index to actuator, and tracking error, respectively. The chosen value for
both payload oscillation and trolley position, where a lower this parameters were 𝛼1 = 0.2, 𝛼2 = 0.3, and 𝛼3 = 0.5.
value of MSE represents a higher swing suppression and In order to summarize the comparison between con-
correct positioning for the cart (according to its set point), trollers Table 3 shows the performances according to error
respectively. measures, maximum transient swing, residual swing, Good-
Together with payload oscillation, maximum transient hart index, settling time, and overshoot.
swing (MT) and residual swing (RS) are also analyzed in
which it is desirable to minimize both MT and RS. It is 6.1. Unconstrained GPC Control. The first examined case is
important to define RS as a payload sway that remains after the application of the Unconstrained GPC to the Gantry
the trolley has reached the steady-state value. Crane System expressed by (8). Figures 3 and 4 illustrate the
Additionally, besides MSE, the control of cart position it payload swing and the trolley position.
is also evaluated considering step responses characteristics, As observed in Figure 4, the trolley position is controlled
as overshoot and settling time. Furthermore, in order to showing an overshoot of 5.89%. However, the UGPC is not
obtain information about the energy consumption during able to keep the cart inside the track limits, since the position
crane operation, the control effort and its variability are used is higher than 0.8m. The same happens regarding payload
Journal of Control Science and Engineering 9

Table 3: Controllers performance.

Performance Index UGPC CGPC EGPC

Oscillation MSE 9.9898 × 10−6 6.7987 × 10−6 6.7987 × 10−6
Oscillation Goodhart 8.6761 9.3991 9.3991
MT 0.0046 0.0025 0.0025
RS 1.0561 × 10−6 1.3208 × 10−7 1.3208 × 10−7
Position MSE 0.3385 0.3698 0.3698
Position Goodhart 8.9758 9.7103 9.7103
Settling time(s) 18 14.4 14.4
Overshoot (%) 5.8855 3.8961 3.8961

Table 4: Design time and overall time simulation. state. About the trolley position, the controller has a lower
overshoot reaching 3.8961%. When one considers the control
Procedure UGPC CGPC EGPC effort, Figures 7(a) and 7(b) show that CGPC keeps the input
Design 89.778𝜇s - 36min signal and input increment inside boundaries.
Overall runtime 0.0291s 0.9901s 2.3360s One limitation found in online optimization is the time
spent to calculate the control action. The overall runtime
simulation was higher than the other controllers, reaching
Crane Control - Unconstrained GPC approximately 2.3360𝑠. If the computation takes longer than
0.9 the sampling time, this may lead the system to instability. For
0.8 this numerical examples, this is the case when 𝑁𝑝 is higher
0.7 and 𝑁𝑦𝑐 = 𝑁𝑝 .
Position (m)

0.6
0.5
0.4
0.3
0.2
0.1
0 handle process constraints while the CGPC may have issues
0 10 20 30 40 50
Time (s)
Figure 4: Trolley displacement.
swing. The ±0.0029rad tolerance is violated, with swing
reaching a maximum value of 0.0046rad during transient
state. The residual swing is nearly zero. The control action
and its increment are presented in Figures 5(a) and 5(b),
respectively. The input constraints were only respected due
to a saturation added in control routine.
6.2. Constrained GPC Control. The Constrained GPC used
the same parameters presented in Table 2. This type of
control, at each iteration, solves an optimization problem
and returns the optimal control action. The advantage of this
technique is that even though the system changes, the control
action calculation can be correctly performed, unless the
problem is infeasible. The CGPC results are shown in Figures
6(a), 6(b), 7(a), and 7(b). As seen from Figures 6(a) and 6(b),
the controller is able to keep both payload swing and trolley
position inside their respective limits. The maximum swing
happens during the transient reaching a value of 0.0025rad
and nearly no RS was left. Comparing with the previous case,
the CGPC has a better performance, mainly during transient
The overall runtime spent during the simulations for all
controllers is shown on Table 4.
6.3. Explicit GPC Control. In this last numerical simulation,
the results obtained by the proposed controller are presented.
It is important to note that the UGPC showed problems to
60 70 80
regarding the time spent at online execution. The EGPC
was able to deal with both situations. Moreover, the tuning
parameters of Table 2 are also used from this approach. With
this tuning configuration, the mp-QP was solved after 34
minutes, generating a total of 57 polyhedral regions. The
greater 𝑁𝑝 , 𝑁𝑢 , 𝑁𝑦𝑐 , and 𝑁𝑢𝑐 are, the greater will be the
number of regions and, consequently, the memory storage
and CPU time needed to solve mp-QP.
The results of proposed EGPC are shown in Figures 8,
9, 10(a), and 10(b). Table 3 shows that EGPC has the same
performance as CGPC, respecting all imposed constraints.
Additionally, note that the overall runtime spent during
simulation to compute all control actions was 1.0274s. This
is around 2.3594 times faster than CGPC and 34.0241 times
slower than UGPC.
The main drawback observed for EGPC was the design
runtime (time spent during the resolution of the mp-QP
problem). That took about 36min to finish. However, this
procedure is an offline operation and it is one time only
execution, whenever the parameters change. After the mp-
QP solution is found, the time spent is lower than online
solutions.
The three studied cases were also evaluated by other
index as settling time MSE, Goodhart. The settling time is
directly related to the DC motor applied voltage. Its amplitude
regulates how fast the trolley can move to reach the set point.
10 Journal of Control Science and Engineering

Crane Control - Unconstrained GPC Crane Control - Unconstrained GPC

5 30

Input increment (V)

Input Voltage (V)

20

0 10

−5 −10
0 20 40 60 80 0 20 40 60 80
Time (s) Time (s)
(a) Control effort (b) Control increment

Figure 5: Control action behaviour for Unconstrained GPC.

Crane Control - Constrained GPC Crane Control - Constrained GPC

0.01 0.8
0.008 0.7
0.006
0.6
Oscillation (rad)

0.004 Position (m)

0.002 0.5
0 0.4
−0.002 0.3
−0.004
0.2
−0.006
−0.008 0.1
−0.01 0
0 10 20 30 40 50 60 70 80 0 10 20 30 40 50 60 70 80
Time (s) Time (s)
(a) Payload oscillation (b) Trolley displacement

Figure 6: System outputs for Constrained GPC.

Crane Control - Constrained GPC Crane Control - Constrained GPC

5 0.5
Input increment (V)
Input Voltage (V)

0 0

−5 −0.5
0 20 40 60 80 0 20 40 60 80
Time (s) Time (s)
(a) Control effort (b) Control increment

Figure 7: Control action behaviour for Constrained GPC.

Since the maximum input voltage is limited to ±5 Volts, it state but provokes a higher overshoot, and, in this case, a
is difficult to obtain a faster control. As shown in Table 3, constraint violation. The UGPC had also lower values for
the CGPC and EGPC controllers took a bit more than 14s to both Goodhart indexes. The control increment variability
reach steady state, while UGPC was slower than the others, explains this behaviour, since UGPC has a lesser oscillatory
taking 18s. On the other hand, the UGPC achieve a lower signal. Regarding payload’s swing error, the CGPC and EGPC
value for position MSE. The main reason for this is his fast had the better performance, reducing almost 1.47 times the
rising time, which reduces error quickly during transient oscillation when compared to the unconstrained case.
Journal of Control Science and Engineering 11

Crane Control - Explicit GPC

0.01
0.008

Oscillation (rad)
0.006
0.004
0.002
0
−0.002
−0.004
−0.006
−0.008
−0.01
0 10 20 30 40 50 60 70 80
Time (s)

Figure 8: Payload oscillation.

Crane Control - Explicit GPC

0.9
0.8
0.7
0.6
Position (m)

0.5
0.4
0.3
0.2
0.1
0
0 10 20 30 40 50 60 70 80
Time (s)

Figure 9: Trolley displacement.

7. Conclusion the proposed formulation, whenever this tuning parameters

In this work, a GPC based on transfer function model was
adapted to calculate an explicit PWA controller that guaran-
tees the same performance as the constrained controller but
with low computational cost.
Results showed that the proposed controller was able to
control the presented Gantry Crane System reducing payload
swing, while positioning the trolley at desired position. When
compared with the Unconstrained GPC, the EGPC was able
to respect both input and output constraints which derive
in better overall performance. When EGPC is compared
to CGPC, Table 3 evaluation indexes show same results;
however online EGPC calculation needs lower computational
effort during simulation.
In comparison with Quadratic Programming, multipara-
metric Quadratic Programming storage need is higher. Also,
online programming is more flexible to parameter changes.
In other words, when using an identification algorithm the
control action calculation that uses Quadratic Programming
may be less expensive than the calculation by multiparamet-
ric programming, since the design step has higher computa-
tional cost.
Future work shall do a further investigation onto multi-
parametric formulation so the parameter 𝑥 becomes inde-
pendent of the prediction and control horizon size. In
increase, the vector 𝑥 also grows, enlarging the amount of
necessary memory and time to solve the mp-QP. Moreover,
future works also shall address the control law adaptation
with a recursive multiparametric programming algorithm,
which would improve the algorithm performance. Finally, the
proposed GPC controller needs to be better investigated and
applied in different scenarios and problems.
Data Availability
The data used to support the findings of this study are
available from the corresponding author upon request.
Conflicts of Interest
The authors declare that they have no conflicts of interest.
Acknowledgments
The authors would like to mention the institutions that con-
tributed to the development of this project: UFRN (Federal
University of Rio Grande do Norte), DCA (Department of
Computation and Automation), UnP (Potiguar University),
12
Crane Control - Explicit GPC
5 0.5
Input Voltage (V)
0 0
−5
0 20 40 60
Time (s)
(a) Control effort
Figure 10: Control action behaviour for Explicit GPC.
IFRN (Federal Institute of Education, Science and Technol-
ogy of Rio Grande do Norte), Capes (National Council for the
Improvement of Higher Education), and LAUT (Laboratory
of Automation in Petroleum).
References
[1] N. Sun, Y. Fang, H. Chen, B. Lu, and Y. Fu, “Slew/translation
positioning and swing suppression for 4-DOF tower cranes with
parametric uncertainties: design and hardware experimenta-
tion,” IEEE Transactions on Industrial Electronics, vol. 630, no.
10, pp. 6407–6418, 2016.
[2] D. Kim and Y. Park, “Tracking control in xy plane of an offshore
container crane,” Journal of Vibration and Control, vol. 230, no.
3, pp. 469–483, 2017.
[3] N. Sun, Y. Wu, Y. Fang, and H. Chen, “Nonlinear antiswing
control for crane systems with double-pendulum swing effects
and uncertain parameters: design and experiments,” IEEE
Transactions on Automation Science and Engineering, vol. 150,
no. 3, pp. 1413–1422, 2018.
[4] E. M. Abdel-Rahman, A. H. Nayfeh, and Z. N. Masoud,
“Dynamics and control of cranes: a review,” Journal of Vibration
and Control, vol. 90, no. 7, pp. 863–908, 2003.
[5] S. C. Duong, E. Uezato, H. Kinjo, and T. Yamamoto, “A hybrid
evolutionary algorithm for recurrent neural network control of
a three-dimensional tower crane,” Automation in Construction,
vol. 23, pp. 55–63, 2012.
[6] G. Rigatos, P. Siano, and M. Abbaszadeh, “Nonlinear H-
infinity control for 4-DOF underactuated overhead cranes,”
Transactions of the Institute of Measurement and Control, vol.
40, no. 7, pp. 2364–2377, 2017.
[7] N. Sun, Y. Fang, and X. Zhang, “Energy coupling output
feedback control of 4-DOF underactuated cranes with saturated
inputs,” Automatica, vol. 49, no. 5, pp. 1318–1325, 2013.
[8] M. R. Ghazali, M. A. Ahmad, M. F. Jusof, and R. M. Ismail, “A
data-driven neuroendocrine-PID controller for underactuated
systems based on safe experimentation dynamics,” in Proceed-
ings of the 2018 IEEE 14th International Colloquium on Signal
Processing & Its Applications (CSPA), pp. 61–66, Batu Feringghi,
Malaysia, March 2018.
[9] H. I. Jaafar, Z. Mohamed, J. Jamian, A. F. Abidin, A. M. Kassim,
and Z. A. Ghani, “Dynamic behaviour of a nonlinear gantry
crane system,” Procedia Technology, vol. 11, pp. 419–425, 2013.
Journal of Control Science and Engineering
Crane Control - Explicit GPC
Input increment (V)
−0.5
80 0 20 40 60 80
Time (s)
(b) Control increment
[10] V. S. Renuka and A. T. Mathew, “Precise modelling of a
gantry crane system including friction, 3d angular swing and
hoisting cable flexibility,” International Journal on Theoretical
and Applied Research in Mechanical Engineering (IJTARME),
vol. 2, pp. 119–125, 2013.
[11] L. Ramli, Z. Mohamed, A. M. Abdullahi, H. Jaafar, and I. M.
Lazim, “Control strategies for crane systems: A comprehensive
review,” Mechanical Systems and Signal Processing, vol. 95, pp.
1–23, 2017.
[12] N. D. Zrnić, V. M. Gašić, and S. M. Bošnjak, “Dynamic
responses of a gantry crane system due to a moving body con-
sidered as moving oscillator,” Archives of Civil and Mechanical
Engineering, vol. 15, no. 1, pp. 243–250, 2015.
[13] R. Piedrafita, D. Comı́n, and J. R. Beltrán, “Simulink imple-
mentation and industrial test of Input Shaping techniques,”
Control Engineering Practice, vol. 79, pp. 1–21, 2018.
[14] D. Fujioka and W. Singhose, “Input-shaped model reference
control of a nonlinear time-varying double-pendulum crane,”
in Proceedings of the 10th Asian Control Conference, ASCC 2015,
pp. 1–6, Malaysia, June 2015.
[15] M. A. Ahmad, R. M. T. R. Ismail, and M. S. Ramli, “Input
shaping techniques for anti-sway control of a 3-D gantry
crane system,” in Proceedings of the 2009 IEEE International
Conference on Mechatronics and Automation, ICMA 2009, pp.
2876–2881, China, August 2009.
[16] L. Ramli, Z. Mohamed, and H. Jaafar, “A neural network-based
input shaping for swing suppression of an overhead crane under
payload hoisting and mass variations,” Mechanical Systems and
Signal Processing, vol. 107, pp. 484–501, 2018.
[17] H. Jaafar, Z. Mohamed, M. Shamsudin, N. Mohd Subha, L.
Ramli, and A. Abdullahi, “Model reference command shap-
ing for vibration control of multimode flexible systems with
application to a double-pendulum overhead crane,” Mechanical
Systems and Signal Processing, vol. 115, pp. 677–695, 2019.
[18] N. Sun, Y. Fang, and H. Chen, “Adaptive antiswing control
for cranes in the presence of rail length constraints and
uncertainties,” Nonlinear Dynamics, vol. 810, no. 1-2, pp. 41–51,
2015.
[19] C. S. Teo, K. K. Tan, S. Y. Lim, S. Huang, and E. B. Tay,
“Dynamic modeling and adaptive control of a H-type gantry
stage,” Mechatronics, vol. 170, no. 7, pp. 361–367, 2007.
[20] Z. Mohamed, A. K. Chee, A. W. I. Mohd Hashim, M. O. Tokhi, S.
H. M. Amin, and R. Mamat, “Techniques for vibration control
of a flexible robot manipulator,” Robotica, vol. 240, no. 4, pp.
499–511, 2006.
Journal of Control Science and Engineering 13

[21] A. Alhassan, K. A. Danapalasingam, M. Shehu, A. M. Abdullahi, [38] T. A. Johansen, W. Jackson, R. Schreiber, and P. Tondel,
and A. Shehu, “Comparing the Performance of Sway Control “Hardware synthesis of explicit model predictive controllers,”
Using ZV Input Shaper and LQR on Gantry Cranes,” in Control Systems Technology, vol. 15, pp. 191–197, 2006.
Proceedings of the 2015 9th Asia Modelling Symposium (AMS), [39] K. I. Kouramas, N. P. Faı́sca, C. Panos, and E. N. Pistikopoulos,
pp. 61–66, Kuala Lumpur, Malaysia, September 2015. “Explicit/multi-parametric model predictive control (MPC) of
[22] M. I. Solihin, Wahyudi, and A. Legowo, “Fuzzy-tuned PID anti- linear discrete-time systems by dynamic and multi-parametric
swing control of automatic gantry crane,” Journal of Vibration programming,” Automatica, vol. 470, no. 8, pp. 1638–1645, 2011.
and Control, vol. 160, no. 1, pp. 127–145, 2010. [40] B. Kouvaritakis and M. Cannon, Non-Linear Predictive Control:
[23] S. Olaru, D. Dumur, and S. Tebbani, “Parameterized polyhedra Theory and Practice, IET, 2001.
approach for robust constrained generalized predictive control,” [41] M. E. Emekli and B. Aksun Guvenc, “Explicit MIMO model
in Proceedings of the 2005 IEEE Conference on Control Applica- predictive boost pressure control of a two-stage turbocharged
tions, (CCA 2005), pp. 428–433, Toronto, Canada, 2005. diesel engine,” IEEE Transactions on Control Systems Technol-
[24] K. Witheephanich, J. M. Escano, and M. J. Hayes, “Explicitly ogy, vol. 250, no. 2, pp. 521–534, 2017.
constrained generalised predictive control strategies for power [42] E. N. Pistikopoulos, L. Dominguez, C. Panos, K. Kouramas,
management in ambulatory wireless sensor network systems,” and A. Chinchuluun, “Theoretical and algorithmic advances
in Proceedings of the 2010 American Control Conference (ACC in multi-parametric programming and control,” Computational
2010), pp. 1856–1861, Baltimore, MD, USA, June 2010. Management Science, vol. 9, no. 2, pp. 183–203, 2012.
[25] A. Gholabi, M. Ebrahimi, G. R. Yousefi, M. Ghayour, A.
[43] A. Bemporad, F. Borrelli, and M. Morari, “Model predictive
Ebrahimi, and H. Jali, “Sensorless anti-swing control for over-
control based on linear programming - the explicit solution,”
head crane using voltage and current measurements,” Journal of
Automatic Control, vol. 47, no. 12, pp. 1974–1985, 2002.
Vibration and Control, vol. 210, no. 9, pp. 1745–1756, 2015.
[44] M. Herceg, M. Kvasnica, C. N. Jones, and M. Morari, “Multi-
[26] F. Alasali, S. Haben, V. Becerra, and W. Holderbaum, “Optimal
Parametric Toolbox 3.0,” in Proceedings of the 2013 European
energy management and MPC strategies for electrified RTG
Control Conference (ECC), pp. 502–510, Zurich, Switzerland,
cranes with energy storage systems,” Energies, vol. 10, no. 10,
July 2013, http://control.ee.ethz.ch/mpt.
2017.
[45] S. G. Goodhart, K. J. Burnham, and D. J. James, “Bilinear self-
[27] H. Chen, Y. Fang, N. Sun et al., “A swing constraint guar-
tuning control of a high temperature heat treatment plant,”
anteed MPC algorithm for underactuated overhead cranes,”
Control Theory and Applications, IEE Proceedings, vol. 14, no. 10,
IEEE/ASME Transactions on Mechatronics, vol. 210, no. 5, pp.
1994.
2543–2555, 2016.
[28] E. F. Camacho and C. B. Alba, Model Predictive Control, Springer
Science and Business Media, 2013.
[29] M. Böck and A. Kugi, “Real-time nonlinear model predictive
path-following control of a laboratory tower crane,” IEEE
Transactions on Control Systems Technology, vol. 220, no. 4, pp.
1461–1473, 2014.
[30] G. Galuppini, L. Magni, and D. M. Raimondo, “Model predic-
tive control of systems with deadzone and saturation,” Control
Engineering Practice, vol. 78, pp. 56–64, 2018.
[31] K. Oh, J. Seo, J.-G. Kim, and K. Yi, “MPC-based approach to
optimized steering for minimum turning radius and efficient
steering of multi-axle crane,” International Journal of Control,
Automation, and Systems, vol. 15, no. 4, pp. 1799–1813, 2017.
[32] A. Singh and H. Agrawal, “A fractional model predictive control
design for 2-d gantry crane system,” Journal of Engineering
Science and Technology, vol. 13, pp. 2224–2235, 2018.
[33] Š. Ileš, J. Matuško, and F. Kolonić, “Sequential distributed
predictive control of a 3D tower crane,” Control Engineering
Practice, vol. 79, pp. 22–35, 2018.
[34] J. Smoczek and J. Szpytko, “Particle swarm optimization-based
multivariable generalized predictive control for an overhead
crane,” IEEE/ASME Transactions on Mechatronics, vol. 220, no.
1, pp. 258–268, 2017.
[35] J. Smoczek, “Experimental verification of a GPC-LPV method
with RLS and P1-TS fuzzy-based estimation for limiting the
transient and residual vibration of a crane system,” Mechanical
Systems and Signal Processing, vol. 62, pp. 324–340, 2015.
[36] M. P. Spathopoulos and D. Fragopoulos, “Pendulation control
of an offshore crane,” International Journal of Control, vol. 770,
no. 7, pp. 654–670, 2004.
[37] M. Short and F. Abugchem, “A microcontroller-based adaptive
model predictive control platform for process control applica-
tions,” Electronics (Switzerland), vol. 60, no. 4, 2017.
 International Journal of

Rotating Advances in
Machinery Multimedia

The Scientific
Engineering
Journal of
Journal of

Hindawi
World Journal
Hindawi Publishing Corporation Hindawi
Sensors
Hindawi Hindawi
www.hindawi.com Volume 2018 http://www.hindawi.com
www.hindawi.com Volume 2018
2013 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018

Journal of

Control Science
and Engineering

Advances in
Civil Engineering
Hindawi Hindawi
www.hindawi.com Volume 2018 www.hindawi.com Volume 2018

Submit your manuscripts at

www.hindawi.com

Journal of
Journal of Electrical and Computer
Robotics
Hindawi
Engineering
Hindawi
www.hindawi.com Volume 2018 www.hindawi.com Volume 2018

VLSI Design
Advances in
OptoElectronics
International Journal of

International Journal of
Modelling &
Simulation
Aerospace
Hindawi Volume 2018
Navigation and
Observation
Hindawi
www.hindawi.com Volume 2018
in Engineering
Hindawi
www.hindawi.com Volume 2018
Engineering
Hindawi
www.hindawi.com Volume 2018
Hindawi
www.hindawi.com www.hindawi.com Volume 2018

International Journal of
International Journal of Antennas and Active and Passive Advances in
Chemical Engineering Propagation Electronic Components Shock and Vibration Acoustics and Vibration
Hindawi Hindawi Hindawi Hindawi Hindawi
www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018