556 IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL. 56, NO.

3, MARCH 2011

Robust Model Predictive Control With Integral

Sliding Mode in Continuous-Time
Sampled-Data Nonlinear Systems
Matteo Rubagotti, Student Member, IEEE, Davide Martino Raimondo, Antonella Ferrara, Senior Member, IEEE,
and Lalo Magni

Abstract—This paper proposes a control strategy for nonlinear

constrained continuous-time uncertain systems which combines
robust model predictive control (MPC) with sliding mode control
(SMC). In particular, the so-called Integral SMC approach is used
to produce a control action aimed to reduce the difference between
the nominal predicted dynamics of the closed-loop system and the
actual one. In this way, the MPC strategy can be designed on a
system with a reduced uncertainty. In order to prove the stability
of the overall control scheme, some general regional input-to-state
practical stability results for continuous-time systems are proved.
Index Terms—Constrained control, nonlinear predictive control Fig. 1. Scheme of the overall hierarchical control system.
(NPC), sampled data control, sliding mode control (SMC), stability
of nonlinear systems.
forcing the satisfaction of the constraints for any possible
perturbation [9]–[14]. This approach presents a very high
I. INTRODUCTION computational burden, and, at the moment, can be applied
to systems with small size or slow dynamics.
M ODEL predictive control (MPC) is a control technique
which permits to cope with a constrained system pro-
viding an optimal control strategy. MPC has been widely used
• The open-loop nominal approach with tightened con-
straints, where the real constraints are shrunk in order
in the process industry and also studied in its theoretical aspects to guarantee that the original constraints are fulfilled by
by the research community (see for an overview, [1]–[3], and the real system for any possible uncertainty realization
the books [4]–[7]). In the last years, one of the topics of in- [15]–[19]. This approach has the drawback of being too
terest in MPC has been the definition of robust strategies, in conservative in view of the open-loop off-line approxima-
order to guarantee certain stability properties and the respect of tion of the worst possible effect of the disturbance that is
the constraints also in the presence of uncertainties and external required in order to compute the tightened constraints.
disturbances. Several nonlinear MPC (NMPC) algorithms with Exploiting the idea proposed in [16], the second approach
guaranteed robust stability with respect to different classes of is here followed. The hierarchical control scheme (see Fig. 1),
disturbances have been proposed (see, e.g., [8] and the refer- composed by a NMPC algorithm and a Sliding Mode controller,
ences therein). In order to obtain a robust controller in presence is introduced in order to improve the robustness features of the
of constraints two main approaches have been proposed in the NMPC algorithm. Sliding Mode Control (SMC) [20], [21] is
literature. a well-known robust control technique for nonlinear systems
• The min-max approach, where the objective function is which guarantees the complete elimination of the effect of
minimized for the worst possible uncertainty realization, “matched disturbances” (i.e. disturbances acting on the control
input channel) once the system is in the so-called “sliding
mode,” that is, the state has reached a suitable subspace of
Manuscript received September 21, 2009; revised February 12, 2010; ac-
cepted June 22, 2010. Date of publication September 13, 2010; date of cur- the state space, called “sliding manifold.” Considering that
rent version March 09, 2011. This work was supported in part by the European matched disturbances are very common and can represent both
Commission under the Project Feednetback FP7-ICT-223866 and by the Italian parametric uncertainties and external disturbances, robustness
PRIN project “Model predictive control algorithms for artificial pancreas.” Rec-
ommended by Associate Editor D. Angeli.
with respect to them is a significant benefit, also taking into
M. Rubagotti is with the Department of Mechanical and Structural account that SMC requires a very low computational burden
Engineering, University of Trento, Trento 38123, Italy (e-mail: matteo. and guarantees the finite-time convergence to the sliding man-
[email protected]). ifold, upon which the controlled system exhibits the desired
D. M. Raimondo was with the Automatic Control Laboratory, Swiss Federal
Institute of Technology (ETH), Zürich 8092, Switzerland and is now with the dynamics.
Department of Computer Engineering and systems Science, University of Pavia, Different approaches to merge MPC and Sliding Mode
Pavia 27100, Italy (e-mail: [email protected]). techniques have already been proposed in the literature. For
A. Ferrara and L. Magni are with the Department of Computer Engineering
and Systems Science, University of Pavia, Pavia 27100, Italy (e-mail: antonella.
instance, in [22] a combined scheme was presented for linear
[email protected]; [email protected]). systems in the framework of Generalized Predictive Control.
Digital Object Identifier 10.1109/TAC.2010.2074590 Another approach has been proposed in [23], where MPC was

0018-9286/$26.00 © 2010 IEEE

Authorized licensed use limited to: Nazarbayev University. Downloaded on February 18,2025 at 09:33:44 UTC from IEEE Xplore. Restrictions apply.
RUBAGOTTI et al.: ROBUST MODEL PPC WITH INTEGRAL SM 557

used to update the parameters of the sliding manifold. The id represents the identity function from to . Given a set
Sliding Mode strategy considered in this paper is instead de- , is the point-to-set distance
signed according to the so-called Integral Sliding Mode (ISM) from to while denotes the boundary of . Given
approach [24]–[27]. This technique is a recent development of two sets , then the Pontryagin difference set
the more classical SMC design methodology that presents the is defined as .
advantage of forcing the system state to lie on the sliding man-
ifold from the initial time instant. Relying on the knowledge of The floor function is defined as follows:
the nominal model of the system and of the control signal gen- . A function is of class (or
erated by the NMPC, the ISM controller is designed to produce a ” -function”) if it is continuous, positive definite and strictly
a continuous-time control action aimed to reduce the difference increasing, and . A function is of
between the dynamics of the nominal closed-loop system and class if it is a -function and as .
the actual evolution of the state. In this way the NMPC can A function is of class if, for
be designed on a system with reduced uncertainty, limiting each fixed , is of class , for each fixed ,
the conservativeness of the open-loop nominal approach. In is decreasing and as . Given a matrix
particular, if only matched disturbances affect the system, no with , then its orthogonal complement is
tightened constraints are required in the NMPC formulation, .
while, if also unmatched disturbances are considered, then less
conservative constraints must be introduced with respect to the
pure NMPC control scheme. III. PROBLEM STATEMENT
The continuous-time setting is the most appropriate if the In this paper, it is assumed that the plant to be controlled is
plant model is derived from first principle continuous-time described by the continuous-time nonlinear model
equations. Nonetheless, since solving an optimization problem
in continuous time would be computationally untractable, (1)
some MPC algorithms which use continuous-time models with
sampled data systems have been proposed [28]–[30]. In this where is the state, is the current control
paper, following the approach analyzed in [29] for systems not vector, is the disturbance term, and is a
affected by uncertainty, the optimization is performed in dis- compact set containing the origin as an interior point. Given
crete-time with respect to piecewise-constant control signals.
system (1), which is assumed to be forward complete, assume
The continuous-time control law of the ISM is then added to
also that denotes the nominal model,
the piecewise-constant signal generated by the NMPC. In order
to prove the stability of the overall control scheme, the regional being , with ,
Input-to-State Stability (ISS) and Input-to-State practical Sta- , . The system can then be expressed
bility (ISpS) results introduced respectively in [31] and [32] for as
discrete-time systems are proved for continuous-time systems.
A preliminary version of the theoretical development here (2)
proposed, without mathematical proofs, can be found in [33].
The organization of the paper is the following: the notations where and
used in the paper are reported in Section II, while Section III denotes the additive uncertainty. The solution of system (2) with
deals with the description of the system. In Section IV the initial state and the uncertain signal is denoted by
overall control strategy, including the robust NMPC control .
strategy and the ISM controller is presented. Section V studies Remark 1: The control-affine form
the stability properties of the control system. The simulation is required in order to obtain an explicit
results and the conclusions are reported in Sections VI and control law for the ISM strategy that will be presented in the
VII, respectively. Finally, for the readers’ convenience, all the sequel.
proofs are in the Appendix, together with the introduction of System (2) is supposed to fulfill the following assumption.
the concept of regional ISpS in continuous-time. Assumption 1:
1) System (2) is forward complete.
II. NOTATIONS 2) .
3) The state and control variables are restricted to fulfill the
The Euclidean norm is denoted as . For any symmetric following constraints
matrix , and denote the largest and the
smallest eigenvalue of matrix , respectively. Given a signal (3)
, let be a signal defined from time to time . In (4)
order to simplify the notation, when it is inferrable from the
context, the subscript of the sequence is omitted. The set of sig- where and are compact sets containing the origin
nals , the values of which belong to a compact set as an interior point.
is denoted by , while . Moreover 4) The uncertainty is such that
where denotes the values that
the signal takes in correspondence to the time . The symbol (5)
Authorized licensed use limited to: Nazarbayev University. Downloaded on February 18,2025 at 09:33:44 UTC from IEEE Xplore. Restrictions apply.
558 IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL. 56, NO. 3, MARCH 2011

where is a compact set containing the origin, with components of will be denoted by and
known. , respectively.

Remark 2: This last property can be obtained starting from

IV. ROBUST NMPC WITH ISM: THE
system (1), if it is assumed that the disturbance is such that
OVERALL CONTROL STRATEGY
, where is a compact set containing the origin, with
known. Given (3), (4), one can state that can be modeled Robust NMPC with tightened constraints can lead to very
as a bounded uncertainty as well. conservative solutions or even to unfeasible problems. In order
Given a suitable sampling period , and letting , with to mitigate the conservativeness of the approach, the hierar-
, be the sampling instants, the control objective consists in de- chical control scheme represented in Fig. 1 is here proposed,
signing a piecewise-constant control law that steers the system where the ISM controller is introduced in order to reduce the
state to (a neighborhood of) the origin fulfilling the constraints uncertainty that must be taken into account by the NMPC con-
on the input and the state itself along the system evolution for troller.
any possible uncertainty satisfying Point 4 in Assumption 1, and The control variable , according to Fig. 1, is composed
yielding an optimal closed-loop performance according to a cer- by two parts, i.e.
tain performance index.
Assumption 2: System (2) is defined such that, considering a (9)
generic time instant , ,
1) given two different initial conditions and at where the component is generated by the NMPC con-
time 0, and a signal , it yields troller, while is generated by the ISM controller to com-
pensate for the uncertain terms. As a cost to pay for the uncer-
(6) tainty reduction, one has that part of the control effort must be
used by the ISM controller so that the input constraints imposed
where is a positive continuous function defined in the NMPC optimization problem are smaller than the actua-
in , such that . The term stands tors limits. The ISM controller, based on the knowledge of the
for raised to the -th power. nominal continuous-time model of the system and of the control
2) given an initial condition at time 0, the signals signal generated by the NMPC control law, produces a control
and , one has that action in order to reduce the difference between the dynamics
of the nominal closed-loop system and the actual evolution of
(7) the state. Moreover, the ISM, having a negligible computational
burden, can be used in continuous-time.
where is a constant value and .
A. Integral Sliding Mode Strategy
In order to evaluate the discrepancy between the nominal and
perturbed evolutions of the system at a generic time instant, the Given system (2), define a sliding manifold, i.e. a set
following lemma can be stated. , where, as in [34]
Lemma 1: Suppose that Assumptions 1 and 2 are satisfied.
Then, given and , one has that

(10)
and is a projection matrix, defined such that the
for all , all and . matrix product is invertible. It is important to note that the
Proof: See Appendix C. initial state belongs to the sliding manifold, i.e.
Given the feedback control law . The uncertain term can always be seen as the sum of two
different parts
(8)

following the idea used in [29], the description of the hold mech-
anism implicit in (8) calls for a state augmentation. Letting where , , and .
, the closed loop system (2), (8) is given by The uncertainty is the matched uncertainty, and can be
perfectly compensated by the SMC action [20]. The other term
is called “unmatched uncertainty” and cannot be compensated
by a sliding mode strategy. The control variable can
be defined in several ways, for instance by relying on to the
so-called unit-vector approach

and its solution from initial time 0 and initial state

(11)
is given by . Moreover, the first and the last
Authorized licensed use limited to: Nazarbayev University. Downloaded on February 18,2025 at 09:33:44 UTC from IEEE Xplore. Restrictions apply.
RUBAGOTTI et al.: ROBUST MODEL PPC WITH INTEGRAL SM 559

where is the controller gain (considered constant here for the the matched uncertainty could imply an increasing of the un-
sake of simplicity) large enough to keep from matched one, which must be kept as small as possible with the
the initial time instant. The choice of this particular gain as choice of a suitable sliding manifold.
will permit to dominate the matched uncertainty Remark 4: In case no unmatched uncertainties are present
term: in principle, one could also choose a larger value for , but (i.e. , or , ), then this strategy will com-
this would induce useless wear in the system and would subtract pletely eliminate the uncertain terms, so that the NMPC con-
some usable control amplitude to the NMPC controller. How- troller has to control a system without uncertain terms. If this is
ever, some variations of this technique, e.g. letting be a func- the case, a robust NMPC strategy is no longer needed, and one
tion of the states and/or a time-varying value can be found in can use the technique described in [29].
the literature (see, e.g. [21]). To determine the motion equations
at the sliding manifold, the equivalent control method [20] is B. Robust NMPC Strategy
used. The equivalent control is the continuous signal which can
be determined by solving, with respect to , the equation The robust NMPC controller must be designed for system
, taking into account (2) and (10). Substituting the (12) formed by the system under control complemented with
value of the equivalent control in (2), it yields the ISM control law. Following the idea behind the control al-
gorithm presented in [16] for discrete-time systems and consid-
ering that system (12) is a particular case of system (2), a new
robust NMPC control algorithm for continuous-time systems in
where form (2) is described in this subsection. Some preliminary defi-
nitions and results are first introduced. In particular, in order to
describe a key ingredient of the robust NMPC controller, i.e. the
tightened constraints, define the tightened set
Note that the perturbation has been re-
placed by . So, the amount of reduction of the uncertain (14)
term depends on the choice of .
Lemma 2: Applying to system (2) the control law (11) based where
on the sliding manifold (10), with , one has that the
choice minimizes the norm of , i.e.

This definition of the tightened set guarantees that, if the nom-

inal state evolution belongs to in (14), then the perturbed
trajectory of the system fulfills (3), as will be proved in the se-
Then, the resulting system that the NMPC controller has to con- quel.
sider is The proposed NMPC controller is based on the following Fi-
nite-Horizon Optimal Control Problem (FHOCP) that consists
(12) in minimizing, at any sampling time instant , a suitable cost
function with respect to the control sequence
where the matched uncertainty has been eliminated, while the , with being the predic-
unmatched one has not changed. tion horizon. The associated finite horizon piecewise-constant
Proof: See Propositions 2 and 3 in [27]. control signal is such that
For each component of the control variable
, , given the approach
selected to design the ISM component, the value of the ampli-
tude of the terms can be computed, and then a new set for all and all .
can be accordingly found as Definition 1 (FHOCP): Consider system (2) with .
Given the positive integer , the quadratic stage cost
(13) ( and being positive defi-
nite matrices), the quadratic terminal penalty
Defining as in (11), a quantity equal to must be sub- (being a symmetric positive definite matrix) and the terminal
tracted to each component of the control bounds for the NMPC set , the FHOCP problem consists in minimizing with
controller. respect to the cost function
Remark 3: It is very important to note that it would be pos-
sible to use the ISM strategy here described even if the term
were state dependent; nonetheless, no solution has been pro-
posed up to now in the literature to find the optimal realiza-
tion of the sliding manifold that can avoid any increasing of
the unmatched uncertainty. So, in case , depending
on the specific system, one would have that the elimination of
Authorized licensed use limited to: Nazarbayev University. Downloaded on February 18,2025 at 09:33:44 UTC from IEEE Xplore. Restrictions apply.
560 IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL. 56, NO. 3, MARCH 2011

subject to Since is a compact set, no additional assumption is needed

1) the state dynamics (2) with , for all to state that is Lipschitz with respect to the state variable
; in the domain , i.e.
2) the state constraint , for all ;
3) the control constraint (4); (20)
4) the terminal state constraint .
Remark 6: In the FHOCP, continuous-time state constraints
Remark 5: Since and are bounded, the stage cost is a are considered. It can appear that this approach is only concep-
Lipschitz function with respect to both the state and the control tual, because a numerical implementation would need a time
values, i.e., there exist and such that discretization and the constraints satisfaction could be checked
(15) only at the integration time instants. Nonetheless, this is not
a significant limitation: choosing an integration step small
(16) enough in the optimization phase to simulate the plant (i.e.
), one can still have guarantees on the convergence properties
for all and all .
of the control system without increasing the conservativeness
It is now possible to define the NMPC algorithm:
due to the numerical approximation (see Theorem 3 in [29] for
at any sampling time instant , find the optimal con-
a detailed analysis).
trol sequence by solving the FHOCP.
Then, according to the Receding Horizon strategy, define
, and apply the control law V. INPUT-TO-STATE PRACTICAL STABILITY
OF THE CLOSED-LOOP SYSTEM
(17)
In this section, the Input-to-State practical Stability of the
The following assumption on the choice of the design param- closed-loop system (12), (17) is proved. In order to improve the
eters is introduced. readability, the concept of continuous-time regional ISpS, that
Assumption 3: The design parameters and are such will be used along this section, is introduced and analyzed in
that, given a compact set and an auxiliary control law , Appendix C.
1) , , such that In the following, let denote the set of states for which
and is a positive real number; a solution of the FHOCP exists, and consider the following
assumption.
2) , being , and Assumption 4: Suppose that the system parameters have been
; chosen obtaining a value of such that
3) , for all , with
and . (21)
4) is Lipschitz with respect to the state variable in the
domain with Lipschitz constant , i.e.
Lemma 3: [Feasibility] Suppose that system (2) satisfies As-
(18) sumptions 1–4. Then, is a robust positively invariant set
(see Definition 3 in Appendix B) for the closed-loop system (2),
5) if , then (17).
for all . Moreover, one has that Proof: See Appendix C.
; Lemma 4: [Regional ISpS] Suppose that system (2) fulfills
6) the following inequality holds: Assumptions 1–4. Then, the closed-loop system (2), (17) is re-
gional ISpS in .
Proof: See Appendix C.
Remark 7: Note that Lemmas 3 and 4 are proved for a system
in the general form , and then their
applicability is not limited to control-affine systems.
We are now ready to introduce the main stability result for the
for all ; overall control scheme.
7) consider a generic time instant , ; system Theorem 1: Suppose that system (2) fulfills Assumptions 1–4
(2) is such that with and . Then, the hierarchical
closed-loop system defined by (2), (9), (11), (17) (computed
with respect to system (12)) is regional ISpS in .
(19) Proof: See Appendix C.
Note that, with respect to Lemmas 3 and 4, the assumptions
for all , where is a positive con- must be verified with a modified disturbance . This is due to
tinuous function defined in such that . the presence of the ISM inner loop that, using part of the control
energy, rejects the matched disturbance.
Authorized licensed use limited to: Nazarbayev University. Downloaded on February 18,2025 at 09:33:44 UTC from IEEE Xplore. Restrictions apply.
RUBAGOTTI et al.: ROBUST MODEL PPC WITH INTEGRAL SM 561

Fig. 2. Graphical representation of the simulation example.

VI. SIMULATION RESULTS

In this section, the proposed control strategy is applied to a
cart moving on a plane (Fig. 2). The structure of the model is
the same as in [35], and is represented by

where is the displacement of the cart with respect to the equi-

librium position, and is its velocity. is the mass of the cart,
is the stiffness of the spring, and is the damping
factor. The control variable is the force applied to the cart,
while is the load force given by the wind. The presence of
another disturbance term is assumed. The values of the pa-
rameters are , , .
The uncertain terms are bounded as follows:
and , while the control and state constraints are
, , , respectively. The con-
sidered system is a control-affine system with matched
and unmatched disturbances. The matched one is quite
common in mechanical systems, when both the control variables
and the disturbances are forces or torques. The sampling time of
the NMPC control law is chosen as , while the predic-
tion horizon is . The matrices in the stage cost defined in
the FHOCP are chosen as and . As for Assump-
tion 3, the auxiliary control law and the matrix which defines
are found following the approach suggested in [29], and are Fig. 3. Time evolution of the state variables (x : dashed line, x : solid line), of
equal to the control variable and of the unmatched disturbance for the NMPC controller.

and

respectively. The auxiliary control law is Lipschitz with respect

to the state value because it is a proportional control law applied
in a bounded state region.
Example 1: NMPC Only, No Matched Disturbance: Consider
, . In this case, Assumption 1 is satisfied.
As for Assumption 2, the values of the parameters required for
the controller design have been numerically found. They are
, , . With and
, all the points of Assumptions 3 are satisfied. Indeed,
Fig. 4. Region of attraction for both Examples 1 and 3.
since , one obtains that Assumption 4 is fulfilled. In
Fig. 3, the evolution of the state, the control, and the unmatched
disturbance is depicted, in a simulation example obtained using analyze the region of attraction, i.e. the region formed by the
only the NMPC controller with initial condition states for which a solution of the FHOCP exists. In Fig. 4 the
. If all the assumptions are fulfilled, it is also interesting to region of attraction for this example is reported.
Authorized licensed use limited to: Nazarbayev University. Downloaded on February 18,2025 at 09:33:44 UTC from IEEE Xplore. Restrictions apply.
562 IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL. 56, NO. 3, MARCH 2011

Example 2: NMPC Only, Full Disturbance: We allow now

the matched disturbance to assume values different from zero,
i.e. , using again the NMPC controller only. The
function is the same as in the previous example, while the
numerical evaluation of the needed parameters leads to ,
, , , which implies the non-fulfilment
of Assumption 4, and the consequent loss of the regional ISpS
property. As a practical test, we decided to perform a simulation
example in the same conditions of Fig. 3, but with the increased
disturbance (knowing that there is no a-priori guarantee of con-
vergence). As a result, no feasible solution of the FHOCP can
be found at the initial time instant.
Example 3: NMPC-ISM, Full Disturbance: Following the
procedure of [27], the sliding manifold in (10) is chosen as
, where . In this way the effect of a
matched disturbance is eliminated and the effect of is un-
changed. A pseudo-sliding mode technique, which makes use of
an approximation of the sign function [21], is used to define the
ISM control variable, the maximum amplitude of which is equal
to the maximum amplitude of . In practice, this approxima-
tion is often used to avoid the so-called “chattering” effect (see
[36]–[39], and the references therein) that is a high frequency
oscillation of the sliding variable due to the discontinuous con-
trol input. In this way, the maximum control effort that must be
allocated for the ISM controller is equal to the maximum am-
plitude of that is 1 N and then only 3 N can be used by the
NMPC. The values of , , and are the same as in Ex-
ample 1. In Fig. 5, the evolution of the state and the control
variables is shown starting at the same initial condition of Ex-
ample 1 and 2. The matched disturbance term is shown in Fig. 5
as well, while the unmatched one has the same realization of
the unmatched disturbance reported in Fig. 3. Even though now
a much larger disturbance is considered (from
to ), the state trajectory is not distinguishable from the
one reported in Fig. 3. Fig. 4 shows a numerical evaluation of
the region of attraction using the combined strategy, which co-
incides with that evaluated for the Example 1. The size of the
Fig. 5. Time evolution of the state variables (x : dashed line, x : solid line),
region is not reduced because the available control amplitude of the control variable and of the matched disturbance for the NMPC-ISM
is larger than the one which the NMPC controller needs to find controller.
a feasible solution to the FHOCP in the same region of attrac-
tion. If the total maximum control amplitude had been smaller,
then handling a matched disturbance of non-negligible magni- the regional ISpS of a system. This concept is then exploited
tude would have caused a reduction of the region of attraction to analyze, under suitable assumptions, the stability properties
(however, remember that using only the NMPC with the full of the proposed control scheme. Finally, simulation examples
disturbance Assumption 4 is not satisfied, so that a region of at- are given, showing the advantages of the proposed hierarchical
traction does not exist). scheme.

VII. CONCLUSION APPENDIX A

In this paper, a hierarchical Model Predictive Control scheme RESULT ON TIGHTENED SETS
with Integral Sliding Mode for continuous-time nonlinear sys- The following lemma is useful to prove the properties of the
tems is proposed and analyzed. The contribution of this work proposed NMPC control law.
can be summarized in the following three points. First, an ISM Lemma 5: Let , such that
control strategy is designed in order to reduce the conservative- , then .
ness of the robust open-loop NMPC strategy. Second, a robust Proof: Let , and .
NMPC control strategy with tightened constraints, which gen- Then, one has that
erates a piecewise-constant control law, is proposed for contin- ,
uous-time systems. Third, the concept of regional ISpS property and so . Since ,
in continuous-time is introduced, and it is proved that the ex- , it is verified that . Then
istence of a suitably defined Lyapunov function can guarantee .
Authorized licensed use limited to: Nazarbayev University. Downloaded on February 18,2025 at 09:33:44 UTC from IEEE Xplore. Restrictions apply.
RUBAGOTTI et al.: ROBUST MODEL PPC WITH INTEGRAL SM 563

APPENDIX B 3) there exist a suitable -function ,a -function and

CONTINUOUS-TIME REGIONAL ISPS a constant such that
In this subsection the ISS framework for continuous-time au-
tonomous nonlinear systems is presented and Lyapunov-like
sufficient conditions are provided. This framework is used to (25)
prove the stability properties of the proposed NMPC control
scheme. The piecewise-constant control law introduced in Sec- for all , all , and for almost all . For the
tion IV-B will make the closed-loop system time-varying: it is values , , for which (25) does not hold, the following
then necessary to deal with time-varying systems. condition holds:
Consider a time-varying continuous-time autonomous non-
linear dynamic system described by (26)

(22) 4) there exist suitable -functions and (with such that

is a -function) such that, given an uncertain
where is nonlinear and pos- signal , there exists a nonempty compact set
sibly discontinuous, is the state, is an (including the origin as an interior point) defined
uncertain term. The solution of system (22) with initial state as follows:
and the sequence of the uncertain terms is de-
noted by . This system is supposed to fulfill the (27)
following assumption.
Assumption 5: where ,
1) System (22) is forward complete. with , ,
2) The uncertainty is such that , .

Whenever , is said to be an ISS-

Lyapunov function for system (22) in .
where is a compact set containing the origin with Then, the following sufficient condition for regional ISpS of
known. system (22) can be stated.
Theorem 2: If system (22) admits an ISpS-Lyapunov
Definition 2 (RPI Set): A set is a robust positively function in with respect to , then it is ISpS in and
invariant (RPI) set for system (22) if , for .
all , all , all , and all . Proof: Let . The proof will be carried out in three
A regional version of ISpS [40] is now defined. steps.
Definition 3: [Regional ISpS in ] Given a compact set Step 1) First we show that the set defined in (27) is ro-
including the origin as an interior point, system (22) with bust positively invariant for system (22). From the definition of
is said to be ISpS (Input-to-State practical Stable) in it follows that . There-
with respect to , if is a RPI set for (22) and if there exist fore and hence
a -function , a -function and a constant , . Moreover (see [41])
such that

(28)
where is a -function. Then, the following
for all and . Whenever , system (22) is said holds:
to be ISS (Input-to-State Stable) in with respect to .
Regional ISpS will be now associated to the existence of a
suitable Lyapunov function (in general, a priori, non-contin-
uous) with respect to . (29)
Definition 4 (ISpS-Lyapunov Function in ): A function
is called an ISpS-Lyapunov function in for all , all , and for almost all t, being
for system (22) with respect to if . Let us now assume that . Then, if
1) is a compact RPI set including the origin as an interior is robust positively invariant, , .
point. In order to prove this claim, assume that this is not true. Then,
2) there exist a pair of suitable -functions and a there exist some and such that .
constant such that Let . Then, it
follows that:
(23)
(24)
Authorized licensed use limited to: Nazarbayev University. Downloaded on February 18,2025 at 09:33:44 UTC from IEEE Xplore. Restrictions apply.
564 IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL. 56, NO. 3, MARCH 2011

that implies (we allow the possibility that ). For

, define
if
By substituting this in (29) we obtain if

In order to verify that , we only

need to show that

for all , all . Hence, considering also

that, where does not exist, (26) holds, that is
for some . This contradicts the minimality
of , and hence, as claimed, is robust positively invariant
(i.e. for all ).
Step 2) Next, we show that the state, starting from ,
tends asymptotically to . If , then
. Hence

which is equivalent to

for all , and for almost all , the last step being ob-
tained using (23). Considering also that, where does
not exist, (26) holds, then, such that

(30) By the following change of variables and taking

into account (26), denoting one sees
Therefore, starting from , the state will arrive close to that the previous inequality is equivalent to
in a finite time and to asymptotically. Hence
.
Step 3) Finally we show that system (22) is regional ISpS in
. Let . Then, from the
robust invariance of , it follows that
, for all . By using (23), this implies
that

Noting that, given a -function ,

(see [41]) it follows that:

where are due to the jump discontinuities of and,

(31) for (26), are always negative. By considering (32), the previous
inequality is always satisfied. It only remains to show that
where and is of class . The function is continuous since both and
. For , , that implies are continuous in their domains, and .
It is strictly increasing in for each fixed since both and
are strictly decreasing. Finally as by
(32) construction. So is a -function.
Hence , . By
for almost all . using (23) and (24), it yields
First of all, this inequality guarantees that is defined
for all . Secondly, by the following generalization of (33)
the comparison principle [42], there exists some -function
which only depends on and , such that where and .
for . Combining (31) and (33), one concludes that the system is
ISpS, i.e.
Define for any , . This
is a strictly decreasing differentiable function on , with
. Let . Then,
the range of , and hence the domain of , is the open interval where .
Authorized licensed use limited to: Nazarbayev University. Downloaded on February 18,2025 at 09:33:44 UTC from IEEE Xplore. Restrictions apply.
RUBAGOTTI et al.: ROBUST MODEL PPC WITH INTEGRAL SM 565

APPENDIX C is the value of the state of the NMPC perturbed closed-loop

OTHER PROOFS system at time . To determine the feasibility of such a solu-
Proof of Lemma 1: Using (7), if the nominal system and the tion, one must prove the following three steps.
real one start at the same point at time , then after a sampling Step 1) it is necessary to show that the state value must lay in
time interval it yields at , i.e.

After two sample time intervals, one has that where is the signal associated with the con-
trol sequence . In order to prove this, we show
that . To this
aim, note that Assumption 2 implies

where and .
Using the triangle inequality together with Assumption 2, it where
yields
(37)

is the value of the state of the NMPC nominal (without un-

certainties) closed-loop system at time . Since
, by defining

Analogously, after sampling times, the result can be obtained

as the sum of a geometric series
and

Finally, in order to generalize it for a generic interval ,

it yields
it yields
(38)

and then

(39)
where the last inequality is obtained applying (21). At this
Proof of Lemma 3: To get the feasibility property, one has to point, applying , according to Assumption 3, one obtains
prove that .
Step 2) the control variable must fulfill
. It follows from the fact that by definition,
(34) and , since .
Step 3) in order to assure the respect of the state constraints,
Letting and the associated optimal solution it must be verified that
of the FHOCP at time , a possible (sub-op-
timal) solution at time for the FHOCP is

for all time instants in the prediction horizon,

i.e. and . Consid-
(35)
ering that
,
where then by recursion, for all and for
, considering Assumption 3, it yields
(36)
Authorized licensed use limited to: Nazarbayev University. Downloaded on February 18,2025 at 09:33:44 UTC from IEEE Xplore. Restrictions apply.
566 IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL. 56, NO. 3, MARCH 2011

.
Then, since

according to Lemma 5 (Appendix A.), it yields

Being one can see

that a lower bound on the integral of the state evolution can be
Proof of Lemma 4: In order to prove the regional ISpS of
obtained if the state moves towards the origin with a “velocity”
the closed loop system, first note that Assumption 5 is satis-
equal to and, if the origin is reached, it remains there.
fied for system (2). In order to prove the stability properties one
Taking into account the evolution of the system with this
has to find a suitable ISpS Lyapunov function, as described in
maximum derivative, the value of the integral is smaller if
Theorem 2. At any time instant , define calculated on a smaller interval, so, the minimum amount of
this interval, i.e. is considered. This means
that we can write
(40)
as the value of the extended state of the NMPC perturbed
closed-loop system at time , where is the extended state
defined in Section III, as well as and . Then, for a fixed
value of , introduce the following ISpS Lyapunov function
candidate:

where
(41)
(case )
which is defined at any time value, and takes into account the (case ).
remaining part of the stage cost until the end of the prediction
horizon. In this way, the length of the integral is varying from
at the left neighborhood of each sampling instant , In case , where the state reaches the origin before the end of
to at each sampling instant. This choice of the Lyapunov the interval , solving the integral, one has
function is similar to the one made in [29] for the nominal case.
Now we verify that this function satisfies all the points in
Definition 4.
Point 1 is fulfilled, because includes the
origin as interior point, and is a robust positively invariant set
according to Lemma 3.
In case , where the state does not reach the origin in the con-
Point 2 requires to find two -functions and
sidered interval, solving the integral it yields
which are a lower and an upper bound, respectively, for
the ISpS Lyapunov function candidate. As for the lower bound,
one can see that, at

After noting that the two functions coincide at ,

that they are strictly increasing in their argument, that the
function in case is equal to zero for , and that the
Authorized licensed use limited to: Nazarbayev University. Downloaded on February 18,2025 at 09:33:44 UTC from IEEE Xplore. Restrictions apply.
RUBAGOTTI et al.: ROBUST MODEL PPC WITH INTEGRAL SM 567

function in case tends to infinity for , it is possible

to state that (45)

Using (15), (16), (18), (20) and (45), it yields

is a -function, such that .

Now, a procedure to find the upper-bound is intro-
duced. To begin with, define

where . Define also

(42)

(43)

The following cost function with horizon is defined, for

with , and
. Since a compact set, then the value of
can be bounded as . Then,
(44)
defining , it yields
Relying on Point 6 in Assumption 3, it yields

In order to obtain the upper bound of in the re-

By exploiting the optimality of the FHOCP solu- gion , we recall the arguments stated in [41]. The com-
tion, it is possible to state that pactness of and implies that the predicted evolution of
the system and the feasible control action are bounded. This
fact and Assumption 3 guarantee that the optimal cost is upper
. This inequality bounded, that is, there exists a finite real number such that
holds only for , that is the region where a feasible , . Let ,
solution for the FHOCP with prediction horizon surely , be a ball such that . Note
exists. that this ball exists since the origin is in the interior of . Let
As a preliminary result, note that, taking into account (6), be a positive constant . Then, defining
(18), (19), given , one can note that , one has
, and then

In fact, if , then and hence . Then

Authorized licensed use limited to: Nazarbayev University. Downloaded on February 18,2025 at 09:33:44 UTC from IEEE Xplore. Restrictions apply.
568 IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL. 56, NO. 3, MARCH 2011

Point 3 requires to express the derivative of the ISpS Lya-

punov function candidate. In the following the right derivative (48)
is analyzed (the calculation of the left derivative is specular).
Such a derivative is defined as
It is worth noting that the value of in this case can be calculated
considering only the current values of the state and the control
variable, independently from the past trajectory of the system.
(46)
Moreover, this term is defined so that
Note that . Then, by using the definition of in
(40), in addition to
As for the term in the second line of (48), a time-invariant upper
bound can be obtained substituting to . So, after defining
the term on the numerator of (46), provided that (49)
, can be bounded as follows:
and

(50)

where is a -function and is a -function, one

has

(51)

Note that the derivative of the Lyapunov function is defined only

almost everywhere, because at the sampling time instants a
jump discontinuity of can occur. Then, it is nec-
essary to show that .
Considering of having to solve the FHOCP at time , a fea-
sible (sub-optimal) solution can be obtained from the control
sequence defined in (35). The corresponding value of the ISpS
Lyapunov function candidate at would surely be larger than
the one that could be obtained solving the FHOCP at that time
instant. So, also taking into account (20) defining as

(47)

Taking into account that the value of the integrand in the second
line of (47) is defined for any time instant in and is also and as in (43) (with ), and taking into
bounded, define as its mean value in . By virtue of account Point 6 in Assumption 3, it is possible to state that
(6), (7), (15), (20), after some calculation it yields

where

Substituting this upper-bound into (46), solving the limit (taking

into account that if , then coincides with the value of
the integrand in the second line of (47) at time ), one obtains

which proves that .

Authorized licensed use limited to: Nazarbayev University. Downloaded on February 18,2025 at 09:33:44 UTC from IEEE Xplore. Restrictions apply.
RUBAGOTTI et al.: ROBUST MODEL PPC WITH INTEGRAL SM 569

As for Point 4, it is necessary to guarantee that the region [14] M. Lazar, D. Muñoz de la Peña, W. P. M. H. Heemels, and T. Alamo,
defined in (27) is contained in , which coincides with “On input-to-state stability of min-max nonlinear model predictive
control,” Sys. Control Lett., vol. 57, no. 1, pp. 39–48, 2008.
in our case. depends on through functions and con- [15] L. Chisci, J. A. Rossiter, and G. Zappa, “Systems with persistent dis-
stants of which we are just able to compute very conservative turbances: Predictive control with restricted constraints,” Automatica,
upper-bounds. Then, the set has to be such that the feasibility vol. 37, no. 7, pp. 1019–1028, 2001.
[16] D. Limon, T. Alamo, and E. F. Camacho, “Input-to-state stable MPC
condition (21) holds, and that . Due to the con- for constrained discrete-time nonlinear systems with bounded additive
servativeness in the calculation of , the last condition could uncertainties,” in Proc. Conf. Decision Control, Las Vegas, NV, Dec.
be the most stringent. However, it is necessary just to give an es- 2002, pp. 4619–4624.
[17] S. V. Rakovic, A. R. Teel, D. Q. Mayne, and A. Astolfi, Simple Robust
timation of the region where the state of the closed-loop system Control Invariant Tubes for Some Classes of Nonlinear Discrete Time
converges asymptotically. In order not to limit the applicability Systems. San Diego, CA, Dec. 2006.
of the method only to extremely small uncertainties, we avoid to [18] G. Grimm, M. J. Messina, S. E. Tuna, and A. R. Teel, “Nominally
robust model predictive control with state constraints,” IEEE T. Autom.
estimate , knowing that the actual region where the system Control, vol. 52, no. 10, pp. 1856–1870, Oct. 2007.
will converge is included in , this latter being a robust [19] G. Pin, D. M. Raimondo, L. Magni, and T. Parisini, “Robust model pre-
positively invariant set. dictive control of nonlinear systems with bounded and state-dependent
uncertainties,” IEEE T. Autom. Control, vol. 54, no. 7, pp. 1681–1687,
Proof of Theorem 1: The application of the ISM to system Jul. 2009.
(2), according to Lemma 2, leads to a system in form (12), i.e. [20] V. I. Utkin, Sliding Mode in Control and Optimization. New York:
system (2) with . Moreover, in order to apply the Springer-Verlag, 1992.
[21] C. Edwards and S. Spurgeon, Sliding Mode Control: Theory and Ap-
ISM inner loop, the control variable in the MPC control law is plications. New York: Taylor & Francis, 1998.
limited in the set . Then, since Assumptions 1–4 are satisfied [22] W. Garcia-Gabin, D. Zambrano, and E. F. Camacho, “Sliding mode
for system (2) with and , according to predictive control of a solar air conditioning plant,” Control Eng.
Pract., vol. 17, no. 6, pp. 652–663, 2009.
Lemma 4 the ISpS of the overall control scheme is proved. [23] K. R. Muske, H. Ashrafiuon, and M. Nikkhah, “A predictive and sliding
mode cascade controller,” in Proc. Amer. Control Conf., New York, Jul.
ACKNOWLEDGMENT 2007, pp. 4540–4545.
[24] V. I. Utkin and J. Shi, “Integral sliding mode in systems operating under
The authors would like to thank Dr. E. Vitali, Department uncertainty conditions,” in Proc. Conf. Decision Control, Kobe, Japan,
of Mathematics, University of Pavia, for the useful discussion Dec. 1996, pp. 4591–4596.
about the proof of Theorem 2. [25] A. Poznyak, L. Fridman, and F. J. Bejarano, “Mini-max integral
sliding-mode control for multimodel linear uncertain systems,” IEEE
T. Autom. Control, vol. 49, no. 1, pp. 97–102, Jan. 2004.
REFERENCES [26] M. Basin, J. Rodriguez, L. Fridman, and P. Acosta, “Integral sliding
[1] M. Morari and J. H. Lee, “Model predictive control: Past, present and mode design for robust filtering and control of linear stochastic
future,” Comput. Chem. Eng., vol. 23, no. 4–5, pp. 667–682, 1999. time-delay systems,” Int. J. Robust Nonlin. Control, vol. 15, no. 9, pp.
[2] D. Q. Mayne, J. B. Rawlings, C. V. Rao, and P. O. M. Scokaert, 407–421, 2005.
“Constrained model predictive control: Stability and optimality,” [27] F. Castaños and L. Fridman, “Analysis and design of integral sliding
Automatica, vol. 36, no. 6, pp. 789–814, 2000. manifolds for systems with unmatched perturbations,” IEEE T. Autom.
[3] S. J. Qin and T. A. Badgwell, “A survey of industrial model predictive Control, vol. 51, no. 5, pp. 853–858, May 2006.
control technology,” Control Eng. Pract., vol. 11, no. 7, pp. 733–764, [28] F. A. C. C. Fontes, “A general framework to design stabilizing non-
2003. linear model predictive controllers,” Sys. Contr. Lett., vol. 42, no. 2,
[4] J. M. Maciejowski, Predictive Control With Constraints. Upper pp. 127–144, 2001.
Saddle River, NJ: Pearson Education, 2002. [29] L. Magni and R. Scattolini, “Model predictive control of contin-
[5] E. F. Camacho and C. Bordons, Model Predictive Control. New York: uous-time nonlinear systems with piecewise constant control,” IEEE
Springer Verlag, 2004. T. Autom. Control, vol. 49, no. 6, pp. 900–906, Jun. 2004.
[6] J. B. Rawlings and D. Q. Mayne, Model Predictive Control: Theory [30] L. Grune, D. Nesic, and J. Pannek, “Model predictive control for non-
and Design. Madison, WI: Nob Hill Publishing, 2009. linear sampled-data systems,” in Nonlinear Model Predictive Control:
[7] , L. Magni, D. M. Raimondo, and F. Allgöwer, Eds., Nonlinear Model Towards New Challenging Applications, R. Findeisen, F. Allgöwer,
Predictive Control: Towards New Challenging Applications. New and L. Biegler, Eds. New York: Springer-Verlag, 2007, pp. 105–113.
York: Springer-Verlag, 2009. [31] L. Magni, D. M. Raimondo, and R. Scattolini, “Regional input-to-state
[8] D. Limon, T. Alamo, D. M. Raimondo, D. Muñoz de la Peña, J. M. stability for nonlinear model predictive control,” IEEE T. Autom. Con-
Bravo, A. Ferramosca, and E. F. Camacho, “Input-to-state stability: A trol, vol. 51, no. 9, pp. 1548–1553, Sep. 2006.
unifying framework for robust model predictive control,” in Nonlinear [32] D. M. Raimondo, D. Limon, M. Lazar, L. Magni, and E. F. Camacho,
Model Predictive Control: Towards New Challenging Applications, L. “Min-max model predictive control of nonlinear systems: A unifying
Magni, D. M. Raimondo, and F. Allgöwer, Eds. New York: Springer- overview on stability,” Eur. J. Control, vol. 15, no. 1, pp. 5–21, 2009.
Verlag, 2009, pp. 1–26. [33] M. Rubagotti, D. M. Raimondo, A. Ferrara, and L. Magni, Robust Model
[9] H. Chen, C. W. Scherer, and F. Allgöwer, “A game theoretical ap- Predictive Control of Continuous-Time Sampled-Data Nonlinear Sys-
proach to nonlinear robust receding horizon control of constrained sys- tems With Integral Sliding Mode. Budapest, Hungary, Aug. 2009.
tems,” in Proc. Amer. Control Conf., Albuquerque, NM, Jun. 1997, pp. [34] W. J. Cao and J. X. Xu, “Nonlinear integral-type sliding surface for
3073–3077. both matched and unmatched uncertain systems,” IEEE T. Autom. Con-
[10] P. O. M. Scokaert and D. Q. Mayne, “Min-max feedback model predic- trol, vol. 49, no. 8, pp. 1355–1360, Aug. 2004.
tive control for constrained linear systems,” IEEE T. Autom. Control, [35] L. Magni, G. De Nicolao, Scattolini, and F. R. Allgöwer, “Robust
vol. 43, no. 8, pp. 1136–1142, Aug. 1998. model predictive control for nonlinear discrete-time systems,” Int. J.
[11] A. Bemporad, F. Borrelli, and M. Morari, “Min-max control of con- Robust Nonlin. Control, vol. 13, no. 3–4, pp. 229–246, 2003.
strained uncertain discrete-time linear systems,” IEEE T. Autom. Con- [36] G. Bartolini, A. Ferrara, and E. Usai, “Chattering avoidance by second-
trol, vol. 48, no. 9, pp. 1600–1606, Sep. 2003. order sliding mode control,” IEEE T. Autom. Control, vol. 43, no. 2, pp.
[12] F. Fontes and L. Magni, “Min-max model predictive control of non- 241–246, Feb. 1998.
linear systems using discontinuous feedbacks,” IEEE T. Autom. Con- [37] L. Fridman, “An averaging approach to chattering,” IEEE T. Autom.
trol, vol. 48, no. 10, pp. 1750–1755, Oct. 2003. Control, vol. 46, no. 8, pp. 1260–1265, Aug. 2001.
[13] M. Diehl and J. Bjornberg, “Robust dynamic programming for [38] L. Fridman, “Chattering analysis in sliding mode systems with inertial
min-max model predictive control of constrained uncertain sys- sensors,” Int. J. Control, vol. 9, no. 10, pp. 906–912, 2003.
tems,” IEEE T. Autom. Control, vol. 49, no. 12, pp. 2253–2257, [39] A. Levant, “Chattering analysis,” in Proc. Eur. Control Conf., Kos,
Dec. 2004. Greece, Jul. 2007, pp. 3195–3202.
Authorized licensed use limited to: Nazarbayev University. Downloaded on February 18,2025 at 09:33:44 UTC from IEEE Xplore. Restrictions apply.
570 IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL. 56, NO. 3, MARCH 2011

[40] E. D. Sontag and Y. Wang, “New characterizations of input-to state- Antonella Ferrara (S’86–M’88–SM’03) was born
stability,” IEEE T. Autom. Control, vol. 41, no. 9, pp. 1283–1294, Sep. in Genova, Italy. She received the Laurea degree
1996. in electronic engineering and the Ph.D. degree in
[41] D. Limon, T. Alamo, F. Salas, and E. F. Camacho, “Input to state sta- computer science and electronics from the Univer-
bility of min-max MPC controllers for nonlinear systems with bounded sity of Genova, Genova, Italy, in 1987 and 1992,
uncertainties,” Automatica, vol. 42, no. 5, pp. 797–803, 2006. respectively.
[42] J. Y. Lin, E. D. Sontag, and Y. Wang, “A smooth converse Lyapunov She has been Assistant Professor at the University
theorem for robust stability,” SIAM J Control Optim., vol. 34, no. 1, pp. of Genova since 1992. In November 1998 she be-
124–160, 1996. came Associate Professor of Automatic Control at the
University of Pavia. Since January 2005 she is Full
Professor of Automatic Control in the Department of
Computer Engineering and Systems Science, University of Pavia. Her research
Matteo Rubagotti (S’07) was born in Voghera (PV), activities are mainly in the area of sliding mode control, with application to au-
Italy, in 1982. He received the “Laurea” and “Laurea tomotive control, process control, and robotics. She has authored or coauthored
Specialistica” degrees (with highest honors) in com- more than 230 papers, including more than 70 international journal papers.
puter engineering, and the Ph.D. degree in electronic, Dr. Ferrara was an Associate Editor of the IEEE TRANSACTIONS ON CONTROL
computer science and electrical engineering from the SYSTEMS TECHNOLOGY and, at present, she is Associate Editor of the IEEE
University of Pavia, Italy, in 2004, 2006, and 2010, TRANSACTIONS ON AUTOMATIC CONTROL. She is a member of the IEEE Tech-
respectively. nical Committee on Variable Structure and Sliding Mode Control, a member
From 2008 to 2009, he was a Visiting Scholar at of the IEEE Robotics and Automations Technical Committee on Autonomous
the Ohio State University Center for Automotive Re- Ground Vehicles and Intelligent Transportation Systems, and a member of the
search, Columbus, OH. From July to August 2009 IFAC Technical Committee on Transportation Systems.
he was an Academic Guest at the Automatic Control
Laboratory, ETH Zürich, Switzerland. Since January 2010, he is a Post-Doc
with the Department of Mechanical and Structural Engineering, University of
Trento, Italy. His research is mainly focused on analysis and controller synthesis Lalo Magni was born in Bormio (SO), Italy, in 1971.
of uncertain nonlinear and piecewise-affine systems (in particular, using sliding He received the M.S. degree (with highest honors) in
mode and model predictive control techniques), and on mobile robotics. computer engineering and the Ph.D. degree in elec-
tronic and computer engineering from the University
of Pavia, Pavia, Italy, in 1994 and 1998, respectively.
From 1999 to 2004, he was Assistant Professor at
Davide Martino Raimondo was born in Pavia, the University of Pavia where he has been Associate
Italy, in 1981. He received the Laurea degree (with Professor since 2005. From 1996 to 1997 and in 1998
highest honors) in computer engineering and the he was at CESAME, Universit Catholique de Lou-
Ph.D. degree in electronic, computer and electrical vain, Louvain La Neuve (Belgium). From October to
engineering from the University of Pavia, Italy, in November 1997, he was at the University of Twente
2005 and 2009, respectively. with the System and Control Group in the Faculty of Applied Mathematics. In
From November 2006 to May 2007, he was with 2003, he was a Plenary Speaker at the Second IFAC Conference “Control Sys-
the Escuela Superior de Ingenieros, Universidad de tems Design” (CSD’03). In 2005, he was Keynote speaker at the NMPC Work-
Sevilla, Sevilla, Spain. Since January 2009 he is shop on Assessment and Future Direction. His current research interests include
working as post-doc at Automatic Control Labora- nonlinear control, predictive control, robust control, process control and glu-
tory, ETH Zürich, Switzerland. He is the author or cose concentration control in subjects with diabetes. His research is witnessed
coauthor of more than 30 papers published in refereed journals, edited books, by more than 40 papers published in the main international journals. In 2003,
and refereed conference proceedings. His current research interests include he was Guest Editor of the Special Issue ‘Control of nonlinear systems with
nonlinear control, model predictive control, networked control, robust control, Model Predictive Control’ in the International Journal of Robust and Nonlinear
input-to-state stability, and glycemia control in diabetic subjects. Control.
Dr. Raimondo co-organized the NMPC Workshop on Assessment and Future Dr. Magni was an Associate Editor of the IEEE TRANSACTIONS ON
Direction, Pavia, Italy, in 2008. He was also co-organizer of the invited sessions AUTOMATIC CONTROL. He is Associate Editor of Automatica. He organized the
Nonlinear Model Predictive Control, for IFAC NOLCOS 2010, and New De- NMPC Workshop on Assessment and Future Direction in September 2008 in
velopments in Nonlinear Model Predictive Control, for IFAC NOLCOS 2007. Pavia.
He served as an Associate Editor of IFAC NOLCOS 2010.

Authorized licensed use limited to: Nazarbayev University. Downloaded on February 18,2025 at 09:33:44 UTC from IEEE Xplore. Restrictions apply.