A unified approach to the velocity-free consensus algorithms design for
double integrator dynamics with input saturations
Abdelkader Abdessameud and Abdelhamid Tayebi
Abstract— This paper considers the consensus problem of
double integrator multi-agent systems where each agent is
subject to input saturations, and the velocity (second state) of
each agent is not available for feedback. We present a unified
approach to the consensus algorithms design that extends
most of the existent consensus algorithms developed for double
integrator multi-agents in ideal situations to handel these two
problems simultaneously. To illustrate the effectiveness of the
proposed approach, we present solutions to three different sec-
ond order consensus problems and provide simulation results.
I. I NTRODUCTION
In contrast to multi-agents with first order dynamics,
consensus algorithms for double integrators can be naturally
extended to design cooperative control strategies for complex
physical systems with applications to flocking [1], formation
control of unmanned vehicles [2]-[3], rigid body attitude syn-
chronization [4]-[5] and synchronization of networked Euler-
Lagrange systems [6]-[7]. The consensus problem of double
integrators involves the design of consensus algorithms such
that agents can reach an agreement on their states, or on a
common objective, using local information exchange. This
information exchange is generally restricted to be directed,
dynamically changing, and may be delayed.
In the related literature to the second order consensus
problem, tools from algebraic graph theory have been suc-
cessfully applied to establish conditions under which second
order consensus is reached. In directed networks, it has been
shown that second order consensus will be reached if and
only if the communication graph has a spanning tree and
the control gains are carefully selected [8], [9]. Within a
similar framework, several related problems to consensus
have been considered such as the formation control problem,
[10], consensus with group reference velocity, [11] and
leader-follower problems, [12]. Also, the case of dynamically
changing topologies have been discussed in [8] and [13]. The
effects of communication delays that are inherently present
in communication systems have also been considered in [14]-
[16] and references therein. However, the above consensus
algorithms are based on the assumption that the full state
vector is available for feedback.
In practice, it is sometimes desirable to design consensus
algorithms that do not require full states information. If we
This work was supported by the Natural Sciences and Engineering
Research Council of Canada (NSERC).
The authors are with the Department of Electrical and Computer En-
gineering, University of Western Ontario, London, Ontario, Canada. The
second author is also with the Department of Electrical Engineering, Lake-
head University, Thunder Bay, Ontario, Canada. aabdess@uwo.ca,
tayebi@ieee.org
consider, for example, a group of point masse agents, an
important problem is to design consensus algorithms in the
case where the velocity information (the second state) is
not available for feedback, either because it is not precisely
measured or agents are not equipped with velocity sensors.
Another important problem arises when the input of each
agent is subject to input saturations. Unfortunately, the papers
dealing with these two problems are not numerous and
only the simple case of fixed and undirected communication
topology has been considered. The author in [11] proposed
consensus algorithms that account for input saturations in
the full state information case. In the same reference, the
author presents a second order consensus algorithm that re-
moves the requirement of velocity measurements. Consensus
algorithms that take into account the two above problems
simultaneously have been proposed in [17]. In this reference,
a new approach, based on auxiliary systems, has been
proposed to simplify the consensus algorithm design problem
in this case. However, it is difficult to show that the results
in [17] are applicable under more general communication
topologies, that may be directed, time-varying and/or subject
to communication delays.
The main contribution of this paper is to provide a
unified approach that extends most of the existent consensus
algorithms, developed for double integrator dynamics with
a certain communication topology, to account for input
saturations and remove the requirement of velocity mea-
surements. Instrumental to our approach is the introduction
of two second order auxiliary systems that simplify the
consensus algorithm design. The first auxiliary system is
used to generate an intermediate reference trajectory for
each agent, and its input is designed such that all agents
reach an agreement on their reference trajectories. The
input of the second auxiliary system is designed such that
each agent tracks its corresponding intermediate reference
trajectory without velocity measurements. With this setting,
the control input of each agent is constructed using only
the auxiliary states to account for input saturations. As a
result, the consensus algorithm design problem with the
above mentioned constraints is reduced to the design of
a consensus algorithm in ideal situations, i.e., in the full
state information case and without input saturations. To show
the effectiveness of the obtained results, we consider three
different problems related to second order consensus, and
extend some consensus algorithms developed for double
integrator dynamics to account for input saturations in the
partial state feedback case.
2011 50th IEEE Conference on Decision and Control and
European Control Conference (CDC-ECC)
Orlando, FL, USA, December 12-15, 2011
978-1-61284-799-3/11/$26.00 ©2011 IEEE 4903