UAV Cooperative Decision and Con-

trol: Challenges and Practical Approaches

edited by TAL SHIMA and STEVEN RASMUSSEN
SIAM, 2008, ISBN 0898716640, 186 pages, US$85
Reviewed by Sonia Martı́nez

Broadly considered, the field of cooperative decision and control cov-

ers those interdisciplinary methods that can be used for the operation of
semi-autonomous agents deployed to achieve a common objective. By joint
exploiting the agents’ capabilities, it is expected that the combined effort of
the team can exceed the sum of its parts. Harnessing this potential benefit is
however challenging due to the complexity that dealing with miscellaneous
components in dynamic and uncertain environments brings about.

In general, cooperative decision and control algorithms find application

in the supervision of large networked systems, such as sensor and actuator
monitoring networks, supply-chain management systems, the power grid, and
traffic systems. A special and important use is the command and control
of teams of unmanned aerial vehicles (UAVs) in surveillance and combat
scenarios. This topic is the main focus of this book, which, as one of the first
of its kind, exposes research challenges and practical approaches to operating
these systems in military scenarios.

More specifically, UAV Cooperative Decision and Control: Challenges and

Practical Approaches shows the reader what the main issues of the field are

1
by describing several algorithms that enable a team of heterogenous UAVs
to identify, classify, locate, and attack sets of targets under possibly imper-
fect information. These algorithms, a sample of the authors’ work since the
late nineties, are analyzed according to aspects that include scalability, com-
putational tractability, heterogenous vehicles, coordination under imperfect
communications, and uncertainty due to false perception about the environ-
ment. As a valuable component, the book describes a simulation package,
the MultiUAV2 Cooperative Control and Simulation tool. This simulation
package is a Matlab/Simulink/C++ simulation tool, freely available to the
public, which can be used to test any cooperative control algorithm under
more realistically modeled mission scenarios. For example, the user can test
the algorithms against more realistic models of UAV dynamics, inter-vehicle
communications, and various types of sensors.

As the authors point out, there are several types of cooperative con-
trollers depending on the information available to agents, the organization
of the agents, and the performance expected of the controllers. For example,
controllers can be divided into centralized, hierarchical, or decentralized, re-
spectively, depending on how global information is available to each, some,
or none of the members of the team. The book focuses mainly on centralized
controllers since, at this point, these are the main class of controllers that
can effectively handle inter-vehicle and mission constraints. The methodol-
ogy considered for centralized algorithm synthesis is based on different types
of programming techniques and can deal with sources of uncertainty and false
information.

2
 Although primarily classifiable as a research monograph, the book does
an excellent job in introducing the reader to the main issues of interest as
well as a representative sample of cooperative control techniques employed
in DoD scenarios. With respect to previous volumes [1, 2, 3, 4, 5], the text
goes beyond the mere compilation of specialized papers, making an effort to
present the subject in a unified way that is accessible to students, academi-
cians, and nonspecialized control researchers alike. I personally believe the
book achieves this objective, and parts of the book can be easily included
into the syllabus of a graduate course on the subject. The book is definitely
one of the first that have been published in this area, and complements re-
cent publications such as [6, 7, 8], which pay more attention to decentralized
controllers based on local interaction rules, such as those based on consensus
algorithms, linear iterations, and geometric optimization problems.

Organization and Contents

The book is organized into seven chapters and two appendices. Chapter 1
takes the reader through a nice historical tour on the development of un-
manned aerial vehicles (UAVs) and autonomous navigation systems. Team
cooperation and its advantages in terms of joint efficiency and robustness to
failure are introduced. A classification of cooperative controllers is made ac-
cording to various metrics, such as the information available to UAVs (local
versus global), the structure of the controller (centralized versus hierarchical
versus decentralized), and its performance level (optimal or adaptive team
behavior.) The chapter ends with a description of three example mission

3
scenarios and the bottlenecks associated with them. These scenarios are (i)
combat intelligence, surveillance, and reconnaissance of moving ground tar-
gets, (ii) electronic attack to deceive a network of radars, that is, the phantom
problem, and (iii) target classification in urban environments with micro air
vehicles.

Chapter 2 expands on the discussion of the application of cooperative

teams and challenges of cooperative control. The chapter starts by distin-
guishing among cooperative/non-cooperative behaviors of multivehicle sys-
tems. Then, various distributed decision and control systems are revisited
in terms of their performance capabilities. The discussion is followed by a
description of the complexity spawned by cooperative controllers. Basically,
the objective of reaching stringent performance requirements (task coupling,
robustness, optimality, scalability) with limited controllers (partial local in-
formation, communication delays, high degree of autonomy) in uncertain
environments is what leads to this complexity. As the authors note, there
is no single approach that can address all the facets of cooperative control.
Optimality can be addressed with mixed-integer linear programming and
stochastic dynamic programming in centralized scenarios, but scalability is a
limiting factor. On the other hand, decentralization can be in conflict with a
severe team coupling imposed by the global performance index and objective
constraints, which can lead to an excessive message traffic. Finally, com-
munication limitations and false information raises the question of to what
extent cooperation is beneficial. The rest of the book is devoted to describing
approaches that the authors have found useful in the aforementioned scenar-
ios. These approaches are multidisciplinary and bring together tools from
fields such as operations research, computer science, and control theory.

4
 Chapter 3 focuses on the problem of assigning UAVs to multiple targets.
UAVs are small powered vehicles that have a life span of 30 minutes and can
perform four tasks, namely, search, classification, attack, and verification, in
a specific order. Due to these capabilities, the tasks can be state-constrained.
For example, a constraint is imposed by the fact that an object classifica-
tion can only be successfully performed under a certain UAV angle of attack.
In this situation, the problem is solved through a linear programming for-
mulation, as a capacitated transshipment assignment problem (CTAP). An
iterative algorithm with memory weighting solves the N -target-N -UAV as-
signment problem in a fast way every time a new target is found, while
respecting the task precedence constraints. The algorithm is compared with
Monte Carlo simulations, and the authors point out advantages and limita-
tions of the CTAP approach.

Chapter 4 delves into the problem of multiple assignments for a group

of UAVs through a mixed-integer linear programming approach. In contrast
with the CTAP, this approach allows the inclusion of timing constraints and
the consideration of a wider range of joint task assignment and scheduling
problems. Alternatively, a decision tree formulation together with a breadth
search algorithm and a genetic algorithm is explored to find globally optimal
solutions. The chapter ends with simulations and a discussion on the pros
and cons of all the algorithms introduced so far.

Chapter 5 considers the cooperative moving target engagement (CMTE)

assignment problem. In this scenario, a group of heterogenous UAVs are to
jointly track a moving ground target and intercept it. The various sensing

5
capabilities of the UAVs (types of limited-sensing footprints, restrictions on
relative bearing angles), timing constraints (a UAV can perform a tracking
task only in some situations), and dynamical constraints (for example, one
of the UAVs needs to fly in tight circular orbits) make this problem a highly
complex one. An additional difficulty is given by the fact that the UAVs
need to solve overlapping and joint tasks when planning for their paths. The
authors translate this scenario into a combinatorial optimization problem by
considering “time-availability windows,” in which the UAVs are available to
carry out a certain task. Then, the problem is reduced to how to combine the
times that allow the completion of the task in minimum time. The paths of
the UAVs are calculated to meet the time constraints. In order to solve this
optimization problem, the authors explore the MILP and genetic algorithm
approaches.

Chapter 6 discusses how imperfect communications and uncertainty may

impact cooperative controllers and ways to overcome this. In connection
with the assignment problems of previous chapters, the authors present a
decision-estimation architecture that allows agents use the CTAP approach
from Chapter 3 in dealing with uncertainty in their teammates positions and
their costs to intercept known targets. Here, uncertainty is a consequence of
delays in the transmission of messages between agents, or a possible asyn-
chronous communication architecture.

An important source of uncertainty is of course due to the lack of infor-

mation about the environment and targets. Chapter 7 explores this topic
through effectiveness measurements. In particular, tools from probability

6
and optimal control theory are used to find cooperative rules that maximize
the probability of attacking at least n targets while constraining the proba-
bility of attaching m false targets. A numerical solution method is discussed
and then evaluated in the situation of a single target scenario that is uni-
formly distributed in a linear-symmetric battle space among a Poisson field
of false targets.

The book ends with two appendices. Appendix A presents the Multi-
UAV2 simulation package, which models inter-vehicle communications, six-
degree-of-freedom dynamics, and nonlinear aerodynamic lookup tables to
accurately model the motion of the UAVs. The software can be used to model
a variety of missions through the specification of “managers” to represent
sensors, targets, weapons, and cooperation tactic rules.

Appendix B is on path-planning strategies for UAVs. The resulting al-

gorithms could be part of a high-level module that would allow the UAVs
to execute predefined flight paths. The appendix starts by discussing how
track-guidance autopilots can be used to allow UAVs follow a path. UAV
modeling for path planning is discussed next through the Dubins kinematic-
vehicle system. The appendix ends by discussing how to extend the available
method for planning in the presence of wind.

7
Conclusions

This volume UAV Cooperative Decision and Control: Challenges and Prac-
tical Approaches presents a unified approach to the management of multiple
UAVs in military scenarios. The book describes precisely what the main
challenges associated with the field are, which in turn motivate the high re-
search activity in this area. By gradually increasing the level of complexity, it
presents specific solutions to particular target assignment problems through
a blend of methodologies that mainly draw from the field of optimization,
making the exposition very amenable to adaptation for a graduate course on
the subject. As the authors note, many issues are left out of the scope of
the book. For example, the inclusion of game-theoretic techniques for selfish
agents would warrant a second volume. The contents can also be comple-
mented with other books that deal with popular decentralized approaches,
which are not covered here. Surely, this book is one of many to come in
the area of cooperative decision and control, addressing the many problems
generated by this fascinating and high impact subject.

References

[1] V. Kumar, N. Leonard, and A. Morse, eds., Cooperative Control: A Post-

Workshop Volume, 2003 BBlock Island Workshop on Cooperative Control.
Springer-Verlag, 2007.

[2] J. Shamma, ed., Cooperative control of distributed multi-agent systems.

8
 John Wiley & Sons, 2008.

[3] S. Butenko, R. Murphey, and P. Pardalos, eds., Recent Developments

in Cooperative Control and Optimization. New York: Kluwer Academic
Publishers, 2004.

[4] S. Butenko, R. Murphey, and P. Pardalos, eds., Cooperative Control:

Models, Applications and Algorithms. New York: Kluwer Academic Pub-
lishers, 2003.

[5] R. Murphey and P. Pardalos, eds., Cooperative Control and Optimization.

New York: Kluwer Academic Publishers, 2002.

[6] W. Ren and R. Beard, Distributed Consensus in Multi-vehicle Cooperative

Control. Communications and Control Engineering, London: Springer-
Verlag, 2008.

[7] P. Zhihua Qu, Cooperative Control of Dynamical Systems: Applications

to Multi-vehicle systems. London: Springer-Verlag, 2009.

[8] F. Bullo, J. Cortés, and S. Martı́nez, Distributed Control of Robotic Net-

works. Applied Mathematics Series, Princeton University Press, 2009.
Available at http://www.coordinationbook.info.

9
Reviewer Information

Sonia Martinez is an assistant professor with the Mechanical and Aerospace

Engineering Department at UC San Diego. She received the Ph.D. degree in
Engineering Mathematics from the Universidad Carlos III de Madrid, Spain,
in May 2002. Following a year as a visiting assistant professor of applied
mathematics at the Technical University of Catalonia, Spain, she obtained a
Postdoctoral Fulbright Fellowship and held positions as a visiting researcher
at UIUC and UCSB. Her main research interests include nonlinear control
theory, robotics, cooperative control, and networked control systems. In par-
ticular, her work has focused on the modeling and control of robotic sensor
networks, the development of distributed coordination algorithms for groups
of autonomous vehicles, and geometric control of mechanical systems. For
her work on the control of underactuated mechanical systems, she received
the Best Student Paper award at the 2002 IEEE Conference on Decision
and Control. She was the recipient of an NSF CAREER Award in 2007.
For the paper “Motion coordination with Distributed Information,” coau-
thored with Francesco Bullo and Jorge Cortes, she received the 2008 Control
Systems Magazine Outstanding Paper Award.

10