1) Title
Search and Pursuit/Evasion in the Physical World: Efficiency, Scalability, and Guarantees
2) Abstract
Full-Day Workshop
This workshop will bring together researchers in all areas of pursuit/evasion and search. This includes
(but is not limited to) geometric approaches, probabilistic optimization, graph theoretic search, decision
theoretic planning, and adversarial search. The primary focus will be on multi-robot search, which
includes unmanned ground (UGVs), aerial (UAVs), surface (USVs), and underwater (UUVs) vehicles, as
well as heterogeneous search teams in both outdoor and indoor environments. Various sensors,
including cameras and laser scanners, will be examined, and search for both adversarial and nonadversarial targets will be considered. The theme of this workshop is to unify various formulations of the
search problem and move towards a common theoretical and applied framework for search with
autonomous robots. This workshop also will be of interest to those working in the general areas of
multi-agent systems, distributed computation, and approximation algorithms.
3) Organizers



Geoffrey Hollinger
Carnegie Mellon University
Robotics Institute
Newell-Simon Hall
5000 Forbes Ave
Pittsburgh, PA 15213 USA
+1-904-993-1584, gholling@andrew.cmu.edu
Volkan Isler
University of Minnesota
Department of Computer Science
200 Union St SE, EECS 4-213
Minneapolis, MN 55455 USA
+1-612-625-1067, isler@cs.umn.edu
Timothy H. Chung
Naval Postgraduate School
Department of Operations Research
1411 Cunningham Road
Glasgow-261
Monterey, CA 93940 USA
+1-831-656-7858, thchung@nps.edu
4) Presenters with affiliation and status confirmation
Confirmed Presenters
 David Hsu, Associate Professor, Dept. of Computer Science, National Univ. of Singapore
 Dinesh Manocha, Professor, Dept. of Computer Science, Univ. of North Carolina at Chapel Hill
 Gaurav Sukhatme, Professor, Computer Science Dept., Univ. of Southern California
 Karl Hedrick, Professor, Mechanical Engineering Dept., Univ. of California at Berkeley
 Nicholas Roy, Associate Professor, Dept. of Aeronautics and Astronautics, MIT
 Stefano Carpin, Assistant Professor, Computer Science Dept., Univ. of California Merced
 Seth Hutchinson, Professor, ECE, Univ. of Illinois Urbana-Champaign
 Sven Koenig, Associate Professor, Computer Science Dept., Univ. of Southern California
 Tomonari Furukawa, Associate Professor, Mechanical Engineering Dept., Virginia Tech
Conditionally Confirmed Presenters
 Athanasios Kehagias, Assistant Professor, Dept. of Mathematics, Aristotle Univ. of Thessaloniki
 Brian Gerkey, Research Scientist, Willow Garage
 Francesco Bullo, Professor, Dept. of Mechanical Engineering, Univ. of California at Santa Barbara
5) List of topics










Graph search and theoretical foundations
Geometric approaches to search in polygonal environments
Adversarial search on discrete and continuous environment representations
Game theoretic approaches to coordinated search
Bounded approximation algorithms for search
Distributed planning and estimation for multiple searchers
Analysis of sensing and perception for search
Urban search and rescue and emergency response with mobile robots
Search with Unmanned Aerial Vehicles and/or Unmanned Underwater Vehicles
Combining search and target tracking
6) Motivation and objectives
Motivation:
The motivation for this workshop is to unify various approaches to search and pursuit/evasion in the
physical world. The search problem takes on various forms depending on assumptions made regarding
the environment, searchers, and target. The environment may be continuous or represented as a
discrete graph. The searchers may have perfect or imperfect sensing, and they may be allowed to move
instantaneously or be subject to kinematic or dynamic constraints. The target may be adversarial, nonadversarial, omniscient, and/or capable of moving with bounded or unbounded speed. The target and
searchers may have perfect or imperfect knowledge of each other, and capture may be defined as
touching, gaining line-of-sight to, or surrounding the target.
The organizers of this workshop see a rift between probabilistic search and pursuit/evasion games.
Researchers in the probabilistic search community make a non-adversarial assumption on the target and
formulate the search problem as the optimization of an objective function related to expected capture
time. Good approximations are available for solving this problem using either deterministic or
randomized policies. In contrast, pursuit/evasion games are characterized by an adversarial target
actively avoiding capture. Game theoretic and graph theoretic solutions exist for determining the
number of searchers necessary to find a target and for generating strategies to guarantee capture.
Despite the obvious similarities, probabilistic search and pursuit/evasion games differ significantly in
their formulations, assumptions, and solutions. This workshop seeks to define the state-of-the-art for
both probabilistic search and pursuit/evasion games and to move towards a unified framework. This
unified framework will be essential for robots to find both adversarial and non-adversarial targets in
unstructured environment in the physical world.
Objectives:
1) The main objective is to bring together researchers from the probabilistic optimization
community with those working on theoretical graph search. This will help to move towards a
unified approach to the search problem with mobile robotic agents.
2) In addition, this workshop will define the state-of-the-art for approximation algorithms in both
non-adversarial and adversarial search.
3) New avenues of research will be discussed and evaluated for theoretical and applied
pursuit/evasion and search.
4) The final goal is to bring more researchers into the pursuit/evasion and search community. This
is particularly true for those interested in multi-agent coordination, distributed systems, and
approximation algorithms
7) Primary/secondary audience
Primary audience: Those working in pursuit/evasion and search with unmanned ground (UGVs), aerial
(UAVs), surface (USVs), and/or underwater (UUVs) vehicles.
Secondary audience: Anyone interested in probabilistic optimization or graph search. Researchers with
interest in multi-agent systems, distributed computation, and approximation algorithms.
8) Relation to previous ICRA and IROS workshops
To our knowledge, this is the first search and pursuit/evasion workshop at ICRA or IROS. However, there
have been several related workshops at previous ICRA meetings:
 The “Robotics for Disaster Response” workshop at ICRA09 included presentations on urban
search and rescue topics. Our proposed workshop will include talks on the current state-of-theart in urban search and rescue, which will extend this theme.
 The “Cooperative Control of Multiple Heterogeneous Unmanned Aerial Vehicles for Coverage
and Surveillance” as well as the “Unmanned Air Vehicles: Civilian and Commercial
Opportunities” workshops at ICRA08 dealt with many problems related to pursuit/evasion and
search on aerial vehicles. Our proposed workshop will focus on the search problem and will
incorporate a wider range of unmanned vehicles.