Workshop at 2016 IEEE ICRA supported by the Technical Committee on Multi-Robot Systems.

Venue: Stockholm Waterfront Center, Room: Room 24

Date and Time: Friday May 20th 2016, from 1:30pm to 6pm

This workshop aims at providing an overview of current research related to real world deployment of heterogeneous multi­robot systems. Many operational situations can indeed benefit from the deployment of a team of robots, in which the robots may have different motion, sensing and decision abilities, or complementary operational payloads. The objective of the workshop is to present the most recent results, obtained by the main research groups active in the field, and to highlight what are the main challenges to be faced to achieve effective deployment of heterogeneous multi­robot systems.

• 13:30 Introduction
• 13:40 Challenges for Perception in Heterogeneous Multi-­Robot Environments

Igor Gilitschenski, Autonomous Systems Lab (ETH Zurich, Switzerland) 

As the number of different robotic applications increases, it becomes more and more common to involve different types of robots simultaneously. This results in several advantages for perception systems compared to a setting that involves only one single robot because richer information may become available. In order to benefit from these advantages, several challenges need to be addressed. First, when using more than one robot, it becomes necessary to fuse information obtained from individual robots. Second, use of different types of robotic platforms often requires consideration of a broader variety of sensor types with different sensing modalities, and thus, different types of measurements. In this talk, we will discuss how these challenges can be addressed in practical applications, particularly for estimation and mapping scenarios. This discussion will cover simultaneous consideration of different levels of uncertainties, a sound representation of underlying domains, and cooperation of multiple heterogeneous agents in large­ -scale map building. The presented methodology will involve directional and robust statistics, and use of optimization based estimation approaches.

• 14:05 Heterogeneous Aquatic Robot Teams 
  Geert De Cubber, Royal Military Academy (Brussels, Belgium)

Robotic systems are finding more and more their way to the field. One application field where they have proven their benefits is in the field of crisis management and disaster relief. Of course, when deploying multiple robotic systems into a crisis area, one must consider carefully the interoperability aspects between these fielded systems. Indeed, standardization is still no easy concept in the world of robotics and certainly in the case of multi-domain crisis, it is easy to predict that there may and will be important interoperability concerns between light UAS, medium-size UGVs and potentially heavy UMS, all with their own control and communication architectures and protocols. The EU ICARUS project has investigated these interoperability issues and has proposed potential solutions in the form of standardized interfaces for multi-domain systems. These solutions have been extensively validated using 3 separate field tests. A shipwreck accident was simulated near the Lisbon harbor. The Portuguese Navy intervened to rescue the victims in the water, helped by the ICARUS UAS and UMS. An earthquake was simulated in Belgium. The Belgian First Aid and Support Team intervened to rescue the survivors, helped by the ICARUS UAS and UGVs. Important during all these exercises was the complete integration of the unmanned tools in the toolkit of the rescue workers and into their standard operating procedures. As a third validation test, the ICARUS team participated to the euRathlon2015 trial in Italy, simulating a multi-domain accident requiring the intervention of UAS, UMS and UGVs at the same time. In this presentation, the speaker will show how the ICARUS multi-domain interoperability concept was applied in these three use cases of fielded multi-robot systems, operating on land, air and sea.

• 14:30 The Use of Multirobot Systems in Open Pit Mining Applications 

Luiz Chaimowicz, Universidade Federal de Minas Gerais (Brazil)

In this talk, we will present an ongoing collaboration project between the Vision and Robotics Lab. of the Universidade Federal de Minas Gerais (VeRLab – UFMG) and Vale’s Technological Institute (ITV), in Brazil. Vale is one of the largest mining companies in the world, and has several applications that would benefit from the use of multiple heterogeneous robot systems. The objective of this project is exploring and evaluating the use of multirobot systems in mining applications. More specifically, in this talk, we will present some preliminary results on path planning and control of multiple UAV’s for magnetometric surveying and also on techniques for fast 3D reconstruction from sequences of georeferenced aerial and terrestrial images, focusing mainly on open pit iron ore mines. 

• 14:55 Poster teasers
  • Multirobot Coordination by Multiplayer Games, Arash Tavakoli and Nora Ayanian (Department of Computer Science, University of Southern California, Los Angeles USA)

  • Collaboration Between Similar Crawling Robots of Various Sizes, Moran Sabbag, Gal Trumper, and David Zarrouk (Department of Mechanical Engineering, Ben Gurion University of the Negev, Israel)

  • Multi-Robot Path Planning for Selective Coverage, Sandeep Manjanna, Nikhil Kakodkar, and Gregory Dudek (School of Computer Science, McGill University, Canada)

  • The MAVIS System: Towards the Use of Marsupial Robotic Networks for Automatic Sensing in Polar Regions, A. Petitti, D. Di Paola, R. Colella, C. Patruno, M. Ianigro, A. Milella, R. Maglietta, M. Bibuli, L. Caviglione, D. Chiarella, A. Odetti, A. Ranieri, E. Zereik, M. Caccia, and G. Bruzzone (Institute of Intelligent Systems for Automation, Italy)

  • Active Terrain Mapping for an Autonomous Air-Ground Search and Rescue System, Jeffrey Delmerico, Elias Mueggler, Julia Nitsch, Davide Scaramuzza (Robotics and Perception Group, University of Zurich, Switzerland)

  • Wearable multi-modal interfaces for mixed-initiative interaction in human multi-robot teams, B. Gromov, L. M. Gambardella, Gianni A. Di Caro (Dalle Molle Institute for Artificial Intelligence, Lugano, Switzerland).

  • Development of an Integrated Iceberg Observation System using Unmanned Underwater, Surface and Aerial Platforms, Ralf Bachmayer, Mingxi Zhou, Zhi Li, Armin Strobel and Brad DeYoung (Autonomous Ocean Systems Laboratory, Memorial University of Newfoundland, Canada).

• 15:20 Coffe break

• 15:40 Poster presentations

• 16:00 Deploying teams of heterogeneous robots for surveillance missions 

Charles Lesire-­Cabaniols, Onera (France) 

Surveillance missions, for infrastructure security or after disasters, are becoming a serious application for robotics. They demand high autonomy skills as the environment may be partially known, highly dynamic, or even hazardous. In addition to navigation skills, deliberation and cooperation are then mandatory to manage teams of robots in such missions. In this presentation, I will present the planning and execution architecture we have developed and deployed in several surveillance missions: (1) patrolling with ground and air vehicles with limited communications, (2) wreck localization with submarine, surface and air vehicles, and (3) ad­hoc network deployment for nuclear plant surveillance. (1) and (2) have been conducted during the Action project funded by the French Armaments Procurement Agency (DGA) (3) has been conducted in the French­German ANCHORS project funded by the French National Research Agency (ANR) and the German Federal Ministry of Education and Research (BMBF).

• 16:25 Mixed­-initiative and Multimodal Interaction with a Robotic Heterogeneous Platform for Searching and Rescue Missions 

Vincenzo Lippiello Università di Napoli Federico II (Italy)

We present a multimodal framework suitable for a human rescuer that interacts with a set of co­located drones and a rover during search and rescue missions. In contrast with typical human­multirobot interaction scenarios, here the operator is not fully dedicated to the control of the robots, but directly involved in search and rescue tasks, hence only able to provide fast, incomplete, although high­value, instructions to the robots. This operative scenario requires a framework that supports intuitive multimodal communication along with an effective and natural mixed­initiative interaction between the human and the robots. In this talk, we illustrate the human­robot interaction framework we are currently designing for this domain describing the system at work in simulated and real case studies.