You are here

Networked Robotics at IROS 2009

nikolaus-correll's picture

Although Networked Robots were formally represented in only two sessions at the recent IROS conference in St. Louis, a large body of work in the conference (more than 45 papers) related more or less strongly to our field. This is a non-comprehensive overview over larger thrusts, new directions and novel ideas presented at the conference.

Around ten papers deal with formation control. Robustness to the pertinent challenges imposed by unreliable communication and localization seem to be more and more in the focus of interest. For instance Mead et al. ("Fault-Tolerant Formations of Mobile Robots") present algorithms for formation repair and obstacle avoidance in flocking robots, some of which are implemented on a team of 20 robots. In Gupta et al. ("Collective Transport of Robots: Coherent, Minimalist Multi-Robot Leader-following") flocking of a team of 4 iRobot Create is achieved without explicitly maintaining group coherence, but by following a leading robot and local obstacle avoidance.

Two papers address the relatively new problem of deploying and maintaining mobile wireless networks. For instance Chiu et al. ("TENTACLES: Self-Configuring Robotic Radio Networks in Unknown Environments") describe a distributed algorithm for maintaining multi-hop connectivity between given points in an environment and provide experimental results with 7 iCreates with Ubiquiti radios in a hallway.

An important robot capability in most networked robot applications is relative range and bearing. While using the radio itself for localization is appealing it turns out that signal strength alone provides only a very coarse estimate of range. Schiff et al. introduce novel algorithms ("Nonparametric Belief Propagation for Distributed Tracking of Robot Networks with Noisy Inter-Distance Measurements") for creating collaborative estimates of robot locations using three static beacons.  Sun et al. propose the use of multiple antennas and phase shifts of the wireless carrier ("Robot Localization and Energy-Efficient Wireless Communications by Multiple Antennas"), which - if implemented - might also lead to more efficient robot-to-robot communication by beamforming.

Growing interest in using commodity radios for localization has also led to a series of paper on modeling wireless signal propagation. Mostofi et al. provide extensive data on multi-path fading and distance fading in ("Characterization and Modeling of Wireless Channels for Networked Robotic and Control Systems - A comprehensive Overview"). Fink et al. study radio signal propagation of Zigbee and Bluetooth signals experimentally and show that these measurements match those generated by existing radio signal models after spatial averaging in ("Experimental Characterization of Radio Signal Propagation in Indoor Environments with Application to Estimation and Control"). An implementation of a high-fidelity wireless model in a multi-robot simulator is described by Shell et al. in ("High-Fidelity Radio Communications Modeling for Multi-Robot Simulation").

Two papers also address a rather novel problem of recharging robot teams using mobile charging stations. For instance, Drenner et al. present algorithms for coordinating a limited number of charging stations distributed in a building among a team of robots ("Coordinating Recharging of Large Scale Robotic Teams") and shows a mobile charging station that can charge and transport 6 miniature robots.

At least three papers present implementations of complex networked robot systems. Correll et al. describes a distributed robot garden in which a team of manipulator robots and potted tomato plants communicate wirelessly for negotating tasks such as watering and harvesting ("Building a Distributed Robot Garden"). Shiomi et al. reports on a ("Field Trial of Networked Social Robots in a Shopping Mall"). The system is able to interact with people that are tracked using static laser range finders. The robots then provide directions and guide people through a shopping mall.  Guarnieri et al. presents ("HELIOS System: A Team of Tracked Robots for Special Urban Search and Rescue Operations") where a team of manipulator robots collaboratively localizes, provides sensing information to a command station (combined visual/3D maps), and can do heavy manipulation including lifting obstacles. Barbosa et al. describe ("ISRobotNet: A Testbed for Sensor and Robot Network Systems") and demonstrate cooperative perception and decision making problems using a team of 5 ground robots and 10 static cameras.

Combining networked robots with static cameras was the theme of multiple papers, for instance Pahliani et al. who consider Markov Decision Processes for planning optimal paths that use static cameras for robot localization while keeping the target in sight in ("Decision-theoretic Robot Guidance for Active Cooperative Perception"). Bistry et al. propose active cameras to provide task-relevant data to service robots in ("Task Oriented Control of Smart Camera Systems in the Context of Mobile Service Robots") and Xu et al. uses active cameras that coordinate for tracking people in ("Systems and Algorithms for Autonomously Simultaneous Observation of Multiple Objects Using Robotic PTZ Cameras Assisted by a Wide-Angle Camera").

Other interesting work on networked robot systems investigates sample path planning in sensor networks (Hombal et al., "Adaptive Multiscale Sampling in Robotic Sensor Networks"), data ferrying of sensor networks (Bhadauria et al., "Data Gathering Tours for Mobile Robots"), hopping sensors (Pei et al., "Hopping Sensor Relocation in Rugged Terrains"), or the multi-robot coordination algorithms of the winning team in the middle-size league in the RobotCup 2008 (Lau et al., "Multi-Robot Team Coordination through Roles, Positionings and Coordinated Procedures").
Nikolaus Correll