Maurizio Piaggio

Coordination in multi-agent RoboCup teams

Abstract In this paper, we focus on various aspects of coordination in the framework of the RoboCup competitions. In particular, we address both multi-agent systems that have been developed within the simulation league and multi-robot systems that have been realized in the middle-size league. From the multi-agent perspective we present a behavior-based technique for position selection and the so-called holonic approach. In the multi-robot domain, we address both communication and distributed coordination within a heterogeneous team of autonomous robots.

Communication and Coordination Among Heterogeneous Mid-Size Players: ART99

Distributed coordination among robotic soccer agents has been considered in the recent years within the framework offered by the RoboCup competitions, mostly in the simulation and F-180 leagues. In this paper we describe the methods and the results achieved in coordinating the players of the ART team participating in the F-2000 league. The team is formed by several heterogeneous robots having different mechanics, different sensors, different control software, and, in general, different abilities for playing soccer. The coordination framework we have developed has been successfully applied during the 1999 official competitions allowing both for a significant improvement of the overall team performance and for a complete interchangeability of all the robots.

Pre-emptive versus non-pre-emptive real time scheduling in intelligent mobile robotics

Autonomous and semi-autonomous mobile robots have to perform a multiplicity of concurrent activities in order to carry out useful tasks in unstructured human-populated environments. Even if it is commonly accepted that a successful accomplishment of assigned tasks requires some sort of real time capability to quickly react and adapt to environmental changes, it is not clear which operating system support is best suited for the scheduling and synchronizing of concurrent activities with different timing requirements. This paper discusses this problem, comparing two different real time scheduling policies for autonomous robot applications: pre-emptive rate monotonic and non pre-emptive Earliest Deadline First (EDF). Experimental results are presented and evaluated.

A programming environment for real-time control of distributed multiple robotic systems

In recent years there has been great interest in robot software control architectures. However, although many interesting solutions have been presented, most of the research problems tackled related to a single robot perception, navigation and action in everyday environments. Instead, most of the practical applications of mobile robotics for service tasks in civilian environments consist of systems composed of multiple robots communicating with each other, with external sensing and actuating devices, and with external supervising workstations. RoboCup offers a great opportunity to deal with this problem. In fact the software architecture of a robot soccer player must allow successful intra-robot integration of the different activities (visual perception, path planning, strategy planning, motion control, etc.) spanning many different types of representation (raw sensor data, images, symbolic plans, etc.) and it must also guarantee successful inter-robot integration by supporting communication and cooperation. This paper focuses on this problem, presenting ETHNOS-IV - a programming environment for the design of a real-time control system composed of different robots, devices and external supervising or control stations - which has been successfully used within the Italian ART robot team in the RoboCup-99 competition. ETHNOS provides support from three main point of views which will be addressed in detail: inter-robot and intra-robot communication, realtime task scheduling, and software engineering and code reuse. Experimental results illustrating the advantages of this approach will also be presented.

Distributing a robotic system on a network: the ETHNOS approach

Cognitive activity in intelligent robotic systems has often been modeled as a set of communicating intelligent distributed agents or modules. Some examples in this field are blackboard architectures, hybrid models or subsumption architectures. The rapid progress of communication technology offers the possibility of distributing computation not only on different processes but on a network of computers. This both results in greater available computational power and it allows the robot to merge with the environment it operates in. In suitable intelligent buildings a mobile robot may open doors, turn on/off lights or even avoid obstacles based not only on its sensors and actuators but on the interaction with other robotic entities. In addition the range of robot interactions is now only limited by the network and thus the robot can operate remotely on the environment. Similarly, users can issue commands to remote robots and receive feedback in real-time. In this paper we propose a global approach to distribut...

A Cognitive Hybrid Model for Autonomous Navigation

Action representation and planning is one among the most important research fields in which it has been experienced the failure of single paradigms in isolation to solve real, complex problems. The goal of this paper is to present a system for action representation and reasoning in complex, real-world, and real time scenarios, characterised by the integration of different representation paradigms: symbolic, diagrammatic, and procedural. In this sense the system is called “hybrid”. The paper focuses on the cognitive model and on the representation and reasoning system. A realistic navigation system for the guidance and control of autonomous mobile robots is used as an example to describe the potentiality of the system in solving real complex problems and it is currently being tested in an indoor environment.

An information exchange protocol in a multi-layer distributed architecture

We propose a novel protocol to exchange information within a distributed multi-layer architecture for adaptive intelligent robotic systems. Initially the architecture and its dedicated operating system ETHNOS (Expert Tribe in a Hybrid Network Operating System) are presented. They are both based on experts: symbolic, sub-symbolic and procedural processes that operate concurrently to accomplish different tasks. The OS allows the experts to be distributed over a network to take advantage, depending on the tasks to be carried out, of the particular computational power of a machine. It is the system that transparently handles communication between experts, regardless of where they are in execution using the integrated EIEP (Expert Information Exchange Protocol). This paper focuses on the description of the EIEP, its implementation and usage. Finally experimental results and applications in mobile robotics are illustrated.

Autonomous navigation and localization in service mobile robotics

We address the problem of autonomous navigation and localization in indoor environments, referring in particular to the specific scenario of service mobile robotics applications. The localization system uses active beacons (i.e. active transponders distributed throughout the building) as reference points; the estimate of the position of the robot and its uncertainty, both retrieved by correcting the estimate provided by odometry through an extended Kalman filter, are fed to the navigation system in order to help the robot to plan and execute target-oriented navigation tasks while showing a reactive behavior to handle the unpredictability of the environment.

Fusion of sensor data in a dynamic representation

The task of moving a robot in a complex and totally or partially unknown environment involves the construction of a suitable representation of the robots surroundings. This paper proposes an analogical statistical representation of the world the robot moves in. Its construction, updating and usage is described, focusing on the techniques used for the integration of sensor data and on their fusion with a priori information about the environment. An additional mechanism that allows the map to be efficiently transmitted to a remote workstation, connected using a radio link, has also been envisaged. Thus the computational load of map related processes can be distributed, taking advantage of the remote stations computational power. The experimental setup and some results obtained are also illustrated.

HEIR - A Non Hierarchical Hybrid Architecture for Intelligent Robots

In recent years different control architectures have been proposed for autonomous robots that are required to carry out service tasks robustly in civilian environments. In this paper a novel architecture is presented, which differs from the typical approaches from two main points of view: it is non hierarchical and it introduces an additional level of representation and reasoning based on image based, iconic descriptions. The paper will show how the proposed architecture has been implemented on a prototypical system, discuss its properties in detail and present our experimental experiences demonstrating the plausibility of the approach.