Cooperative control is a key issue for multirobot systems in many practical applications. In this paper, we address the problem of coordinating a set of mobile robots in the RoboCup soccer middle-size league. We show how the coordination problem that we face can be cast as a specific coalition formation problem, and we propose a distributed algorithm to efficiently solve it. Our approach is based on the distributed computation of a measure of satisfaction (called Agent Satisfaction) that each agent computes for each task. We detail how each agent computes the Agent Satisfaction by acquiring sensor perceptions through an omnidirectional vision system, extracting aggregated information from the acquired perception, and integrating such information with that communicated by the teammates. We empirically validate our approach in a simulated scenario and within RoboCup competitions. The experiments in the simulated scenario allow us to analyse the behaviour of the algorithm in different situations, while the use of the algorithm in real competitions validates the applicability of our approach to robotic platforms involved in a dynamic and complex scenario.

Keywords: Coordination, Multirobot system, RoboCup, Task assignment

In the last decade the development of multirobot systems has shown with growing evidence how a well-balanced deliberative-reactive coordination is the key issue to provide the group with an efficient and robust collective behavior. Usually, but not necessarily, deliberation is linked to the planning dynamical role exchange whereas reactivity is associated with the ability of the system to stress any condition which could fertilize cooperation. Using a chemical metaphor we could say that reactive cooperation should emerge from the interaction between individiduals through an appropriate catalyst whose availability increases as cooperation increases. In the paper we present an attempt to satisfy the constraint by maintaining at the subsymbolic level, as long as it is possible, all the relevant information to coordinate a multirobot system. To this aim we model the group of robots using the roboticle framework and we assume a thermal metaphor where temperature and conductivity are the quantities which trigger the individual dynamics to make emerging the collective behavior of the group.

Keywords: behaviour-based control, thermal diffusion metaphor, roboticle, robot colony

In the last decades the theory of the complex dynamical systemshas come to maturation providing a lot of important results in the field ofmany applied sciences. Also robotics has taken advantages from this new approach in what the behavior-based paradigm is particularly suitable to devise specific sensing activity since sensors usually provide information about the environment in a form which depends on the physics of the interaction. It is not required to be immediately converted into some symbolic representation but, on the contrary, it can be maintained at some physical level as a metaphor of the events observed in the environment. The close connection between the motor schema with its companion perceptual schema seems suggesting the presence of a substratum which underlies both perception and action activities, driving the flow of information accordingly. In the paper we consider a colony of robots immersed in a well-specified thermodinamical substratum where enthalpy and heat flux are devised to go vern the diffusion/merging behavior of a swarm.

Keywords:

Autonomous robots, like living systems, must be adaptive in nature if we want them to preserve their integrity while completing their mission. The challenge to survive in their environment is better accomplished if they are open systems, interacting with the environment by exchanging matter, energy, information, and so on. The roboticle framework, presented here forth, is an attempt to model how the autonomous robot control unit works. It borrows from living systems the idea that sensing and acting on the environment can be recognized as a mechanism exchanging energy with the environment in order to maintain an highly organized internal control structure to resist to external applied perturbations. The necessary energy balancing is provided by an autopoietic loop which is fed by the energy entering the robot through its sensor devices and it is dissipated by its effectors for properly acting in the environment. The autopoietic loop is also responsible of the adaptive properties of the robot.

Keywords: Autonomous robots, Autopoietic loop, Roboticle framework

In this paper, we suggest an hybrid architecture where the deliberative part takes advantages from the reactive one and vice versa, to make a multi-robot system to exhibit some assigned cooperative task. We explain our architecture in terms of schemas and a set of firing conditions. To experiment our approach, we have realized an implementation that tries to exploit the resources of our robot team participating to Middle-size RoboCup tournaments. Each individual exhibits both reactive and deliberative behaviors which are needed to perform cooperative tasks. To this aim we have designed each robot to become aware of distinguishing configuration patterns in the environment by evaluating descriptive conditions as macroparameters. They are implemented at reactive level, whereas the deliberative level is responsible of a dynamic role assignment among teammates on the basis of the knowledge about the best behavior the team could perform. This approach was successfully tessted during the Middle-size Challenge Competition held in Padua on last RobCup2003.

Keywords:

This paper considers the properties a multirobot system should exhibit to perform an assigned task cooperatively. Our experiments regard specifically the domain of RoboCup middle-size league (MSL) competitions. But the illustrated techniques can be usefully applied also to other service robotics fields like, for example, videosurveillance. Two issues are addressed in the paper. The former refers to the problem of dynamic role assignment in a team of robots. The latter concerns the problem of sharing the sensory information to cooperatively track moving objects. Both these problems have been extensively investigated over the past years by the MSL robot teams. In our paper, each individual robot has been designed to become reactively aware of the environment configuration. In addition, a dynamic role assignment policy among teammates is activated, based on the knowledge about the best behavior that the team is able to acquire through the shared sensorial information. We present the successful performance of the Artisti Veneti robot team at the MSL Challenge competitions of RoboCup-2003 to show the effectiveness of our proposed hybrid architecture, as well as some tests run in laboratory to validate the omnidirectional distributed vision system which allows us to share the information gathered by the omnidirectional cameras of our robots.

Keywords: Cooperative behaviors, distributed sensing, distributed vision system, multirobot systems, Robocup middle-size league competitions

Autonomous robots, like most biological systems, use locomotion as very distinguishing property to dynamically interact and learn about the environment. In the biological context this activity implies some form of intelligence suggesting a similar approach for artificial systems. Recently Asama has coined the term mobiligence referring to the intelligence which emerges through the interaction between an agent and its environment due to its mobility. This kind of intelligence requires abduction, the property by which autonomous agents can import environmental information inside them. The roboticle framework provides a similar property in what it supplies the internal representation of the sensed world by dealing with sensor information as perceptual energy. It feeds the agent governor's unit which delivers effector commands triggered by the so called autopoietic loop. In this perspective, previously assimilated perception is converted into effort by adjusting the direct and inverse gains of the autopoietic loop, balanced by mobility itself.

Keywords:

The aim of this paper is to investigate how stigmergic information allow each individual of a group of autonomous robots to take advantages from other individual behaviors. The proposed analysis is based on the roboticle model where sensor data and effector commands are treated as energy exchange between the robot and its environment, eventually populated by other robots. Without explicit communication, the collective behavior of a group of teammates can be forced only if the robot designer makes each robot to become aware of distinguishing configuration patterns in the environment. Usually, the job is accomplished both by evaluating descriptive conditions as macroparameters and an appropriate dynamic role assignment among teammates. Since observed individual behaviors can affect the normal course of operations for each robot propagating to other teammates, we want to address some issues on how a collective behavior is fired and maintained.

Keywords:

We investigate the problem of how to make a multi-robot system performing a cooperative task by inducing a set of emergent actions.We model the environment dynamics by considering some parameters that express the ability of each robot to perform its task. Thus, the members of a group of robots become aware of their ability to realize some tasks by simply computing some quality function Q of the cofiguration pattern of the environment.A role assignment schema allows roles to be swapped among the robots of the group in order to select the best behaviors able to performthe task cooperatively. We illustrate this approach by showing how two soccer robots were able to exchange a ball, during a real game, by combining the use of efficient collision avoidance algorithms with role swapping triggered by the value of the above quality function Q.

Keywords: Multi-robot, emergent behavior engineering, RoboCup

The robotic model considered hereafter is an attempt to deal with the problem of mobility for simplified holonomic mobile robots, such as Braitenberg vehicles. The aim is to find out which behavioural constraints are needed to make emerging a collective behaviour in the form of "mobility task". In our model "situatedness" and "embodiment" are defined as in the traditional behaviour-based approaches. The situated body of such a simplified robot is reach enough to be more than a merely symbol manipulating process. The model provides sensor data and effector commands with the same meaning of energy exchange between the robot and its environment, eventually populated by other robots. When a group of robots want to cooperate for a common goal it uses "stigmergy" to allow each individual to take advantage from other individual behaviours. Stigmergy detection is evaluated by the so called "macroparameters" and triggered by a dynamical assignment of "roles", which can force a collective behaviour. The concept of "perceptual perturbation" introduced in our robotic model is a useful abstract tool to figure out collective mobility tasks. This property stems from the "autopoietic loop" shown by each individual and which, under specified conditions, can interact with other ones to form one or more autopoietic loops involving two or more individuals. Collective behaviours eventually emerge as a consequence of a non-linear interaction among behaviours.

Keywords:

The design of an autonomous robot should start from the available hardware where microelectronics is just a part of a more complex architecture. The complementary roles of electrical and mechanical devices are such that the latter cannot be ignored at all, as it has usually made in the world of computers, because the interactions between the physical components of the robot are not neglectable. This is why terms such as embodiment or situatedness are so meaningful during the design phase of its governor's unit: its behaviour cannot be controlled as if the robot were a merely symbol manipulating process. The physics of the world is reflected on the way the interation between the component parts propagates to and from the environment.

The development of autonomous robots requires a continuous demand for a deeper understanding of how living systems operate with respect of their internal structure and their interaction with the environment. The aim of this paper is to show that the behaviour of an autonomus robot can be designed within a framework which refers to some general quantities, borrowed from thermodynamics, such as energy and entropy. The general mechanism we have devised is that, like any living system, an autonomous robot operates in a stationary state by triggering the flow of energy, supplied by its sensors and dissipated by its effectors.

The aim of this paper is to show that the behaviour of an autonomus robot, either in isolation or as a part of a group, can be described within the language of complex dynamical system theory. In a framework which deals with entities built on several component parts and specified by one or more quantities changing over time, we have introduced some general quantities borrowed from thermodynamics, such as internal energy and temperature. By so doing, we have devised a mechanism by which an autonomous robot, like any living system, operates in a stationary state by triggering the flow of energy, supplied by its sensors and dissipated by its effectors. Moreover, the implicit coordination of a group of autonomous robots, which partecipate to a collective action, can be described as a self-organizing mechanism where the energy of the group plays the role of cue-based communication. The way it flows among the group components is the consequence of the entropy decrement with the respect of that ascribed to the environment.

The challenge to provide a qualitative depiction of autonomous robot behaviour, performing either individual or collective tasks, has suggested us to consider robots as open systems that receive a stream of input sensor signals from and generate a stream of output actions to the environment. Within the framework of complex dynamical systems we have defined the "energy", a robot is fed with, as the amount of sensor information acquired from the environment which is later dissipated by the action of its effectors. Moreover, the dynamical law can be interpreted as a mechanism which controls robot operations by triggering the flow of energy between sensors and effectors, whose current value can be considered, in some sense, the internal world representation handled by the agent.

The behaviour-based approach deals with robot operations, inside an unstructured and partially unknown environment, as they were accomplished thinking at the robot as a collection of interacting parts, its behaviours, competing for common resources, its sensors and actuators. The challenge to provide a qualitative depiction of autonomous robot behaviour, performing either individual or collective tasks, has suggested us to consider robots as open systems that receive a stream of input sensor signals from and generate a stream of output actions to the environment. Using the framework of complex dynamical systems we have introduced the concepts of "energy" and "entropy" with the aim to make emerging a "collective behaviour" from a robot team as if it were caused by specific patterns of self-organizing groups.

Since its origins, Artificial Intelligence has been faced with the challenge to control robot operations through the so called deliberative thinking paradigm. Robot actions are governed by a reasoning process which needs robots to acquire information from the environment to update their internal world model causing the failure to generate an useful action in a finite amount of time. The framework of roboticles, appearing in this paper and borrowed from the theory of complex dynamical system, is a tool to deal with quantities like "energy" or "effort", to symbolize the amount of sensor information a robot is fed with, later dissipated by the action of its effectors. The dynamical law works as a triggering mechanism which controls the flow of energy between sensors and effectors so that its current value can be interpreted, in some sense, as the internal world model handled by the agent. Environment changing, detected through sensor signals, results in moving the representation point of the system on the energy surface. Moreover, actions issued by robot effectors dissipate energy in a way to maintain the working point of the system in a stationary state where the energy supplied by sensor signals is balanced by the effort delivered to the effectors.

We illustrate the Cooperation through Implicit Communication behavior-based approach used for developing the PaSo-Team The University of Padua Simulated Soccer Robot Team, a multi-robot software system for soccer robot competitions promoted by the RoboCup Simulation League. The configuration of the environment, namely the robots' relative positions depending on both the global task and the game dynamics, provides a source of implicit information about the robots' intention to be involved in collective actions, making them able to cooperate implicitly. The soccer team performance can be tuned by triggering the arbitration module of any single robot to generate, as many as possible, suitable situations which hint to the team the action of scoring the goal. Some macroscopic parameters can be usefully introduced to evaluate the evolution of the whole multi-robot software system.

We discuss how to plan the actions of a multirobot system acting in the real world by using task constraints for implicit coordination. Classifying and modelling task constraints is shown to be a very powerful mechanism for giving each robot an extended capability of coordinating itself with the other robots of the whole system. Thus, we investigate the problem of performing collective tasks with emphasis on the use of task constraints to get intelligent coordination among the individual members of the robot group. We have also introduced a set of macroscopic parameters that allow us to evaluate the evolution of the whole system. They depend on both the global task, and the environment dynamics, allowing the robots to implicitly communicate their intention to be involved in complex actions. Other group members become aware of their task simply by recognizing configuration patterns of the environment. We illustrate some examples of such a coordination taken from the scenario of both service robotics and of simulated robot games.

The challenge to implement a robot team for participating to a simulated soccer competition has motivated us to develop a multiagent architecture where single agents have a set of selected behaviors triggered by an arbitration module and use implicit communication to give rise to implicit coordination. Every agent is provided with a set of states, partly referred to its own acting (local flag) and partly referred to its team acting (global flags). The latter can issue actions such that, modifying the environment, implicitly inform about its intention to be involved into the acting of another agent. This is performed simply by making an agent to be aware of a pattern that it can be easily recognized by looking at the environment. The soccer-team performance can be tuned by triggering the arbitration module of any single agent to generate, as many as possible, suitable situations which hint the team the action of scoring a goal.

We illustrate PaSo-Team (The University of Padua Simulated Robot Soccer Team), a Multi-Agent System able to play soccer game for participating to the Simulator League of RoboCup competition. PaSo-Team looks like a partially reactive system built upon a number of specialized behaviors, just designed for a soccer play game and generating actions accordingly with environmental changes. A general description of the architecture and a guideline of main ideas is presented in the paper, whereas a more detailed description of actual implementation is given in the appendix.

Classical planners, which automatically generate plans to be executed as a sequence of primitive actions cannot follow their actual execution because they cannot compare and modify its behaviour continuously using sensor data flow. On the other hand, autonomous systems need to interact with the environment because they have an incomplete knowledge about it. So, the planner doesn't pretend to generate explicitly all the actions. In this paper we propose an alter- native way to design and build an autonomous system introducing the metaphor of biochemical machine. We think of the whole system as a family of behaviours, each implementing a set of actions which specify, at each step, the most appropriate response to a perceptual issued stimulus. Moreover, each behaviour is equipped with appropriate receptors which can cause its inhibition, or suspension, if it is not currently requested. Later, the behaviours can be resumed. Such an inhibition schema is directly driven by the system itself using sensor data and the knowledge it has about its state.

The aim of this paper is to outline a planning system architecture which allows robots to exhibit varying degrees of autonomous behaviour. While several systems have been developed to cope with specific classes of robot tasks, a litte effort has been made towards the autonomy itself. Looking at the behaviour of animals from the ethological point of view we can suppose that even robots need to exibit a wide variety of specific behaviours. Starting from Brooks and Rosenschein's approach we can think of an autonomous system as a vertical composition of its basic behaviours, or "instincts", to produce the overall "emergent activity". The key point, however, is how to really obtain it considering that robot actions require to be planned in some way to complete their mission. In this paper we propose an analternative way to design and build an autonomous system introducing the metaphor of a chemical machine. We th ink of the whole system as a set of behaviours, each implementing a specific response to incoming environmental stimuli, equipped with appropriate receptors which can be inhibited if a behaviour is not currently requested. Such an inhibitor schema is directly driven by the system itself using sensor data and the knowledge it has about its state. The advantage of this robot design lies in its ability to make explicit the adaptive capabilities of the system during its implementation.