This paper describes the Player/Stage software tools applied to multi-robot, distributed-robot and sensor network systems. Player is a robot device server that provides network transparent robot control. Player seeks to constrain controller design as little as possible; it is device independent, non-locking and language- and style-neutral. Stage is a lightweight, highly configurable robot simulator that supports large populations. Player/Stage is a community Free Software project. Current usage of Player and Stage is reviewed, and some interesting research opportunities opened up by this infrastructure are identified.

Distributed Artificial Intelligence (DAI) has existed as a subfield of AI for less than two decades. DAI is

Successful distributed sensing and control requffe data to flow effectively between sensors, processors and actuators on single robots, in groups and across the Internet. We propose a mechanism for achieving this flow which we have found to be powerful and easy to use; we call it Player. Player combines an efficient message protocol with a simple device model. It is implemented as a multi-threaded TCP socket server that provides transparent network access to a collection of sensors and actuators, often comprising a robot. The socket abstraction enables platform and language independent control of these devices, allowing the system designer to use the best too1 for the task at hand. Player is freely available from http://robotics .usc. edu/player.

The swarm intelligence paradigm has proven to have very interesting properties such as robustness, flexibility and ability to solve complex problems exploiting parallelism and self-organization. Several robotics implementations of this paradigm confirm that these properties can be exploited for the control of a population of physically independent mobile robots. The work

Potential function approaches to robot navigation provide an elegant paradigm for expressing multiple constraints and goals in mobile robot navigation problems [9]. As an example, a simple reactive navigation strategy can be generated by combining repulsion from obstacles with attraction to a goal. Advantages of this approach can also be extended to multi-robot teams. In this paper we present a new class of potential functions for multiple robots that enables homogeneous largescale robot teams to arrange themselves in geometric formations while navigating to a goal location through an obstacle field. The approach is inspired by the way molecules "snap" into place as they form crystals; the robots are drawn to particular "attachment sites" positioned with respect to other robots. We refer to these potential functions as "social potentials" because they are constructed with respect to other agents. Initial results, generated in simulation, illustrate the viability of the approach. 1 Int...

Cooperative multi-agent systems are ones in which several agents attempt, through their interaction, to jointly solve tasks or to maximize utility. Due to the interactions among the agents, multi-agent problem complexity can rise rapidly with the number of agents or their behavioral sophistication. The challenge this presents to the task of programming solutions to multi-agent systems problems has spawned increasing interest in machine learning techniques to automate the search and optimization process. We provide a broad survey of the cooperative multi-agent learning literature. Previous surveys of this area have largely focused on issues common to specific subareas (for example, reinforcement learning or robotics). In this survey we attempt to draw from multi-agent learning work in a spectrum of areas, including reinforcement learning, evolutionary computation, game theory, complex systems, agent modeling, and robotics. We find that this broad view leads to a division of the work into two categories, each with its own special issues: applying a single learner to discover joint solutions to multi-agent problems (team learning), or using multiple simultaneous learners, often one per agent (concurrent learning). Additionally, we discuss direct and indirect communication in connection with learning, plus open issues in task decomposition, scalability, and adaptive dynamics. We conclude with a presentation of multi-agent learning problem domains, and a list of multi-agent learning resources. 1

We introduce a novel framework called "artificial physics", which provides distributed control of large collections of agents. The agents react to artificial forces that are motivated by natural physical laws. This framework provides an effective mechanism for achieving self-assembly, fault-tolerance, and self-repair. Examples are shown for various regular geometric configurations of agents. A further example demonstrates that self-assembly via distributed control can also perform distributed computation. 1.

Abstract. As research expands in multiagent intelligent systems, investigators need new tools for evaluating the artificial societies they study. It is impossible, for example, to correlate heterogeneity with performance in multiagent robotics without a quantitative metric of diversity. Currently diversity is evaluated on a bipolar scale with systems classified as either heterogeneous or homogeneous, depending on whether any of the agents differ. Unfortunately, this labeling doesn’t tell us much about the extent of diversity in heterogeneous teams. How can it be determined if one system is more or less diverse than another? Heterogeneity must be evaluated on a continuous scale to enable substantive comparisons between systems. To enable these types of comparisons, we introduce: (1) a continuous measure of robot behavioral difference, and (2) hierarchic social entropy, an application of Shannon’s information entropy metric to robotic groups that provides a continuous, quantitative measure of robot team diversity. The metric captures important components of the meaning of diversity, including the number and size of behavioral groups in a society and the extent to which agents differ. The utility of the metrics is demonstrated in the experimental evaluation of multirobot soccer and multirobot foraging teams.

Quantitative relationships between performance and behavioral diversity are investigated in a multirobot foraging task. The task, referred to as multi-foraging, requires robots to collect different types of object and deliver them to different locations according to type. Multi-foraging was selected for investigation because it offers even more opportunities for agent specialization than simpler foraging tasks. Three team foraging strategies are evaluated: homogeneous, where each agent is capable of delivering all types of object; specialize-by-color, where each robot specializes in collecting one type of object; and territorial, where most of the robots drop objects off near the delivery area, while the remaining agent completes the sorting and delivery. Each strategy is evaluated for diversity and performance using quantitative metrics. Data is gathered in thousands of simulation runs and the behaviors are also verified on mobile robots. In contrast to the results of a similar study ...

Many mobile robot tasks can be most efficiently solved when a group of robots is utilized. The type of organization, and the level of coordination and communication within a team of robots affects the type of tasks that can be solved. This paper examines the tradeoff of homogeneity versus heterogeneity in the control systems by allowing a team of robots to coevolve their high-level controllers given different levels of difficulty of the task. Our hypothesis is that simply increasing the difficulty of a task is not enough to induce a team of robots to create specialists. The key factor is not difficulty per se, but the number of skill sets necessary to successfully solve the task. As the number of skills needed increases, the more beneficial and necessary heterogeneity becomes. We demonstrate this in the task domain of herding, where one or more robots must herd another robot into a confined space. 1