A Control Agent Architecture for Cooperative Robotic Tasks

In Robotics a multi-robot approach is mandatory when the performance, robustness needed or functionality cannot be fulfil with only one robot. In this context, cooperation is a very important aspect to be taken into account because it allows that a set of autonomous robots to achieve the task by adding their skills and resources. A natural approach to accomplish a multi-robot task is by decomposing it into cooperative actions, each one executed by a group of the robots where every robot takes a well defined role. To have an intentional cooperation level working with a Multi-Robot System (MRS), explicit mechanisms and control architectures have to be defined. In order to analyse or design a multi-robot system, it can be viewed as a Multi-Agent System (MAS) composed of physical agents. Many aspects have influence in the design of these complex systems in robotics: task decomposition, task allocation, role assignment, inter-agent interference and competition, agent cooperation, coordination, conflict resolution, negotiation and inter-agent communication. Cooperation in the context of MAS has emerged to provide better use and performance of the agents and their capabilities. Following the approach proposed by Ferber (Ferber, 1999), cooperation can be seen as the conjunction of three components: Collaboration: it is centered in task allocation. In order to assign which agent has to accomplish a specialized task, interaction protocols can be used. These dialogues allow to take into account not only the capabilities of the agents, but also their availability. Coordination: it deals with the synchronization and planning issues. For this, it is necessary to determine at what time an agent should perform a task. The global team performance depends on providing good timing to each agent. Conflict Resolution: usually agents share resources in a concurrent way, which easily can produce conflicts and even dead locks; thus it is necessary to incorporate mechanisms to prevent, avoid or solve conflicts. Normally, these three components are obtained by the use of interaction protocols that define well structured dialogues between the cooperative agents. In just one sentence, cooperation is achieved by using collaboration, coordination and conflict resolution mechanisms supported by structured communications between agents. In order to achieve cooperation, different architectures and techniques have been proposed, as shown in section 2. In this chapter, the Multi-Resolution Cooperation Control (MRCC)