RUBICON - Control Layer Architecture

This deliverable (D3.1) takes place at the end of the first task (T3.1 - Functional Design and Specifications) of WP3 - the RUBICON Control Layer. The goal of the Control Layer is to exercise high level control over the sensing/acting/communication capabilities of the RUBICON system. This control focuses on devising suitable action and configuration strategies while exploiting the RUBICON Learning Layer (WP2) to adapt these strategies to the environment and improve their quality over time. Together with D1.1, D2.1, and D4.1, this deliverable presents a set of requirements and specifications to support the development of the RUBICON system . The requirements reported here provide a collection of statements to inform research directions for the RUBICON project. They are based on several inputs: the Description of Work (DoW) document, a Closed Workshop (June 20-23, OÌ^rebro), the case studies described in Section 2.1 (which combine contributions from all the RUBICON partners), the requirements described in deliverables D2.1 and D4.1, as well as input from several informal discussions among the project consortium. These inputs are used to carry out domain analysis, leading to the identification of the requirements for the Control Layer. After that, this document examines the state of the art in control solutions applicable to robotic ecologies in order to provide a first approximation of the design of the control layer and its most important interactions with the other layers of the RUBICON architecture. Finally, this deliverable provides the specifications of a test-plan to be used for the development and the evaluation of the various releases of the RUBICON middleware.