Axon: A Middleware for Robotics

Abstract

The area of multi-robot systems and frameworks has become, in recent years, a hot research area in the field of robotics. This is attributed to the great advances made in robotic hardware, software, and the diversity of robotic systems. The need to integrate different heterogeneous robotic components and systems has led to the birth of robotic middleware. A robotic middleware is an intricate piece of software that masks the heterogeneity of underlying components and provides high-level interfaces that enable developers to make efficient use of the components. A large number of robotic middleware programs exist today. Each one comes with its own design methodologies and complexities. Up to this moment, however, there exists no unified standard for robotic middleware. Moreover, many of the middleware in use today deal with low-level and hardware aspects. This adds unnecessary complexity in research involving robotic behavior, inter-robot collaboration, and other high-level experiments which do not require prior knowledge of low-level details. In addition, the notion of structured lightweight data transfer between robots is not emphasized in existing work. This dissertation tackles the robotic middleware problem from a different perspective. The aim of this work is to develop a robust middleware that is able to handle multiple robots and clients within a laboratory environment. In the proposed middleware, a high-level representation of robots in an environment is introduced. Also, this work introduces the notion of structured and efficient data exchange as an important issue in robotic middleware research. The middleware has been designed and developed using rigorous methodologies and leading edge technologies. Moreover, the middleware’s ability to integrate different types of robots in a seamless manner, as well as its ability to accommodate multiple robots and clients, has been tested and evaluated.