Communication Architecture for Robotic Applications

In robotic applications, the communication protocol is one of the main features that dictate scalability and network topology. The communication architecture must consider heterogeneity (i.e. network, hardware, operational system and programming language) and provide programming abstraction to simplify its development. There are several middlewares and frameworks that can be applied in robotic applications, differing considerably in complexity, programming languages and approach. In this context, this paper presents a communication architecture that fulfils such requirements and ensures information exchange through the network. It was evolved from a previous study, providing more flexibility and easily to adapt to other applications. An interface definition language (IDL) was conceived that enables users to define and deploy services, and also adjustable constraints (service request timeout, message size, maximum number of connected nodes) that restrict provided functionalities. The middleware is based on a multi-threaded service-oriented hybrid peer-to-peer architecture that uses concepts of object-oriented programming in a layered structure to provide flexibility for the communication implementation, minimizing code changes when ported to other robotic systems. Tests of availability and network response time were performed to evaluate its time constraints. The middleware applicability was proven when implemented in an AGV distributed system, designed to operate an intelligent warehouse.