SO-MRS: A Multi-robot System Architecture Based on the SOA Paradigm and Ontology

Abstract

A generic architecture for a class of distributed robotic systems is presented. The architecture supports openness and heterogeneity, i.e. heterogeneous components may be joined and removed from the systems without affecting its basic functionality. The architecture is based on the paradigm of Service Oriented Architecture (SOA), and a generic representation (ontology) of the environment. A device (e.g. robot) is seen as a collection of its capabilities exposed as services. Generic protocols for publishing, discovering, arranging services are proposed for creating composite services that can accomplish complex tasks in an automatic way. Also generic protocols for execution of composite services are proposed along with simple protocols for monitoring the executions, and for recovery from failures. A software platform built on a multi-robot system (according to the proposed architecture) is a multi-agent system.

Human integration in an ontology-based IoT system

“…For a given task, Planner returns abstract plans that, when arranged and executed, may realize the final situation specified by the given task. More details can be found in [8].…”

“…As in the case of the CO2 example, the conflict of action itself is not as important, as ensuring the safety of a person in a potentially dangerous environment. Based on common ontology in the Autero system [8], services and tasks are described in a declarative manner, and each task is subject to the planning and arrangement process, which allows to assess the result of the service before its execution, and thus to determine whether its performance would violate the defined safety (comfort) rules.…”

A generic ontology and recovery protocols for Human-Robot Collaboration (HRC) systems

“…The simplicity should be viewed here as an advantage. Such ontology was proposed in Skarzynski et al (2018) [1] for adaptive multi-robot systems without humans.…”

“…The constrains, corresponding safeguards, and the recovery protocols can be defined on the basis of the proposed ontology. They naturally extend the SO-MRS architecture (Skarzynski et al (2018) [1]) where no humans were involved in task accomplishing.…”

“…Critical and hazardous (for humans) situations, as well as corresponding safeguards are defined on the basis of the ontology. The ontology is an extension of the ontology introduced in Skarzynski et al ( 2018) [1]. The architecture of SO-MRS [1], a software platform for automatic task accomplishment, is extended to HRC systems.…”