A Distributed Semantic Knowledge Framework for Collaborative Robotics

Abstract

Collaborative work between robots are likely to be a keystone in Industry 4.0 which talks about large scale automation in industrial scenarios. The robots in warehouses, retail stores, manufacturing floors may be guided by a central system, but in unknown scenarios such as disaster management, or areas where connectivity may be intermittent, the robots should be able to coordinate the task(s) at hand among themselves without having to rely on a central coordinating entity. Further, there may be a likelihood of situations where the higher round-trip latency of cloud-based systems may be undesirable. In this work, we propose a distributed knowledge framework deployed modularly across multiple heterogeneous robots that enables them to perform tasks by real time knowledge sharing and collaboration. We also have performed extensive experiments to compare performance of such a distributed framework with a centralised one and reported the results here in detail.