Edge Computing Architectures for Enabling the Realisation of the Next Generation Robotic Systems

TL;DR

This paper reviews various edge computing architectures and platforms, highlighting their advantages and disadvantages, to guide the development of next-generation robotic systems leveraging edge and 5G technologies.

Contribution

It provides a comprehensive analysis of existing edge architectures and platforms, aiding robotic developers in selecting suitable solutions for their applications.

Findings

Different edge architectures offer varied capabilities and limitations.

Edge computing can significantly enhance robotic system performance with 5G integration.

Guidelines are provided for choosing appropriate edge platforms based on application needs.

Abstract

Edge Computing is a promising technology to provide new capabilities in technological fields that require instantaneous data processing. Researchers in areas such as machine and deep learning use extensively edge and cloud computing for their applications, mainly due to the significant computational and storage resources that they provide. Currently, Robotics is seeking to take advantage of these capabilities as well, and with the development of 5G networks, some existing limitations in the field can be overcome. In this context, it is important to know how to utilize the emerging edge architectures, what types of edge architectures and platforms exist today and which of them can and should be used based on each robotic application. In general, Edge platforms can be implemented and used differently, especially since there are several providers offering more or less the same set of services with some essential differences. Thus, this study addresses these discussions for those who work in the development of the next generation robotic systems and will help to understand the advantages and disadvantages of each edge computing architecture in order to choose wisely the right one for each application.