Survey on Cloud Robotics Architecture and Model-Driven Reference Architecture for Decentralized Multicloud Heterogeneous-Robotics Platform
Akila Siriweera, Keitaro Naruse
Abstract
Robotics engineering is gradually becoming an essential part of our daily life. However, it has been generating Big Data and seeking large computation cost because of the diversified sensors and processing requirements involved in heterogeneous robotics and their workflows. Therefore, cloud computing has become the incumbent platform for robotics. There are numerous works related to the architecture for cloud robotics. However, most cloud robotics architectures are ad hoc and are not based on a model. Inherent drawbacks of ad hoc approaches include being strictly domain specific and minimally customizable and adaptable. Moreover, heterogeneous cloud robotics platforms have been operating diverse requirements of industries and households. Nevertheless, there are certain benchmarks set to be achieved by Industry 4.0 and norms by Society 5.0. Those benchmarks and norms lead to new products and services in cross industries and alleviate the impending drawbacks. However, those should be achieved while retaining both the sovereignty and security of the respective systems and industries. This is equally applicable and an enormous challenge to the system-of-systems involving the cloud robotics domain. Therefore, we surveyed cloud robotics architectures. Then, we learned a top-down design approach involving a unified architectural framework as the cognitive approach for the highly variable and systematically complex challenges to be achieved in the next-generation cloud robotics domain. Reference architecture is a well-known approach for instantiating top-down unified architectural framework processes. Therefore, we proposed an architectural design process and modeling for the reference architecture for next-generation cloud robotics platforms.