A Distributed Cloud Robotics Middleware Architecture for Trust-minimized Storage and Computation

We introduce a novel, distributed architecture utilizing consensual, blockchain-secured computation and verification that enables a scalable, transparent, and semantically interoperable cloud robotics middleware, capable of powering an emerging internet of robots in an automated way, without human intervention. Our architecture combines a set of modular key technologies that distinguishes it from existing, centralized robotics middleware approaches whilst enhancing their value. This includes technologies to (i) render data available across a distributed network; (ii) enact re-usability of data and applications, and their interoperability; (iii) universally execute data-manipulating applications yielding reproducible results, independently of the underlying infrastructure; and (iv) enable trust-minimized verification of derived data and the computations generating it. On an implementation level, the proposed middleware architecture integrates with ontologies, contentaddressable distributed storage, a blockchain database, and software containerization. In this paper, we motivate and justify the need for such a system and all underlying building blocks. Design and interplay of the components are explained in detail and embody the main contribution of our work. The benefits of our approach are highlighted by means of a sample application.