The BRAIN-IoT technical objectives
The BRAIN-IoT vision is realized through seven Technical Objectives (TOs), as described in the following:
BRAIN-IoT approach to interoperability is based on the adoption of shared semantic models, dynamically linked to concrete IoT devices (sensors, actuators, controls, etc.) operating autonomously in complex scenarios. Binding of models to concrete implementations leverages mapping to open industry standards for semantic device description.
Building upon shared models (TO1) BRAIN-IoT facilitates the deployment of smart cooperative behaviour, realized by means of modular AI/ML features which can be dynamically deployed to heterogeneous IoT devices in mixed edge/cloud IoT environments. Smart behaviour features are enriched by distributed data processing, federated learning, virtualization/aggregation of data/events/objects, resolution of mixed-criticality situations and conflicts, verification and context-aware self-adaptation of connectivity and real-time event-oriented, reactive approaches.
This is achieved by leveraging fully de-centralized peer-to-peer approaches providing linkage between modular, ad-hoc IoT self-hosted and cloud-based services through existing open standards.
BRAIN-IoT introduces a holistic end-to-end trust framework for IoT platforms suitable to be employed in scenarios characterized by strict security and safety requirements, associated with actuation and semi-autonomous operations, and by special needs for secure identification, authentication of data and devices, encryption, non-deniability, as well as detection of cyber-attacks and protection against them. This is done by adopting established security protocols, joint with distributed security approaches derived by peer-to-peer systems e.g. block-chain.
BRAIN-IoT develops new patterns for interaction between users and IoT solutions, leveraging semantic mapping of privacy requirements towards data and service models in use in each specific use case, introducing privacy-related APIs and models. This enables the possibility to programmatically inform users about privacy policies in place, as well as enabling them to exercise fine-grained privacy controls.
Building upon shared models (TO1), BRAIN-IoT provides tools to ease rapid prototyping (development, integration) of smart cooperative IoT systems. This is achieved by extending available tools for development, integration, commissioning and management of IoT and Cyber-Physical systems.
Building upon its federation capabilities (TO3), BRAIN-IoT enables end-users to dynamically commission and reconfigure their modular IoT instances, choosing among the available platforms, modules implementations and services. This is achieved by extending existing open marketplace of IoT services and data jointly with available catalogues providing open IoT enablers, and integrating them with its federation framework.