IoT Core – Industrial Edge Intelligence Platform
Fog/ Edge Real-time Analytics and Machine Learning
The IoT Core enables a broad range of industrial IoT applications providing new levels of autonomy, resilience, data security and lightning fast analytics on-the-spot.
Despite its broadly recognized power, the applicability of Cloud Computing for time-critical operations is highly limited. Especially in manufacturing, automotive or telemedicine scenarios, where latencies caused by the roundtrip to cloud servers can have fatal consequences. Real-time intelligence at the fog/edge level is required.
Fog/ Edge intelligence also provides new levels of resilience and autonomy, as poor connectivity or network outages do not affect local, mission-critical control loops. Wireless Fog-to-Fog communication mechanisms allow for autonomous device-to-device communication which, unlike cloud-based communication still functions during network performance degradation.
By aggregating and analyzing data locally fog/edge-based data processing yields several data-security related advantages compared to traditional cloud-computing mechanisms. Not only can communication and exposition of sensitive data to external data centers be fully omitted, but also finely configured, which data is being sent to external parties, how data is being filtered, anonymized and encrypted and how the communication channels between local and external infrastructures are being secured.
Among several other advantages, fog-computing allows for significant network capacity optimization, as less bandwidth is utilized on shop-floor-, enterprise-network and internet uplinks, significantly reducing bottlenecks and enabling fewer network congestions.
Depending on the number of networked sensor devices and controlled actuators, the frequency and format of data exchanged (sensor on video data) and on the complexity of analytic computations, the IoT Core can be deployed on embedded mini-computers, industrial gateways and edge nodes, micro-datacenters as well as Cloud backends.
As topologies and heterogeneity of distributed fog-/ edge-/ cloud-computing infrastructures can vary considerably, in many cases intelligent application orchestration systems are required, capable of dynamically deploying, managing and elastically scaling the compute infrastructure.
Edge Intelligence Use Cases
With edge intelligence a plethora of use cases from automotive to smart manufacturing, telehealth and smart city safety are enabled. The following few example applications realized within the Transfer Center IoT show the power of fog-/ edge-based data processing and local actuation.
Edge Intelligence enables Public Safety & Privacy
Through intelligent, distributed, autonomous Fog-based (video) data analysis and public surveillance systems are enabled to autonomously analyze public surveillance video data and trigger safety and defense actions without violating data protection and privacy regulations. Thereby an entirely new generation of public surveillance and safety services can readily be rolled out. By enforcing strict video data protection rules at the fog/edge level, the overall system complies with German privacy regulations. Through Fog-to-Fog communication mechanisms ultra-low latency levels are achieved. Highest levels of autonomy, resilience and robustness are achieved by local data processing and local actuator control.
Edge Intelligence enables Industry 4.0
“Retrofitting” Industry 3.0 Shop Floor Assets allows for the monitoring of equipment performance and factory processes thus enabling the “digital twin”. Combined with hard real-time networking (TSN), hard & soft PLC interworking and fast stream analytics the IoT Core provides cost-effective means for monitoring and increasing Overall Equipment Effectiveness (OEE), enables Condition Monitoring & Predictive Maintenance as well as industrial Safety applications. Factory integration involves
- Deployment of IIoT Core Gateways in the field,
- monitoring and control through OPC UA based PLCs,
- Mounting of e.g. Micro-Electro-Mechanical Systems (MEMS) and many other sensors for aggregation of missing asset/ equipment data
- Local, real-time data analytics for rapid and direct asset control (safety, hazards, etc.) exploiting Machine Learning techniques
Mission-Critical Intelligence on-the-Spot
Combining fog-/ edge-/ cloud computing mechanisms with advanced real-time stream analytics and machine learning techniques enables a broad range of mission-critical industrial IoT applications. By providing advanced M2M connectivity through latest low-power wide area networks (LPWAN) and deterministic time-sensitive networking (TSN) technologies, edge intelligence is brought to smart cities in just the same way as to Industry 4.0 environments.
Industrial IoT Standards, M2M Protocols and APIs (OPC UA, OneM2M) are provided to assure seamless integration and interoperability. Time-Critical Stream Analytics is enabled by latest Complex Event Processing (CEP), Machine Learning and Video Analytics modules, deployed on COTS fog-/ edge- as well as cloud-hardware infrastructures. Stream analytics algorithms (CEP, classification and regression models, decision trees, clustering, statistical analysis), as well as video analytics functions (object, motion detection, tracking, counting & diagnosis, 3D video analysis) enable a broad range of use cases. The IoT Core is provided as a module of the OpenIoTFog Industry 4.0 toolkit on www.openiotfog.org.
- IoT Core Solution: Hardware selection, Deployment, Integration and Licensing
- IoT Testbeds: Feasibility-, Stress- Safety-, Security-Testing and Quality Assurance
- CPS-Enablement and Retrofitting: Mounting and interconnecting industrial sensors, drives and controllers with subsequent real-time data aggregation and controller logic programming
- Predictive Analytics & Artificial Intelligence: Modelling, Calibration and Application of advanced AI and Machine Learning Techniques for condition monitoring, predictive maintenance and safety applications
- IoT and Data Analytics Training, Consultation & Knowledge Transfer