Nokia boosts Industry 4.0 with MX Grid, Visual Position and Object Detection

Nokia has unveiled technology to enable far edge compute and artificial intelligence (AI) to boost operational technology (OT) responsiveness and decision-making in advanced industrial applications, as well as new services to enable AI-powered applications for worker safety and industrial automation.

Nokia cited a recent report by research organisation Omdia highlighting the growing industrial trend of data pre-processing at the edge to improve productivity, efficiency and agility, which requires processing power located in or near the device, including machine learning (ML) and AI. Newly connected nodes with converged functionalities show the highest global growth rate, a CAGR of 11.3% from 2022–2027, driving industrial digitisation and IT/OT convergence.

As part of its response to these trends, Nokia has introduced MX Grid, what it claims is the world’s first on-premise, hyper-interconnected and distributed AI/ML offering that enables organisations to improve OT responsiveness and decision-making by processing and analysing data closest to the source.

MX Grid is designed to accelerate Industry 4.0 adoption, building on the Nokia MX Industrial Edge (MXIE) platform and using wireless connectivity, to offer a new OT data processing architecture that facilitates IT/OT convergence, supporting mission-critical industrial applications.

Compared with most current AI/ML assets that run in cloud environments, MX Grid brings OT-compliant AI/ML processing capabilities closer to the OT data source. It uses a pool of orchestrated compute-capable field devices, micro-edges, with a specialised, AI-capable software stack. These micro-edges are connected by private wireless networks and/or Wi-Fi.

Nokia said MX Grid can fundamentally enhance the efficiency of enterprise OT operations using decentralised workload processing and real-time, agile decision-making, bringing intelligence to legacy OT assets. Key industrial applications supported include predictive maintenance, security and surveillance, worker safety, tracking and positioning, and quality assurance.

In a quality assurance use case, the application on the micro-edge could analyse real-time sensor data and video feeds coming from the connected machine. Depending on the deviation level, either an immediate action can be triggered directly by the micro-edge or the MXIE peer application takes over the monitoring for deeper analysis of real-time data for later corrective action. This results in improved latency and optimised network load.

Omdia principal research analyst for manufacturing technology Anna Ahrens said: “We are observing a significant trend of decentralised processing in industrial technology moving compute and AI/ML capabilities to devices near the OT data source.

“In the upcoming years, we anticipate a steady increase in these connected edge nodes that are the true enablers of the industrial digitisation,” she said. “Nokia addresses these industry needs introducing a transformative decentralised AI/ML processing platform harnessing the power of private wireless connectivity and on-prem OT edge. This innovation facilitates emerging use case scenarios with low-latency requirements.”

The ability to integrate and use connected worker data and situational sensory information in MX Grid brings new capabilities for worker safety use cases. The Visual Position and Object Detection (VPOD) application is designed to enhance industrial tracking and positioning, and contextual awareness to enable Industry 4.0 applications and increase worker safety in industrial plants.

Explaining the rationale for its launch, Nokia said that in complex industrial environments it’s often difficult to equip machinery and tools with active tags, and challenging for people to wear and maintain those units (for example, battery change and faulty units).

It added that by using a real-time video data feed, VPOD eliminates the need to equip assets and people with powered devices, facilitating improved situational awareness.

Using Bell Labs’ patented technologies, VPOD traces industrial assets by using locally deployed AI algorithms that analyse real-time camera feeds to deliver insights into industrial operations to improve workplace conditions. VPOD utilises the MX Grid architecture to process video data next to the camera, together with MXIE, for more accurate asset tracking and positioning. This is said to improve significantly worker safety and contextual awareness, when combined with other OT data sources.