Modern automotive factories are becoming increasingly connected.
Machines are able to report telemetry, sensors can identify and track equipment, and vehicles pass through production phases and RFID, BLE, or UWB location systems. At the same time, MES , ERP, PLCs , and SCADA systems are used to track and control elements of the wider manufacturing process.
The real challenge with AIoT systems in industry is not simply the collection of data - but the aggregation and contextual analysis of information generated from disparate sources. How can AIoT provide new ways of connecting and interpreting the physical world?
Consider the following data flow in a typical automotive AIoT use case:
Physical Assets → Sensors → Connectivity → Data → AI/Analytics → Insights → Action
The physical asset layer can consist of things like machines and equipment, vehicles, tools, production equipment, materials, etc; and sensors and identification systems like RFID, BLE, UWB, can track:
Location, movement
Equipment status, temperature, vibration
Production events
Vehicle/VIN status, materials inventory, or utilization
These can be passed (or processed) through a connectivity layer before being used in higher-level manufacturing, analytics, or control systems.
Note that while data from sensors and identification systems is valuable, the most useful aspect of an AIoT architecture is the ability to connect diverse lines of information with contextual meaning.
Context is everything
Imagine a vehicle moving through the final stages of production.
The vehicle's location could be monitored with a location system.
MES or production software may indicate the vehicle's progress.
An inventory system may note if any particular parts are available.
Production equipment may report fault or operating status.
Taken individually, these data points may offer limited insight. Combined together, and through the use of analytics and AI, operations teams can observe patterns and correlations in the production environment. For example, repeated issues in production timing or performance could be evaluated in conjunction with equipment status, inventory status, vehicle movement, and production events in order to determine potential root causes.
This can enable a joined-up view of the production process to help manufacturers identify and resolve bottlenecks, or highlight areas for improvement.
OT and IT unification
A major challenge in industrial AIoT solutions is the integration between operational and information-technology systems. In a traditional manufacturing setup, there are two key classes of systems: operational technology (OT) and information technology (IT). This can include everything from PLCs or other industrial equipment (using OPC UA, MQTT or similar) in the factory, to manufacturing and enterprise management systems (MES, ERP, etc.) in back-end operations.
An effective AIoT system can provide a unifying view of the information from both sides of the manufacturing operation. Rather than treating OT/IT systems as isolated silos, an AIoT architecture can provide a way to unify:
Machines, sensors and location information, production events and MES systems, inventory management with ERP, analytics and AI, and more.
Edge-processing can also be leveraged as a way to perform local processing and filtering of data at the equipment or production line level before passing this on to wider manufacturing systems.
From data to action: AI and analytics add meaning
Connectivity and identification alone are not enough for AIoT in industry. Once information is captured, it must be processed and understood in order to deliver actionable insight.
For example, analytics and AI systems can help identify patterns in equipment performance (predictive maintenance), production bottlenecks (using analytics on equipment status, materials inventory, and production events), or material utilization (using similar methods). Once again, these examples highlight the importance of contextual data - the greater the insight into physical assets and the production process, the more that manufacturers can optimize their operations.
Automotive traceability
In many cases, car manufacturers need a deep understanding of the status and history of the vehicles they are manufacturing. AIoT can provide a way to build a traceable view by connecting together location, RFID or other identifiers, production events, and more. With the push towards connected cars and greater electrification in the automotive industry, this can allow manufacturers to understand the status of their vehicles in the production environment and beyond.
Companies looking to implement an AIoT solution in the automotive industry should look at OEMNex AI as a possible starting point for connected, data-driven manufacturing.
Conclusion
AIoT solutions are about connecting the physical world with data, context, and analytics. An effective system can bridge the gap between the manufacturing floor and executive decision-making. In automotive manufacturing, this means moving from data to insights to drive operations and decision-making. At the deepest level, this is about building systems that create value by aggregating and analyzing connected data.
Top comments (0)