The EdgeBrain™ inline AI module from de Jonge GmbH allows existing industrial systems to be quickly expanded with AI functionality without the need for extensive modifications. The updatable plug-and-play solution is primarily aimed at manufacturing companies, machine builders and system integrators. The EdgeBrain™ module is integrated directly into the data path between sensors, controllers and machines. There, it analyses and evaluates data in real time using AI. Data is not first sent to a cloud or a separate industrial PC, but is processed immediately and locally in the module as it flows through the system in the ongoing data stream. This allows immediate, deterministic decisions to be made without additional delays. The module reads the current signal data, evaluates it using a trained AI model and returns the result to the control system in real time. The existing plant structure remains unchanged. This distinguishes the module from other AI approaches such as cloud AI, where data is first transmitted, analysed externally and returned with a delay, or industrial PCs, which require a high level of integration.
The EdgeBrain™ module detects quality deviations during production as well as machine errors and anomalies. Other features include the fusion of sensors and consolidation of signals, as well as real-time optimisation of process parameters. The key advantage is that existing machines can be retrofitted with AI functionality without modification or downtime. This gives companies immediate decisions in real time, stable latencies for real-time systems, local data processing with high data sovereignty, and lower integration costs than with traditional AI architectures.
The EdgeBrain™ module addresses typical challenges in machine parks, such as a lack of transparency in the process, delayed error detection and complex integration projects. The offline-capable inline module makes it possible to modernise existing lines without downtime and to feed data-based decisions directly back into the control system. It combines several concrete benefits for the user. Anomalies and quality deviations are detected directly in the data stream before rejects occur. Deterministic real-time inference ensures reproducible decisions without additional latency. Plug-and-play integration without new PLC structures or GPU servers ensures minimal integration effort. Data sovereignty remains within the company, as the data is optionally processed locally with highly secure connectivity without cloud dependency. The modernisation of machines and lines takes place without interrupting production during ongoing operations. This makes EdgeBrain particularly suitable for manufacturing companies with high quality requirements, retrofit projects or heterogeneous existing plants.
The target groups for EdgeBrain™ are production companies with existing machine parks (brownfield) as well as machine and plant manufacturers who want to offer retrofit-capable AI functionality. Other potential customers include OT and automation managers who want to use real-time analytics without the cloud and system integrators who need standardised, repeatable AI solutions. Typical areas of application include industrial series production, process industries such as chemicals, pharmaceuticals and food, as well as high-throughput environments where stable latencies and robust systems are crucial.
The EdgeBrain™ module is designed for harsh industrial environments. It is vibration and shock resistant, EMC/EMI robust and resistant to oils, coolants and cleaning agents. The compact, all-aluminium housing with passive cooling allows it to be used in a working temperature range from -40 °C to +85 °C. Power consumption is max. 5 watts. The low-power FPGA SoC architecture enables energy-efficient continuous operation with high power density. Local data processing significantly reduces bandwidth requirements, computing load and system complexity, which is a decisive advantage for scalable Industry 4.0 applications. With dimensions of 90 mm × 75 mm × 25.5 mm and a weight of 300 g, the module can be integrated into existing systems in a space-saving manner without additional industrial PCs or control cabinets. Unlike cloud-centric solutions, EdgeBrain™ brings deterministic intelligence directly into the data path of the machines and operates entirely locally. A highly secure, bidirectional connection can be activated as an option.
At the Embedded World Exhibition & Conference from 10 to 12 March 2026 in Nuremberg, de Jonge GmbH will present its EdgeBrain™ inline AI module at the TRS-STAR GmbH stand in Hall 4, Stand 4-467.