The Industrial AI and Robotics department develops applied methods and deployable prototypes combining artificial intelligence, human-centred workflows and robotic or automation integration. Its work prioritises robustness, traceability and practical validation with industrial stakeholders.

Mission

Translate research into components and validated workflows for AI-enabled industrial operations. Decision support, human-in-the-loop systems and transferable integration patterns help connect experimentation with use across industrial sites.

Core focus areas

  • Industrial computer vision: inspection, classification, robustness to changing conditions, limited-labelled-data strategies and operator-oriented explainability.
  • Human-in-the-loop decision support: confidence-aware recommendations, escalation, approval and override flows, and feedback capture.
  • Evidence and traceability: structured recording of inputs, decisions and outcomes for auditability, quality assurance and reproducible validation.
  • Edge AI: containerised inference, performance and latency constraints, and resilient operation under variable conditions.
  • Robotics and automation: modular workcell interfaces, safe interaction and integration architectures, including ROS 2 where appropriate.
  • Operator experience: interfaces that account for time pressure, variability and cognitive load.

Applications and methods

Applications include visual inspection and defect triage, quality decision support, rework guidance, inspection-to-action workflows, process-improvement data pipelines and operator-guided robotics. Methods include AI benchmarking, dataset and taxonomy design, labelling guidelines, validation protocols, software-stack integration and realistic performance testing. Security and data protection are considered from the design stage.

Collaboration

The department works with SMEs, integrators, industrial end-users and research partners through working groups, technical briefings, pilot co-development and validation of measurable operational outcomes. Non-confidential results and good practices are shared with the wider community. Connections with AI, interaction, cybersecurity and advanced computing support interdisciplinary delivery.