MTC Innovation Accelerator

Using AI-powered cameras, sensors and measurement tools to reliably catch defects and verify quality in real environments, even when lighting, part shapes, surfaces and tolerances vary from one unit to the next. Such as:

• Machine vision and other sensing technologies (3D scanning, tactile sensing, multi-sensor fusion) that inspect parts as they're made
• Systems that keep working reliably despite everyday variation: different lighting conditions, part geometries, surface finishes and defect types
• Quality checks built directly into the production line, rather than performed afterwards by hand
• AI models that can be adapted to new products or defect types without being rebuilt from scratch

Robots, automation and control systems that can physically handle, assemble and adjust to different products, batch sizes and tasks, without needing to be manually reconfigured or reprogrammed each time. Such as:

• Robotic arms, grippers and end-effectors designed to handle a range of shapes, sizes and materials
• Control and learning approaches that let robots handle irregular, flexible or previously unseen items
• Tools and methods that make it fast to switch from one product or task to another
• Ways of testing and refining behaviour in simulation before deploying it and scaling it reliably in real environment.

Robots and physical AI systems that make safe, understandable decisions instantly, especially in busy environments where people, equipment and conditions are constantly changing. Such as:

• Systems that sense a situation and decide how to act in real time, without noticeable delay
• AI that can explain why it made a particular decision, rather than acting as an unexplainable “black box”
• The ability to safely adjust plans on the fly when the base environment suddenly changes
• Approaches and solutions that meet the safety standards required to operate around people and valuable equipment

Tools that connect machines, robots, sensors and software together so they can reliably share information, work together and be rolled out easily across a factory or multiple sites, including tools simple enough for non-specialists to use. Such as:

• Ways of connecting different machines, robots and sensors, even when they come from different manufacturers, so they can share data and work as a system
• Simple, low-code or no-code tools that let operators configure and adjust automation without needing an engineer or programmer every time
• Central dashboards or systems that show what's happening across a whole plant, not just a single machine
• Approaches that make it realistic to replicate a working setup across multiple production lines or factories, not just one