Cranfield University and Plextek Services Limited KTP 24_25 R2

A new radar sensor, small and cheap enough to fit on a forklift or a warehouse drone, will use machine learning to help autonomous vehicles navigate safely through cluttered industrial sites. Current radar sensors that work well in factories and warehouses are often too large, heavy, or expensive for widespread use on smaller autonomous machines. They also struggle to distinguish between a stationary steel beam, a moving person, and a pallet of goods in a dusty, noisy environment. This project combines a compact millimetre-wave radar—which uses short-wavelength radio waves to create high-resolution images even in smoke or darkness—with the latest machine learning algorithms that can interpret those images in real time. If successful, the sensor could allow autonomous forklifts, delivery robots, and inspection drones to operate safely alongside human workers without expensive infrastructure changes. The technology would improve logistics and manufacturing efficiency by reducing the need for dedicated, fenced-off robot zones. It could also lower the cost of automating smaller factories and warehouses that cannot currently afford high-end navigation systems. The project is applied research with a clear industrial endpoint: a working prototype that can be commercialised.