Virtual Care & Invisible Safety: How Tech Guards Our World

Virtual reality (VR) technology provides an innovative solution, enabling users to interact with three-dimensional environments through multiple sensory channels. Growing evidence suggests that VR may improve mental health and wellbeing, but further research is needed to evaluate its acceptability, usability, and feasibility for individuals with learning disabilities.

Inspecting welded structures - especially in places like nuclear power plants - is a bit like trying to see through a foggy window. It’s essential for safety, but the job isn’t easy: the inside of a weld is full of twisting, turning grains that bend and blur ultrasonic waves. Instead of travelling in neat, straight lines, the waves distort, change speed, and fade, making defects harder to spot
But what if we could predict how ultrasound moves through this maze?
We developed a deep‑learning approach that uses the travel time of ultrasonic waves to uncover the weld’s hidden features: its grain orientations, shape, probe position, and elastic properties. Using thousands of simulated welds and a fast ray‑tracing model, we train neural networks to recognise how different weld structures affect wave paths. We validate the method with detailed simulations and real experiments, then use the recovered weld maps to correct ultrasonic images. The result? Clearer images and better defect identification.