For many autistic kids and teens, anxiety is common but hard to notice and communicate in the moment. Wearables that track heart rate could help, but they often can’t tell if a racing heart means “I’m anxious” or “I’m just walking fast,” leading to lots of false alarms.

This research solves that problem by combining heart‑rate data with movement data from a wearable sensor. The algorithm constantly learns each child’s normal heart pattern while still, walking slowly, and walking faster, then looks for unusual spikes on top of those patterns. That means it can detect anxiety‑related arousal even when the child is moving, and ignore heart‑rate increases that come from normal activity.

The team tested the system with 15 autistic children and youth aged 8–16, who wore a chest sensor while standing, slow‑walking, and fast‑walking on a treadmill. In each condition, they had a calm video “baseline” and then a stressful colour‑word (Stroop) task to raise anxiety. The new algorithm reached about 91% overall accuracy and greatly improved specificity (fewer false positives), especially during walking, compared with their previous method.

In plain terms: this is a motion‑aware way for wearables to say, “This spike really matters,” instead of triggering every time a child moves. It’s a key step toward real‑time, objective anxiety support in everyday life for autistic kids and teens.