The full site goes deep. This page doesn't. Here's the shape of the idea, the headline numbers, and where to go for more.
Standard wearable sensors report peak linear acceleration and peak rotational velocity — two numbers summarizing an entire impact. That's a rough measure of intensity. It says nothing about the shape of the event: how long it lasted, how rotation and translation were related, what frequencies it contained.
Impact Feature Space (IFS) represents each recorded head-motion event as a point in a multidimensional space instead — reading the waveform for what it was and how it happened, not just how big it was.
Rotation was identified as the likely mechanism of concussive brain injury back in 1943. The instruments used to monitor athletes today still mostly report a single peak number. Even the newest scalar criterion, BrIC (2013), still only answers "how much" — just about a different axis. IFS is an attempt to finally ask "what kind."
Shape, timing, amplitude — is the event a sharp transient or a sustained load?
Frequency content, and how it evolves over the event.
How rotation and translation relate — timing, correlation, shared information.
Which direction the motion concentrates in, and how stable that direction stays.
Rotational energy split into axial, coronal, and sagittal components.
How patterned or irregular the waveform is, independent of size or frequency.
Validated so far on elite women's ice hockey. Under exploratory, unpublished consideration: soccer heading, American football, boxing, figure skating falls — each would need its own validation before it means anything.
The complete site covers the framework, the evidence, applications, and the full reference list.