The helmet and head are too loosely mechanically coupled. Let's say you have acccelerometer data from a sensor on the helmet and you pass it through an LPF: I

The helmet and head are too loosely mechanically coupled. Let's say you have acccelerometer data from a sensor on the helmet and you pass it through an LPF: If you have the helmet bobbying at really high forces you over-represent real G forces and you are right that you could negate that with filtering. _However_, if you have really high forces transmitted from the bobsleigh up to the torso/neck/head it is possible that the inertia of the loosely coupled helmet makes it not register the full force as propagated through the tightely coupled torso/neck/head and you underrepresent the real G forces experienced by the brain/skull. Accelerometer sensors should be coupled with the skull for proper risk/hazard assessment.