Ich zitiere mal aus obigen Artikel:
The primary sensor is comprised of an LED array that passes light through the muscle belly of the gastrocnemius (also known as the calf muscle), and a detector. As the light passes through the muscle tissue, it is reshaped by elements within the tissue (known as chromatophors) to produce a unique signal (similar to a fingerprint) that contains information about the local metabolic activity.
This profile is then processed by embedded BSX algorithms to analyze identifiable signal features which exist within the profile. When combined together, these are used to accurately generate a lactate threshold curve. Real-time monitoring of this curve allows the athlete to always know where they are along that curve and their proximity to crucial training thresholds.
Wenn ich da richtig verstehe, dann wird nicht direkt
Laktat gemessen, sondern der gesamte NIR-sichtbare Metaboliten-Mix. Das ergibt anscheinend ein fingerprint-artiges Spektrum, aus welchem man den Laktatgehalt ableiten kann.
Vermutlich so ähnlich wie neuronale Netze, die trainiert werden, bis sie eine sehr gute Vorhersage liefern.
