Belgian research spin-off Brailsports has officially presented the findings from the real-world deployment of its AI platform with professional UCI cycling team Lotto Dstny. Created as a joint spin-off by research institute imec alongside Ghent University and the University of Antwerp, the venture focuses on overhauling physiological monitoring in endurance sports. Rather than depending on manual assessments, the platform applies machine learning algorithms to evaluate athletic stress continuously and objectively. The developers thereby demonstrate how academic research from Flanders can transition directly into the demanding routines of elite international cycling.
At the core of this technological transition lies the replacement of conventional, invasive diagnostic methods. Historically, coaches and sports scientists have depended heavily on incremental lactate step tests, which require drawing blood samples from cyclists during exhaustive exertion intervals to analyze metabolic shifts. These traditional approaches carry notable drawbacks, as they deliver only isolated snapshots in laboratory settings or on rigid test tracks while interrupting regular preparation cycles. By deploying the new algorithmic model, the team can dispense with uncomfortable blood sampling without forfeiting critical physiological feedback.
From a technical standpoint, the AI system continuously ingests diverse metrics recorded throughout daily rides. The machine learning models combine extensive historical training logs with live streams from sensors fitted to both the bicycles and the riders. Primary parameters analyzed by the platform include power output measured in watts, heart rate variability as an indicator of physiological recovery, and pedaling cadence. To ensure an accurate representation of physical exertion, the platform also incorporates external elevation profiles directly into its computational pipeline.
Drawing on this comprehensive dataset, the software generates reliable projections regarding each athlete's current physical state. The algorithms calculate real-time fitness levels and determine the exact fatigue rate for every cyclist on the Lotto Dstny roster. This capability enables coaching staff to abandon rigid, pre-planned weekly schedules in favor of a flexible regime. During ongoing operations, daily training prescriptions are updated dynamically whenever the sensor data indicates that a rider requires additional rest or possesses remaining capacity for higher workloads.
The overarching objective of this data-driven methodology centers on proactive risk mitigation. By detecting subtle fatigue patterns early, the system helps coaches prevent chronic overtraining, an issue that frequently leads to prolonged performance drops during a racing season. Furthermore, calibrating workloads with precision helps minimize injury-related absences across the squad and optimizes recovery following grueling mountain stages. Through its partnership with Brailsports, Lotto Dstny illustrates how non-invasive data modeling can redefine performance diagnostics across modern endurance athletics.