Characterizing ILKATBO trigger detection in Volvo Trucks using XGBoost and SHAPLEY values

Sammanfattning

This thesis examines how ILKATBO trigger frequency varies across the Volvo truck fleet and which operating conditions and vehicle characteristics are associated with elevated trigger behavior. Using logged vehicle data, trigger rate was analyzed at fleet level and within different driving-condition subsets. The results show that trigger behavior is systematic rather than random. At fleet level, road condition is the most important factor, with rougher driving environments associated with higher trigger frequency, while altitude and temperature also show clear associations. Some vehicle-configuration effects appear consistently across conditions, particularly for front axle load, while others seem more closely tied to regional or operational clusters. Overall, the study shows that ILKATBO trigger frequency is concentrated in identifiable parts of the fleet and that both operating environment and vehicle setup contribute to that variation. Results describe vehicles at the extremes of the trigger rate distribution and do not necessarily characterize the full fleet.