Mobile road weather sensor calibration by sensor fusion and linear mixed models

摘要: Mobile, vehicle-installed road weather sensors are becoming ubiquitous. While mobile often capable of making observations on a high frequency, their reliability and accuracy may vary. Large-scale observation forecasting still mostly based stationary stations (RWS). Though expensive, sparsely located relatively low RWS' well-known accommodated for in the models. Statistical analysis revealed that conditions indeed have great effect how temperature differ from each other. Consequently, we calibrated with linear mixed model. The model was fitted fusing ca. 20 000 pairs RWS same location at time, following rendezvous sensor calibration. calibration nearly halved MSE between types. Computationally very light, can be embedded directly sensors.