loading page

Skedastic boosting: exploiting sensor diversity for hidden signal estimation
  • Celestine Lawrence
Celestine Lawrence
University of Groningen

Corresponding Author:[email protected]

Author Profile


Here we propose a novel mathematical method to construct an analog gauge using a weighted-sum combination of several sensor measurements, where the sensors have significant skedasticity in their response functions. We provide a theoretical derivation for the weights and an empirical demonstration of the method (skedastic boosting) on a real-world dataset with applications to gas concentration estimation.