Dynamic Design of Experiments for Engine Pollutants Emissions Modeling and Optimization

Designs of Experiments (DoE) that are appealing to system identification are by and large optimized with respect to model structure. In spite of this, for a nonlinear dynamic system, an optimal design is still an open question particularly when dealing with different model structures independent of any known methods. This article presents a new heuristic to build a Dynamic Design of Experiments based on classical designs but in a higher dimension feature space. The complete heuristic of data preprocessing and dynamic modeling is also presented in the particular case of diesel engine emissions optimization process.