Experimental study on sparse modeling of a diesel engine air path system

Toward model-based development of automotive engines, it is crucial to develop an effective method to construct their accurate and low-complexity models. In this paper, we conduct an experimental study on sparse modeling of a diesel engine air path system. First, we apply two standard methods and examine their trade-offs; the Gaussian process regression that can provide a confidence interval; the ℓ¹-regularized regression that can reduce the number of basis functions. Next, we attempt to take advantages of both methods via the so-called Relevant Vector Machine.