Fast in-cylinder pressure reconstruction from structure-borne sound using the EM algorithm

The paper addresses the problem of reconstructing the low-frequency part of the in-cylinder pressure of spark ignition engines by analyzing structure-borne sound signals measured on the surface of the engine block. The new pressure trace model proposed yields accurate approximations with a minimal number of parameters. When combined with the EM algorithm, a processing scheme results that provides fast pressure estimates and efficiently exploits the information contained in the sound signal. Experimental results with real measurement data show the potential of the reconstructed signal to perform misfire detection and closed-loop spark ignition timing tasks.