Abstract
Recent advances in knock control suggest that stochastic controllers can deliver significantly improved regulatory response while achieving similar transient responses relative to conventional deterministic strategies. The new controllers have used the Cumulative Summation (CumSum) of knock events to determine when control actions need to be taken, and a lookup table based on run length to determine the magnitudes of these actions. In this paper a different stochastic approach is used where the likelihood ratio of the observed knock events is used both to determine when control actions are necessary and the magnitude of these actions. The likelihood-based approach is arguably easier to tune, and its performance, like the CumSum controller, is significantly better than traditional deterministic strategies.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call