An efficient censoring scheme for lifetime of connected solid-state lighting based on entropy measures

Abstract

Beyond energy saving, LED connected lighting systems allows for real-time data streaming, generating intelligence and creation of intelligent buildings systems. Lifetime experiments is crucial for such a connected system, as early termination would lead to big errors in estimating the lifetime model parameters, and late termination would waste time and cost. In this paper sup-entropy is applied in quantifying the amount of information in censored solid state lighting (SSL) luminaire experimental data that has Weibull distribution with respect to complete data. We use a quantity called efficiency of the censored sample as an indicator of the quality and goodness of the censoring scheme. This allowed us to determine the suitable termination time of the experiment subject to a desired efficiency. An intensive simulation study is conducted to validate our procedure.