Interval state observer based on a time-variant transformation for LPV systems and application to induction machines

For the state estimation of safety-critical systems, interval observers are an emerging alternative to classical state estimation methods due to their provision of guaranteed information on unmeasurable quantities. One possible application of interval observers are induction machines used as a traction drive whose common models belong to the class of linear parameter-varying (LPV) systems. To overcome the design conditions of existing interval observers for LPV systems, which are too strict for an application to the aforementioned system, we propose a new interval observer design based on a time-variant transformation. The effectiveness of the method is shown by experimental results of an induction machine.