KAI Kompetenzzentrum Automobil- und Industrieelektronik
KAI Kompetenzzentrum Automobil- und Industrieelektronik GmbH was founded in 2006. Today it is a well-established competence center with 46 employees (04/2018). The main focus is on power electronics reliability, materials research, modeling and simulation.
KAI Kompetenzzentrum Automobil- und Industrieelektronik GmbH is a legally independent subsidiary, 100% owned by Infineon Technologies Austria AG. KAI is a well-established industrial research center with a large national and international network of partners. Besides core competences in the area of power electronics reliability, test concept and methodology development, advanced semiconductor materials research, Bayesian statistical lifetime modelling and multi-physics FEM simulation, KAI maintains a well-equipped electronics design laboratory, dedicated measurement equipment for materials characterization as well as state of the art simulation computing resources, complemented by a long proven experience in the coordination of interdisciplinary research projects. – Statistical modelling of device life time with physical regression models as well as the development of data analysis methods for semiconductor production data based on machine learning techniques are additional branches of academic research at KAI that are currently supported by national scientific funding programs.
KAI Kompetenzzentrum Automobil- und Industrieelektronik GmbH
KAI is a well-established industrial research center with a large national and international network of partners.
KAI develops and applies advanced statistical methods for semiconductor reliability and manufacturing data since 2008. Since 2013, the statistical methods and degradation modeling research group at KAI develops data mining concepts and advanced statistical screening methods based on machine learning approaches to identify risk devices in the electrical measurements after the semiconductor Frontend production. With the tasks defined in iDev40, KAI will go one step further by analyzing additional data sources, collected during semiconductor Frontend manufacturing. Since these data are more complex and significantly larger, new statistical methods need to be developed and skills from other disciplines (e.g. information technology) are required. Via internal experts and a large scientific network, KAI has access to this know-how and will coordinate the activities within UC1.
In this project KAI acts as research partner for Infineon Technologies Austria AG for the development of a data-driven root cause analysis using semiconductor manufacturing data (see UC1). Within iDev40 KAI investigates and validates machine learning and advanced statistical methods which are capable to deal with the complexity of the available data to support business domain experts on their root cause investigations. Hence, manual effort of data analysis is decreased, fast learning cycles are enabled and further, this knowledge is preserved. To achieve this aim the following tasks are addressed:
- Development and application of intelligent data pre-processing methods for complex and heterogeneous data from different sources, including the setup of SW infrastructure to allow efficient and robust combination of large data sets
- Development of machine learning methods for root cause investigations, supporting experts in their evaluations
- Implementation, application and validation of developed data-driven methods
- Proposals for countermeasures in the event of detected root causes (e.g. adaption of control limits)