Research project RAPIDS – Reliable Adaptive Predictive maintenance and Intelligent Decision Support
Less accidents on the road, and more operational uptime. That is the expected outcome of this research project which uses data from trucks to develop new machine learning models. The models will let us know when maintenance is needed – before the vehicle breaks down.
Customer demands on operational uptime of vehicles have increased in recent years and are expected to be further accentuated with the introduction of autonomous and electrified vehicles. Forecasting, and data analysis in general, are areas where machine learning has a strong industrial potential.
The project revolves around developing machine learning models based on increased availability of streamed log data from vehicles, and integration of these models in the decision-making processes for maintenance. It specifically deals with how uncertainty in predictions can be estimated and weighed in order to make robust individual-based decisions. A central feature is also how new data can be fed back to the models in order to improve performance and predictive power.
We will develop theory and generally applicable methods which are then tested and demonstrated on real use cases. The project will help to strengthen the research fields of machine learning and forecasting, specifically in the areas of models for streamed log data, data-driven decision making under uncertainty, and efficient model feedback.
The project's main applicant is Scania CV AB and other parties are Linköping University, Stockholm University and the Royal Institute of Technology. The project is planned to run for three years starting in January 2022 and with a total budget of SEK 19.4 million.
Project description
Customer demands on operational uptime of vehicles have increased in recent years and are expected to be further accentuated with the introduction of autonomous and electrified vehicles. Forecasting, and data analysis in general, is an area where machine learning has a strong industrial potential. In a previous research project – CODA, funded by FFI and Scania –the potential of using data-driven and interpretable methods to plan maintenance was demonstrated.
The RAPIDS project revolves around developing machine learning models based on increased availability of streamed log data from vehicles and integration of these models in the decision-making processes for maintenance. It specifically deals with how uncertainty in predictions can be estimated and weighed in order to make robust individual-based decisions. A central feature is also how new data can be fed back to the models in order to improve performance and predictive power.
The project will develop theory and generally applicable methods which are then tested and demonstrated on real use cases. The project will help to strengthen the research fields of machine learning and forecasting, specifically in the areas of models for streamed log data, data-driven decision making under uncertainty, and efficient model feedback.
The project's main applicant is Scania CV AB and other parties are Linköping University, Stockholm University and the Royal Institute of Technology. The project is planned to run for three years starting in January 2022. It has a total budget of SEK 19.4 million, of which SEK 9.7 million is funded by the program and the rest is funded by Scania. Funding is sought to cover the expenses of the academic parties, as well as the project's costs for external services, travel and other expenses.
Project members
Project managers
Tony Lindgren
Unit head SAS
Olof Steinert
Data Scientist at Strategic Product Planning and Advanced Analytics
Members
Zahra Kharazian
PhD student
