RAPIDS - Reliable Adaptive Predictive maintenance and Intelligent Decision Support

Important results from the project

The goals were largely met. RAPIDS has developed new methods for predictive modeling of vehicle components with a focus on streaming data, uncertainty quantification and decision support under uncertainty. The project has delivered processes for feedback and continuous learning, a public dataset with guidelines for responsible sharing, and a Continuous Integration/Delivery-pattern for machine learning. The results have strengthened the research field and contributed to industrial application.

Expected long term effects

The project is expected to have long-term effects by establishing a strong knowledge base and methodology for predictive maintenance. The results contribute to continued research in uncertainty quantification, model updating and feedback, and to the development of self-learning multimodal systems. The public data set and the developed processes for data sharing can be used in new projects. Methods, frameworks and working methods from the project are now being implemented within Traton R&D.

Approach and implementation

The project was carried out in close collaboration between industry and academia and was structured around five work packages. The iterative working method with a tight connection between research and application enabled the rapid conversion of ideas into practical results. The implementation largely followed the plan, but was extended by nine months due to recruitment delays and high workload within Scania. Regular meetings and workshops provided clear steering and good coordination.