This content is not included in your SAE MOBILUS subscription, or you are not logged in.
Scenario Uncertainty Modeling for Predictive Maintenance with Recurrent Neural Adaptive Processes (RNAP)
Technical Paper
2021-01-0191
ISSN: 0148-7191, e-ISSN: 2688-3627
Sector:
Event:
SAE WCX Digital Summit
Language:
English
Abstract
For commercial-vehicle original equipment manufacturers, predictive maintenance has drawn attention for the benefits to facilitate money-saving and increased road safety. Data-driven models have been widely explored and implemented as predictive maintenance solutions. However, the working scenarios for different commercial-vehicles vary a lot, which makes it difficult to build a universal model suitable for all the cases. In this paper, we propose a Recurrent Neural Adaptive Processes network to adapt to different scenarios by modeling the uncertain at the same time. The ensemble network combines the traits of neural processes, recurrent neural network and meta learning together. Neural processes could achieve prediction in stochastic based on the input dataset. Meta-learning is good at dealing with few-shot multi-tasks learning, and recurrent networks are utilized as the encoder of the neural processes to fit for the input time-series data. The proposed network is used to do both anomaly detection using regression with threshold and fault code classification tasks. It improved the accuracy compared to the existing approaches that do not ensemble all the parts together.