A deep LSTM autoencoder-based framework for predictive maintenance of a proton radiotherapy delivery system

作者：

Highlights：

• A novel predictive maintenance framework for proton therapy system is demonstrated.

• Data-driven modeling is based on Long short-term memory deep neural network.

• No positive sample is required for the training of the proposed framework.

• Treatment machine interruptions can be predicted with up to 51-fold enhancement.

摘要

•A novel predictive maintenance framework for proton therapy system is demonstrated.•Data-driven modeling is based on Long short-term memory deep neural network.•No positive sample is required for the training of the proposed framework.•Treatment machine interruptions can be predicted with up to 51-fold enhancement.

论文关键词：Deep neural network,LSTM autoencoder,Predictive maintenance,Anomaly detection,Class imbalance,Proton radiotherapy

论文评审过程：Received 12 January 2022, Revised 23 August 2022, Accepted 24 August 2022, Available online 30 August 2022, Version of Record 9 September 2022.

论文官网地址：https://doi.org/10.1016/j.artmed.2022.102387