Chao Li, Weimin Zhai, Weiming Fu, Jiahu Qin and Yu Kang
This study aims to introduce a method for predicting the remaining useful life (RUL) of bearings based on parallel feature extraction. The proposed model provides prior knowledge…
Abstract
Purpose
This study aims to introduce a method for predicting the remaining useful life (RUL) of bearings based on parallel feature extraction. The proposed model provides prior knowledge and removes redundant handcrafted feature information, additionally, which focuses on the important features at different time scales.
Design/methodology/approach
Distinct from traditional parallel feature extraction methods, which can lead to information redundancy, a one-dimensional convolutional autoencoder is introduced to process selected indicators to remove redundancy and retain useful feature information. To fully capture the important degradation information within different stages in the feature sequences, a novel multi-scale attention feature fusion module is proposed to extract degradation features at different time scales. Considering the impact of degradation modes on RUL prediction, a dual-task prediction module based on no degradation mode labels is designed to obtain accurate RUL.
Findings
Comparative experiments and ablation studies on the PHM2012 bearing dataset verified the effectiveness of the proposed method. Furthermore, the rationality of the selected parameters is confirmed through model parameter analysis.
Originality/value
The novelty of the proposed method is that it not only provides prior knowledge but also further removes redundant information from prior knowledge. In addition, the distribution differences between the original features and their multi-scale convolution results are measured through Kullback–Leibler divergence as the attention scores, which allows the proposed method to focus on important information at different time scales.
Details
Keywords
Sijie Tong, Qingchen Liu, Qichao Ma and Jiahu Qin
This paper aims to address the safety concerns of path-planning algorithms in dynamic obstacle warehouse environments. It proposes a method that uses improved artificial potential…
Abstract
Purpose
This paper aims to address the safety concerns of path-planning algorithms in dynamic obstacle warehouse environments. It proposes a method that uses improved artificial potential fields (IAPF) as expert knowledge for an improved deep deterministic policy gradient (IDDPG) and designs a hierarchical strategy for robots through obstacle detection methods.
Design/methodology/approach
The IAPF algorithm is used as the expert experience of reinforcement learning (RL) to reduce the useless exploration in the early stage of RL training. A strategy-switching mechanism is introduced during training to adapt to various scenarios and overcome challenges related to sparse rewards. Sensor inputs, including light detection and ranging data, are integrated to detect obstacles around waypoints, guiding the robot toward the target point.
Findings
Simulation experiments demonstrate that the integrated use of IDDPG and the IAPF method significantly enhances the safety and training efficiency of path planning for mobile robots.
Originality/value
This method enhances safety by applying safety domain judgment rules to improve APF’s security and designing an obstacle detection method for better danger anticipation. It also boosts training efficiency through using IAPF as expert experience for DDPG and the classification storage and sampling design for the RL experience pool. Additionally, adjustments to the actor network’s update frequency expedite convergence.