Judging the Normativity of PAF Based on TFN and NAN
- PDF / 1,473,867 Bytes
- 9 Pages / 612.284 x 810.709 pts Page_size
- 35 Downloads / 177 Views
Judging the Normativity of PAF Based on TFN and NAN
),
LI Zhiqiang (
ª ),
BAO Jinsong ∗ (
LIU Tianyuan (
Í),
WANG Jiacheng (
(College of Mechanical Engineering, Donghua University, Shanghai 201600, China)
)
© Shanghai Jiao Tong University and Springer-Verlag GmbH Germany, part of Springer Nature 2020 Abstract: The normativity of workers’ actions during producing has a great impact on the quality of the products and the safety of the operation process. Previous studies mainly focused on the normativity of each single producing action instead of considering the normativity of continuous producing actions, which is defined as producing action flow (PAF) in this paper, during operation process. For this issue, a normativity judging method based on twoLSTM fusion network (TFN) and normativity-aware attention network (NAN) is proposed. First, TFN is designed to detect and recognize the producing actions based on skeleton sequences of a worker during complete operation process, and PAF data in sequential form are obtained. Then, NAN is built to allocate different levels of attention to each producing action within the sequence of PAF, and by this means, an efficient normativity judging is conducted. The combustor surface cleaning (CSC) process of rocket engine is taken as the experimental case, and the CSC-Action2D dataset is established for evaluation. Experiment results show the high performance of TFN and NAN, demonstrating the effectiveness of the proposed method for PAF normativity judging. Key words: producing action normativity, sequential model, attention mechanism, deep learning CLC number: TP 391.7 Document code: A
0 Introduction With the development of manufacturing automation and the transformation of intelligent manufacturing mode, growing number of intelligent machines and robots participate in workshop operations and play increasingly important roles. Despite this, human beings, as the highest level of intelligent agent, still dominate the majority of producing at this stage. There is no substitute for manual operation especially in product lines which have complicated processes, high degree of refinement, high automation cost or high technical requirements[1]. However, humans are prone to inherent disadvantages such as inattention, physical fatigue and mood swings when compared with robots, making workers’ behaviors uncontrollable[2] , which may impact the normativity of production negatively by lowering product’s quality or increasing operating risk. Therefore, it is necessary to study the normativity of workers’ producing behavior in workshop, which is also one of the important research directions in production management science[3] . Since the video recording technology is developed, the scheme of “record by camera, monitoring by mankind” has become the most popular and effective Received date: 2019-05-23 Foundation item: the National Natural Science Foundation of China (No. 51475301) ∗E-mail: [email protected]
way for decades on production behavior analysis. But it consumes a lot of human resour
Data Loading...
