RFID-enabled indoor positioning method for a real-time manufacturing execution system using OS-ELM

详细信息    查看全文

• 作者：Zhixin Yang ; ^{zxyang@umac.mo" class="auth_mail" title="E-mail the corresponding author} ; ^{mb35515@umac.mo" class="auth_mail" title="E-mail the corresponding author} ; ^{mb15550@umac.mo" class="auth_mail" title="E-mail the corresponding author}Author Vitae ; Pengbo ZhangAuthor Vitae ; Lei ChenAuthor Vitae
• 关键词：Manufacturing execution system (MES) ; RFID ; Real-time signal processing ; Indoor positioning ; Online sequential extreme learning machine (OS-ELM) ; Smart manufacturing objects (SMOs)
• 刊名：Neurocomputing
• 出版年：2016
• 出版时间：22 January 2016
• 年：2016
• 卷：174
• 期：part_PA
• 页码：121-133
• 全文大小：2408 K

文摘

Manufacturing execution systems (MES) have recently been introduced to monitor various manufacturing objects (MOs) in dynamic shop floors; they can leverage the efficiency of information flow across functional layers for planning and control. However, current MES practices using traditional indoor positioning algorithms face several difficulties in tracking MOs for wireless manufacturing: (1) inefficient wireless data acquisition in the shop-floor environment, (2) lack of a reliable and accurate real-time signal processing method for handling massive signal data, and (3) the positions of reference objects cannot be treated as in a static environment when unknown manufacturing orders arrive in a streaming manner. This paper proposes to handle the first challenge by adopting RFID technology that can constantly capture the wireless signals sent from tags mounted on various MOs. The second difficulty can be solved by applying the online sequential extreme learning machine (OS-ELM) that inherits the elegant properties of ELM in terms of extremely fast learning speed and high generalization performance. The OS-ELM based positioning method also addresses the third issue in which an online localization model has been constructed in a streaming manner. The proposed method can greatly reduce the training time without costly retraining of the previously trained data together with the newly arrived data. With the novel OS-ELM based RFID positioning framework, the MOs are upgraded to smart manufacturing objects (SMOs), and the processes are enhanced with real-time signal processing and intelligent tracking capabilities. The experimental results verify that the proposed positioning method is superior to other state-of-the-art algorithms in terms of accuracy, efficiency, and robustness.