Masters' Theses A study on Real-Time Collection of Shop Floor Data Using Speech Recognition
2015.07.29 11:10
| year | 1995 |
|---|---|
| author | Namjoo Kim |
| Keyword | Manufacturing database, Shop floor data, Speech Recognition, Multi-Layered HMM, Phoneme Extraction |
| Abstract | Nowadays, many researchers try to improve the intelligence of manufacturing systems as CIM and IMS proliferate. One of important performance measures about manufacturing is the efficiency of the systems. In order to make the manufacturing systems more efficient, the role of database is very important. Especially, real-time data collection is getting more attention from many researchers. Today, magnetic card readers and bar code readers are becoming more popular for updating the database and many deficiencies of hand-written documents are overcome. Those methods, however, have deficiencies that they are not flexible enough to accommodate diverse situations. The only flexible method to deal with dynamic situations is to use keyboards as input devices. Using keyboards, however, may cause inaccurate database because workers not familiar with keyboard input may make mistakes. So, in order to input real-time data more accurately and in more timely manner, this research proposes a methodology based on speech recognition, where workers can input shop floor data in real time easily. Existing algorithms for speech recognition are studied, categorized and implemented in this research. And a new architecture like Multi-Layered HMM and a new algorithms like Phoneme Extraction Algorithm are devised to improve the efficiency of the speech recognition system, and finally those implemented algorithms are compared to select the most appropriate one in CIM environment. |
| c | MS |
Downlaod :
