International Journal of Production Research, Vol. 45, No. 10, 15 May 2007, 2295–2311 Data-reduction method for spatial data using a structured wavelet model MYONG K. JEONG*y, JYE-CHYI LUz, WEIXIN ZHOUz and SUJIT K. GHOSH§ yDepartment of Industrial and Information Engineering, The University of Tennessee, Knoxville, TN 37996-0700, USA zSchool of Industrial and Systems Engineering, Georgia Institute of Technology, Atlanta, GA 30332-0205, USA §Department of Statistics, North Carolina State University, Raleigh, NC 27695-8203, USA (Revision received January 2006) Recent advances in sensor instrumentation have provided opportunities for process engineers to collect data at various process steps in order to detect process problems and develop remedial procedures. This article presents a structured wavelet model for the reduction of two-dimensional data having distinct structures. The wavelet component of our model can handle irregular data patterns exhibiting many peaks and valleys, while the existence of a distinct data structure prompts the use of polynomial functions on wavelet coefficients. The two-dimensional antenna data is reduced with a structured wavelet model followed by some procedures for the detection of process defects based on the reduced-size data. A real-life example is presented to illustrate the usefulness of the proposed tools in detecting process problems from a potentially large volume of data exhibiting many peaks and valleys. Keywords: Fault detection; Intelligent manufacturing; Process control; Structured model; Wavelets 1. Introduction The increasing popularity of wireless communications in recent years has increased the demand for antenna equipment for sending and receiving signals. The technologically sophisticated antennae developed for this market require a high degree of quality during the production process. To monitor antenna manufacturing quality, our team of students and faculty worked with engineers at the Nortel production facility in Research Triangle Park (RTP), North Carolina, in order to collect antenna data sets for the purpose of developing procedures to detect process problems. *Corresponding author. Email: mjeong@utk.edu International Journal of Production Research ISSN 0020–7543 print/ISSN 1366–588X online # 2007 Taylor & Francis http://www.tandf.co.uk/journals DOI: 10.1080/00207540600793547