June 15, 2019
June 15, 2019
October 19, 2019
Electrical and Computer
In this paper, a Linear Systems laboratory project is described that involves designing a simplified speech recognition system to recognize the 5 long vowel sounds for a team of students. This laboratory project has several goals including (1) to solve an engineering problem using the frequency domain, (2) practice using the matlab language and development environment, and (3) provide a setting where a number of ABET outcomes can be practiced and measured, specifically Student Outcomes (a), (b), (d), (e), (h) and (k). This is a first semester, junior year laboratory that is a co-requisite of a standard Linear Systems course.
The students are directed to use a solution strategy that involves collecting training data, converting the training data into the frequency domain, analyzing the frequency domain data, and developing a decision tree that implements the decision logic. The decision tree nodes use spectral energy in specific frequency bands to create a metric vector and use support vectors to define decision regions. The nearest neighbor decision rule is used to select a path in the decision tree.
To provide a real-world context to this engineering design, student teams are put together that mix accents and gender in order to provide a diverse data set for each group. Besides determining system performance, students are required to perform a simple market study to estimate the number of potential users assuming that their system performance translates to the performance of a full speech recognition system. This requires some research into the number of English speaking persons with accents from the team and also with consideration of their system performance.
Ebel, W. J., & Carroll, C. (2019, June), A Speech Recognition Linear Systems Lab Paper presented at 2019 ASEE Annual Conference & Exposition , Tampa, Florida. 10.18260/1-2--31999
ASEE holds the copyright on this document. It may be read by the public free of charge. Authors may archive their work on personal websites or in institutional repositories with the following citation: © 2019 American Society for Engineering Education. Other scholars may excerpt or quote from these materials with the same citation. When excerpting or quoting from Conference Proceedings, authors should, in addition to noting the ASEE copyright, list all the original authors and their institutions and name the host city of the conference. - Last updated April 1, 2015