S1-COMP8-2 - Phone Keypad Voice Recognition (PKVR): An Integrated Experiment for Digital Signal Processing Education

1. Innovative Practice Work In Progress
Xiaoyan Luo1 , Lu Li1, Han Wan1, Fugen Zhou1
1 Beihang University

Title: Phone Keypad Voice Recognition (PKVR): An Integrated Experiment for Digital Signal Processing Education

Keywords: Digital Signal Processing; Practice; Curriculum Design; Experiment Design

Topics: Curriculum Design; Teaching & Learning Experiences in Engineering Education; Discipline Specific Issues: Information Engineering

Paper abstract (Work in Progress):

Digital signal processing (DSP) is always a core course for information engineering related majors. Facing the classical DSP theory, its corresponding practical training is fragmented and difficult to design. We had considered and discussed how to design the practical experiment in DSP course at AutumnSim 2019. In this paper, we present an integrated practice framework for training experiment in DSP course. Since the DSP course focuses on 1 dimension signal processing, voice signal has a great advantage. We provide an integrated voice signal processing experiment named as Phone Keypad Voice Recognition (PKVR), which includes the following parts: phone keypad voice collection, signal denoising preprocessing, Discrete Fourier Transform (DFT) and analysis, filter design, digital query table establishment, number recognition of any keypad voice. This experiment design can cover all the important knowledge points of whole DSP course. For examples, the voice acquisition step is used to illustrate the concept and properties of digital signal. The DFT operation can aid to understand the principle and effect of the frequency transformation and frequency spectrum analysis. Filter design part also helps students reveal the essence of filter. Finally, the DSP system integrity and complete application case is embodied into the establishment of digital query table and the number recognition of any keypad voice part. Moreover, the introduction of these knowledge points is progressive, which is consistent with the experiment progress requirement. In our DSP course, students can use any software platform without limitation, whose course autonomy in the practical programming is ensured and increased. And then, they can use their own mobile phone as experiment acquisition equipment to collect their keypad voice from 0 to 9. Each number voice signal should be kept in same length, which is transformed into frequency domain by DFT. According to the frequency distribution, the students need to design low-pass and high-pass filters. Based on dual-tone multi-frequency strategy, digital query table is constructed referring to the filtered high frequency group and low frequency group. After these operations, the teacher can test the PKVR of each student with different keypad voice input. For each student, the practice performance is measured by the consistency between the real input numbers with the recognized numbers. This paper presents the rationality and detail about PKVR experiment setup. The students who joined our DSP course reflected that PKVR could be very interesting and very helpful to understand the theory.