Software-Oriented Collaborative Project-Based Learning of Biomedical Signal Processing in Simulated Industry-Like Conditions

Abstract

Most academic approaches to teaching biomedical signal processing (BSP) focus either on acquisition procedures of electrophysiological data, predefined menu-driven signal processing methods using virtual laboratories, or ordinary signal processing procedures applied to biomedical signals without reaching the nature of the data and reflecting the future workplace. This paper introduces a new software-oriented project-based learning approach to BSP for graduate biomedical engineering students. The course simulates industry-like practices under an imposed work environment and acquaints the participants with biomedical signal evaluation and quality assurance (QA) procedures required for software standardization. The students develop fully functional multifaceted software to analyze real-world electrocardiograms that can be run on a desktop computer without external numerical tools. The project is led by a project manager (PM) and assisted by a software architect (SA) and project coordinator (PC). The paper hypothesizes that learning the BSP through a proposed simulated industry-like approach improves understanding of BSP principles, computer programming skills, and social competencies in developing a collaborative project. The results show that the participants significantly enriched BSP-related knowledge after the course, improved computer programming skills (p < 0.0001; nonparametric Wilcoxon signed-rank test), and enhanced soft skills in collaborative work (p < 0.001) and public presentations (p < 0.001). The course participants valued the role of algorithm prototyping stages and QA procedures according to official standardization rules. The PM and SA enabled the smooth software development process, while the PC has proven helpful in resolving intrinsic conflicts.