Case Study: Daihatsu achieves skilled-worker-level engine knock classification with MathWorks (MATLAB)

Daihatsu - Customer Case Study

Daihatsu faced the challenge of automating judgment of engine knocking—a task historically done by skilled workers—while needing a tool that combined deep learning with acoustic analysis and could help train AI-literate engineers. To address this, Daihatsu chose MathWorks’ MATLAB, leveraging Audio Toolbox, Deep Learning Toolbox, and Statistics and Machine Learning Toolbox for easy data import, visualization, and rapid model development.

Working with MathWorks experts, Daihatsu used feature extraction and machine learning to build classification models that identify knocking sounds, achieving the same accuracy as skilled workers. The MathWorks solution also enabled fast iterative analysis, increased in-house AI expertise through MATLAB training, and improved visualization and awareness of AI across the engineering team.

Daihatsu

Takuya Kumagae

Daihatsu