← Retour au blog
tech 1 July 2026

I Built a mmWave Material Classification Radar (2025)

Learn how I designed an innovative radar capable of detecting asbestos in buildings, overcoming the challenges of hardware development.

Article inspired by the original source
I built a mmWave material classification radar (2025) ↗ gauthier-lechevalier.com

Introduction

Innovation never sleeps, especially in the field of material detection. While software has become a commodity thanks to advances like Claude Code, hardware remains a major challenge. Let me tell you how I built a mmWave material classification radar—a journey that started as a graduation project and evolved beyond my expectations.

The Idea Behind the Project

In Europe, asbestos is a common plight. Found in many buildings, it's often detected through costly and invasive methods. My idea was to develop a radar capable of non-invasively detecting asbestos. Using my knowledge in material sciences and wave physics, I designed a device that could revolutionize this sector.

The Technological Choice

To rapidly prototype, I chose the Texas Instrument IWRL6432 BOOST and an ESP32 development kit. These choices allowed for reduced development and iteration time. I began working on DSP algorithms to detect materials and built a test bench to evaluate the electromagnetic response of materials.

The Digital Signal Processing Chain

The radar I constructed uses FMCW (Frequency Modulated Continuous Wave) technology. It emits a frequency that linearly sweeps over time, generating a "chirp." The DSP chain transforms the echoes of these chirps into precise material signatures. This process begins with chirp generation and characterization, a crucial step as everything depends on the exact shape of this frequency sweep.

Classifier Training

After developing an approach based on Capon beamforming to obtain a density spectrum, I integrated a neural network to classify material surfaces. Under the hypothesis that "same surface, same layer" and "material change is sudden and discontinuous," I could determine the composition of material layers.

Challenges and Results

Although the project was hindered by a lack of funding, the results obtained with the prototype were promising. The cost of asbestos detection could be significantly reduced, making this technology accessible and cost-effective for end users.

Conclusion

Creating a mmWave material classification radar has been a journey of innovation and challenges. As a tech entrepreneur, this project taught me resilience and the importance of rapid iteration. I hope this example inspires you to explore and innovate in your own projects.

Let's discuss your project in 15 minutes.

mmWave radar material classification DSP asbestos detection
Deepthix newsletter · 100% AI · every Monday 8am

An AI agent reads tech for you.

Our AI agent scans ~200 sources per week and ships the best articles to your inbox Monday 8am. Free. One click to unsubscribe.

Visit the newsletter page →

Want to automate your operations?

Let's talk about your project in 15 minutes.

Book a call