Student builds mmWave radar for asbestos detection

// 45d agoTUTORIAL

Student builds mmWave radar for asbestos detection

Gauthier Lechevalier spent six months building a hardware startup around an FMCW mmWave radar designed to detect asbestos in European buildings. By leveraging capon beamforming to process chirp echoes and feeding the data into a neural network, the radar classifies material surfaces to identify hazardous layers without physical intrusion.

// ANALYSIS

–Applying mmWave radar and neural networks to non-intrusive asbestos detection is a highly innovative solution to a pervasive health hazard.
–Rapid prototyping with off-the-shelf boards (IWRL6432, ESP32) demonstrates a lean approach to hardware development.
–The project underscores the difficulty of securing funding for deep-tech hardware startups compared to software companies.
–Combining advanced DSP (FMCW, beamforming) with edge AI for material classification represents a compelling use of modern embedded systems.

// TAGS

radarrfdspembeddedbeamforminghardwarellm

DISCOVERED

45d ago

2026-06-30

PUBLISHED

45d ago

2026-06-30

RELEVANCE

8/ 10

AUTHOR

GL26

// KEEP READING

More AI developer news from the feed

EXPLORE FULL FEED

LAUNCH1h ago

ChatCut turns Codex into video editor

ChatCut brings a full AI video editor into Codex, letting users cut footage, remove silences, add captions, generate B-roll, and build motion graphics through natural-language prompts. Edits remain adjustable on a real timeline, with exports available for professional tools.

NEWS1h ago

three.ws Highlights AI Heavyweight Backing

three.ws founder Nich highlights the platform’s growing user base and relationships with OpenAI, IBM, Google Cloud, and NVIDIA. The project positions itself as a 3D agent platform combining browser-native avatars, AI tooling, on-chain identity, and micropayments.

SECURITY3h ago

Z.ai Delays GLM-5.3 Over Cyber Risk

Z.ai is delaying GLM-5.3’s public weight release by two weeks after the model demonstrated unusually strong vulnerability discovery and exploitation capabilities. The lab says it will use the delay to strengthen safety controls before open-sourcing the model.