Keeping humans safe around robots with Rust
summary
Sonair builds an functional safety-certified ADAR sensor concept using Rust on a multicore Cortex‑A platform, moving from Raspberry Pi prototypes to bare‑metal quad‑core A53 compute with an M4 integrity core. The system handles >64 configurable safety zones, meets a 100 ms stop‑decision window by starting processing as samples arrive, uses an atomic counter multi‑core scheduler, and even benchmarks an AI‑generated FFT faster than vendor DSP. ADAR 1.0 ships functionally, evaluation is ongoing, safety certification targets Q1.
AI-written summary · about
chapters (14)
- 0:00ADAR and functional safety
- 2:333D ultrasound sensing importance
- 3:36Real-time processing challenges
- 5:46Hardware prototypes evolution
- 7:33Algorithms in Rust
- 9:21Certifiable hardware platform
- 11:09Bare-metal Rust choice
- 13:17Multi-core Rust parallelization
- 16:48Functional safety in Rust
- 18:34Safety and integrity separation
- 20:18FFT performance with Rust
- 21:43ADAR 1.0 safety zones demo
- 23:09Safety certification path
- 26:05Summary and Q&A
related talks
41m"Rust for Robotics: Safer, Faster Systems for Autonomous Applications"
20m"Rust in Motion: Building Reliable and Performant Robotics Systems"
02mRustbots - Robots powered by Rust
01h 09mThe Rust Programming Language: Origin, History and Why it's so important for Arm today
39mReal-Time Rust in Robotics
38mFearless Single-Photon Detection: Leveraging Rust for Scientific Instruments
33mBuild your own (Rust-y) robot!
56mRust for Foundation Software
