概述
描述
Cameras, Light Detection and Ranging (LiDAR), and radar can’t see around corners. And, they cannot see emergency vehicles that are far away but fast approaching. The best line-of-sight sensors in the world are limited by the fact that they can’t see threats that are hidden from view. With the addition of obstructed view scenarios into the safety standards of Euro NCAP and other international testing bodies, finding new ways to detect threats is critical.
The Reality AI solution uniquely combines both hardware and software to give passengers a new level of protection.
Automotive Safety Warning System (SWS) Detects and Localizes:
- Emergency vehicles
- Cars
- Trucks
- Motorcycles
- Bicycles
- Joggers
Learn about RealityCheck AD for Industrial Anomaly Detection
特性
- MEMS Mic Array integrated into components or placed on the roof. Flexible geometry.
- Automotive MCU runs AI detection and localization software on inexpensive hardware
- Detection provided by AI models that classify different threats accurately at 1.5km distance for sirens, 35m+ for cars, trucks and motorcycles, and 10m for bicycles and joggers
- Localization through AI models that compute the angle of arrival, estimate distance, and detect whether threats are approaching or receding
- AI Firmware engineered by Reality AI (now Renesas), the foremost specialist in the application of machine learning to complex signal processing problems
目标设备
文档
|
|
|
|
|---|---|---|
| 类型 | 文档标题 | 日期 |
| 产品简述 | PDF 354 KB | |
| 白皮书 |
PDF
951 KB
Existing ADAS solutions like cameras and LiDAR require a clear line of sight, limiting their effectiveness. This paper introduces a passive acoustic solution that "listens" to the environment to detect road participants, even when they are around corners or out of sight. It discusses the use cases, challenges, and architecture of this AI-based system, which uses low-cost, automotive-grade components to provide early warnings and supplement existing ADAS technologies for improved safety.
|
|
2 项目
|
||
探索
视频和培训
A look at how data-driven engineering tools can learn from examples to make a decision and ultimately be deployed automatically.
新闻和博客
博客 2025年6月25日 |
博客 2024年10月4日 |
