Samsung patents self-checking system for car safety warnings
Samsung describes a method to verify that safety warning lights and sounds in vehicles or robots work correctly. The system uses a neural processing unit to compare the actual visual and audio output against a pre-trained model to detect errors.
In the filing’s own words
Methods and systems for verifying safety compliance of Safety Warning Sound and Graphic Sign systems to achieve ASIL B or D level in automotive and robotics applications. The method includes providing a predetermined safety warning sign and sound set comprising graphic symbols and auditory signals, transmitting commands to hardware components such as a display processing unit (DPU) and an audio unit (AUD) to generate the safety warning outputs, and capturing output data including displayed frames and sound clips. A neural processing unit (NPU) processes the captured data using a pre-trained inference model generated from the safety warning set and valid sensor data. The processed data is compared against predetermined safety criteria to detect deviations, triggering error reporting and hardware reset commands if necessary. The system supports over-the-air updates, multi-layered display composition, and adaptive sound mixing, ensuring compliance with safety standards under varying environmental conditions.
What exactly they claim (claim 1)
1. A method for verifying safety compliance of a safety warning sound and graphic sign system for achieving ASIL B or D level, the method comprising: providing a predetermined safety warning sign and sound set comprising predetermined graphic symbols and auditory signals for alerting an operator; transmitting, by a processor, a command to safety warning sign and sound hardware including a display processing unit (DPU) and an audio unit (AUD) to generate the safety warning sign on a display and the safety warning sound through speakers in accordance with the predetermined safety warning sign and sound set; capturing output data, the output data including an output frame of the displayed safety warning sign and a corresponding sound clip of the safety warning sound from the safety warning sign and sound hardware; processing the captured output data using a neural processing unit (NPU) executing a pre-trained inference model generated from the predetermined safety warning sign and sound set, and valid sensor data; comparing the processed output data with predetermined safety criteria corresponding to the predetermined safety warning sign and sound set; and in response to detecting a deviation between the processed output data and the predetermined safety criteria, reporting an error signal, and issuing a reset command to the safety warning sign and sound hardware, whereby the method provides an end-to-end checking mechanism ensuring that the safety warning sign and sound output meets ASIL B or D safety requirements.
Facts
- Company
- Samsung (Samsung Electronics Co., Ltd.)
- Official title
- METHODS AND SYSTEMS TO ACHIEVE AUTO SAFETY INTEGRITY LEVEL B AND D FOR SAFETY WARNING SIGN AND SOUND SYSTEM
- Publication number
- US20260330474A1
- Application number
- 19306586
- Filed
- Aug 21, 2025
- Published
- Oct 22, 2026
- Inventors
- 2
- Claims
- 21
Classified under
G06Q 30/018 Physics & computingMore from Samsung in Cars & transport
- Samsung patents AI system for real-time autonomous parking Sep 8, 2026
- Samsung patents map compression for self-driving cars Sep 29, 2026
- Samsung patents car system that adjusts data size to keep remote control fast Oct 6, 2026
- Samsung patents car memory that separates safety-critical tasks from basic ones Oct 6, 2026
- Samsung patents radar method to predict cut-in vehicles using wheel motion Oct 20, 2026
Summaries are simplified from the filing’s own abstract and first claim. A patent filing is not a product announcement.
Patent data and drawings: U.S. Patent and Trademark Office (USPTO). Nebula is not affiliated with or endorsed by the USPTO.