Method and system for multiple sensor correlation diagnostic and sensor fusion/DNN monitor for autonomous driving application
US10551838B2 · kind B2 · utility
Assignee
Inventors
Key dates
| Filing date | Aug 8, 2017 |
| Grant date | Feb 4, 2020 |
| Priority date | — |
| Expiry date | Mar 2, 2038 |
Classification
- Technology area (CPC G)Physics
- CPC primaryG01S2013/93274
- WIPO fieldMeasurement
- WIPO sectorInstruments
Abstract
The system can diagnose sensor issues for a vehicle that can perform automatic driving functions based on areas where sensors overlap coverage and sensor performance limitation (SOTIF (Safety Of The Intended Functionality) in ISO 26262). First, the system can identify sensor overlap zones for advanced driver assistance systems/automated driving (ADAS/AD) applications that can be used for multiple sensor correlation and diagnostics. Once the zones have been identified, the system can define detailed multiple sensor correlation and diagnostics (e.g., the same kind of sensor, different kind of sensors, two-way plausible check, three-way voting, etc.). Finally, verified sensor output may be used in monitoring sensor fusion/deep neural network (DNN) outputs within defined a monitor zone.
Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.