TY  - JOUR
PY  - 2023//
TI  - Analytical redundant virtual sensor design for hyper-safe driving platform
JO  - Transactions of the Korean Society of Automotive Engineers
A1  - Park, Beomhyuk
A1  - Choi, Seungyong
A1  - Kim, Dohyun
A1  - Cho, Wanki
A1  - You, Seung-Han
SP  - 815
EP  - 825
VL  - 31
IS  - 10
N2  - In order to advance the commercialization of autonomous vehicles, it is imperative to prepare fallback for sensor and actuator failures. This study focuses on the design of robust virtual sensors to counteract failures in key sensors associated with vehicle motion, namely the yaw rate, lateral acceleration, and steering angle sensors. Analytical redundant models were developed and utilized to create alternative virtual sensors. Each virtual sensor incorporates four analytical redundant models: the left/right speed difference of the front wheel, the left/right speed difference of the rear wheel, the lateral physical model, and the induced signal from other sensors. The determination of the final alternative virtual signal is achieved through weighting the redundant models, employing a combination of vehicle domain knowledge and artificial intelligence techniques. The efficacy of the proposed virtual sensors was validated through extensive testing across various driving scenarios using the CARSIM software.<p />  <p>Language: en</p>
LA  - en
SN  - 1225-6382
UR  - http://dx.doi.org/10.7467/KSAE.2023.31.10.815
ID  - ref1
ER  -