Detection of Unintended Acceleration in Longitudinal Car Following
Academic Article
Overview
Research
Identity
Additional Document Info
Other
View All
Overview
abstract
Copyright 2015 SAE International. This paper presents a model-based approach to detect unintended acceleration (UA) as well as other vehicle problems. A diagnostic system is formulated by detecting several specific vehicle events such as acceleration peaks and gear shifting. Mathematical models are created for these events based on simulation data and the final diagnostic conclusion is drawn from the voting result of all these models. The detection algorithm is validated using independent data sets obtained from Matlab/Simulink. A three dimensional vehicle model is built to implement traffic simulation. Vehicle problems and drivers' reactions are simulated and added during the process. Sensor noise is also considered and corresponding filters are designed and applied. The results show that the fault diagnostic system is successful in detecting UA.