Among all multi-automotive sensor target tracking algorithms, statistical sequential estimation techniques are in the dominant position. For these techniques, statistical sensor measurement uncertainty model is essential. However, for camera, which is a primary automotive sensor, there exists a problem that vision measurements are usually not statistical model-based. In this paper, to open up a possibility of establishing an access between the vision measurements and statistics-based methods, we try to describe vision distance measurement (VDM) uncertainty in a statistical way. As the foundation for statistically analysing VDM uncertainty, we first derive the expression of VDM uncertainty. Based on the expression, we then model the probability distribution of VDM uncertainty, thus the statistical VDM uncertainty model is obtained and VDM is connected with the statistics-based fusion methods. The proposed statistical model is validated by experimental data.
Modeling and Statistical Analysis of Vision Distance Measurement Uncertainty for Multi-Automotive Sensor Target Tracking
2022-10-08
658292 byte
Conference paper
Electronic Resource
English
V2V Communications in Automotive Multi-Sensor Multi-Target Tracking
British Library Conference Proceedings | 2008
|Demonstrator: V2V Communications in Automotive Multi-Sensor Multi-Target Tracking
British Library Conference Proceedings | 2008
|Realtime Vision Based Multi-Target-Tracking with Particle Filters in Automotive Applications
British Library Conference Proceedings | 2006
|Statistical uncertainty in automotive emissions testing
SAGE Publications | 1999
|