Improved sensor error definitions for reliability analysis of multi-sensor systems

Qiu M, Bazan P, Antesberger T, German R (2023)

Publication Language: English

Publication Type: Conference contribution, Conference Contribution

Publication year: 2023

Event location: Bologna IT

Abstract

Defining sensor errors is the key issue of the reliability assessment of perception sensor systems. Recently, the most common sensor error types considered are false negatives (FNs) and false positives (FPs). Under certain circumstances, a sensor perception result may trigger a pair of correlated FN and FP, e.g., when detecting an object with inaccurate depth estimation. To investigate the frequency of the occurrence of this type of error, and to provide a more explainable interpretation of the sensor errors, we redefine them into nondetections, ghost objects, and localization errors, where localization errors are caused by low-quality perception results. Furthermore, we propose a three-step association algorithm to identify these error types from perception results and the ground truths, utilizing the Hungarian algorithm with distance matrices based on the Generalized Intersection over Union (GIoU) and Intersection over Union (IoU). Moreover, we extend a proven reliability model to account for the new error types and fusion methods, e.g., Kalman filters. Finally, we conduct a case study on the nuScenes dataset to show the significance of the proposed sensor error definitions and reliability model.

Authors with CRIS profile

Minhao Qiu Computer Science 7 (Computer Networks and Communication Systems) Peter Bazan Computer Science 7 (Computer Networks and Communication Systems) Reinhard German Computer Science 7 (Computer Networks and Communication Systems)

Related research project(s)

Reliability design of multi-sensor systems Jan. 1, 2020 - Dec. 31, 2022

Involved external institutions

Audi AG DE Germany (DE)

How to cite

APA:

Qiu, M., Bazan, P., Antesberger, T., & German, R. (2023). Improved sensor error definitions for reliability analysis of multi-sensor systems. In Proceedings of the The 7th International Conference on System Reliability and Safety. Bologna, IT.

MLA:

Qiu, Minhao, et al. "Improved sensor error definitions for reliability analysis of multi-sensor systems." Proceedings of the The 7th International Conference on System Reliability and Safety, Bologna 2023.

BibTeX: Download