Anomaly Detection in Autonomous Vehicle's Lidar Sensor Data Using Variational Autoencoders

Nourhane Sboui; Mohamed Hadded; Hakim Ghazzai; Mourad Elhadef; Gianluca Setti

doi:10.1109/VTC2024-Fall63153.2024.10757444

Anomaly Detection in Autonomous Vehicle's Lidar Sensor Data Using Variational Autoencoders

Nourhane Sboui^*, Mohamed Hadded, Hakim Ghazzai, Mourad Elhadef, Gianluca Setti

^*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

1 Scopus citations

Abstract

LiDAR sensor data is essential for autonomous vehicle navigation, traffic flow monitoring, obstacle detection, and passenger safety. However, the reliability of LiDAR data can be compromised by anomalies caused by sensor malfunctions, environmental conditions, or unexpected road events. To address this, detecting anomalies in spatial-temporal (ST) LiDAR data is critical for ensuring safety. This paper proposes a novel low-complexity unsupervised framework named CNN-BiLSTM VAE for anomaly detection (AD) in non-image LiDAR data. The framework combines variational auto-encoder (VAE) reconstruction, CNN for spatial learning, and bidirectional LSTM for time-series learning in a mirror-to-mirror (M2M) architecture. Experimental results show that this method effectively detects anomalies in multidimensional ST LiDAR data, thereby maintaining robustness under various environmental conditions.

Original language	English (US)
Title of host publication	2024 IEEE 100th Vehicular Technology Conference, VTC 2024-Fall - Proceedings
Publisher	Institute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)	9798331517786
DOIs	https://doi.org/10.1109/VTC2024-Fall63153.2024.10757444
State	Published - 2024
Event	100th IEEE Vehicular Technology Conference, VTC 2024-Fall - Washington, United States Duration: Oct 7 2024 → Oct 10 2024

Publication series

Name	IEEE Vehicular Technology Conference
ISSN (Print)	1550-2252

Conference

Conference	100th IEEE Vehicular Technology Conference, VTC 2024-Fall
Country/Territory	United States
City	Washington
Period	10/7/24 → 10/10/24

Bibliographical note

Publisher Copyright:
© 2024 IEEE.

Keywords

Anomaly Detection
Autonomous Driving
LiDAR Data
Sensors
Variational Auto-Encoder

ASJC Scopus subject areas

Computer Science Applications
Electrical and Electronic Engineering
Applied Mathematics

Access to Document

10.1109/VTC2024-Fall63153.2024.10757444

Cite this

Sboui, N., Hadded, M., Ghazzai, H., Elhadef, M., & Setti, G. (2024). Anomaly Detection in Autonomous Vehicle's Lidar Sensor Data Using Variational Autoencoders. In 2024 IEEE 100th Vehicular Technology Conference, VTC 2024-Fall - Proceedings (IEEE Vehicular Technology Conference). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/VTC2024-Fall63153.2024.10757444

@inproceedings{db2b4b29e24d4c7a8f78f3db1df1fb87,

title = "Anomaly Detection in Autonomous Vehicle's Lidar Sensor Data Using Variational Autoencoders",

abstract = "LiDAR sensor data is essential for autonomous vehicle navigation, traffic flow monitoring, obstacle detection, and passenger safety. However, the reliability of LiDAR data can be compromised by anomalies caused by sensor malfunctions, environmental conditions, or unexpected road events. To address this, detecting anomalies in spatial-temporal (ST) LiDAR data is critical for ensuring safety. This paper proposes a novel low-complexity unsupervised framework named CNN-BiLSTM VAE for anomaly detection (AD) in non-image LiDAR data. The framework combines variational auto-encoder (VAE) reconstruction, CNN for spatial learning, and bidirectional LSTM for time-series learning in a mirror-to-mirror (M2M) architecture. Experimental results show that this method effectively detects anomalies in multidimensional ST LiDAR data, thereby maintaining robustness under various environmental conditions.",

keywords = "Anomaly Detection, Autonomous Driving, LiDAR Data, Sensors, Variational Auto-Encoder",

author = "Nourhane Sboui and Mohamed Hadded and Hakim Ghazzai and Mourad Elhadef and Gianluca Setti",

note = "Publisher Copyright: {\textcopyright} 2024 IEEE.; 100th IEEE Vehicular Technology Conference, VTC 2024-Fall ; Conference date: 07-10-2024 Through 10-10-2024",

year = "2024",

doi = "10.1109/VTC2024-Fall63153.2024.10757444",

language = "English (US)",

series = "IEEE Vehicular Technology Conference",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

booktitle = "2024 IEEE 100th Vehicular Technology Conference, VTC 2024-Fall - Proceedings",

address = "United States",

}

Sboui, N, Hadded, M, Ghazzai, H, Elhadef, M & Setti, G 2024, Anomaly Detection in Autonomous Vehicle's Lidar Sensor Data Using Variational Autoencoders. in 2024 IEEE 100th Vehicular Technology Conference, VTC 2024-Fall - Proceedings. IEEE Vehicular Technology Conference, Institute of Electrical and Electronics Engineers Inc., 100th IEEE Vehicular Technology Conference, VTC 2024-Fall, Washington, United States, 10/7/24. https://doi.org/10.1109/VTC2024-Fall63153.2024.10757444

Anomaly Detection in Autonomous Vehicle's Lidar Sensor Data Using Variational Autoencoders. / Sboui, Nourhane; Hadded, Mohamed; Ghazzai, Hakim et al.
2024 IEEE 100th Vehicular Technology Conference, VTC 2024-Fall - Proceedings. Institute of Electrical and Electronics Engineers Inc., 2024. (IEEE Vehicular Technology Conference).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Anomaly Detection in Autonomous Vehicle's Lidar Sensor Data Using Variational Autoencoders

AU - Sboui, Nourhane

AU - Hadded, Mohamed

AU - Ghazzai, Hakim

AU - Elhadef, Mourad

AU - Setti, Gianluca

PY - 2024

Y1 - 2024

N2 - LiDAR sensor data is essential for autonomous vehicle navigation, traffic flow monitoring, obstacle detection, and passenger safety. However, the reliability of LiDAR data can be compromised by anomalies caused by sensor malfunctions, environmental conditions, or unexpected road events. To address this, detecting anomalies in spatial-temporal (ST) LiDAR data is critical for ensuring safety. This paper proposes a novel low-complexity unsupervised framework named CNN-BiLSTM VAE for anomaly detection (AD) in non-image LiDAR data. The framework combines variational auto-encoder (VAE) reconstruction, CNN for spatial learning, and bidirectional LSTM for time-series learning in a mirror-to-mirror (M2M) architecture. Experimental results show that this method effectively detects anomalies in multidimensional ST LiDAR data, thereby maintaining robustness under various environmental conditions.

AB - LiDAR sensor data is essential for autonomous vehicle navigation, traffic flow monitoring, obstacle detection, and passenger safety. However, the reliability of LiDAR data can be compromised by anomalies caused by sensor malfunctions, environmental conditions, or unexpected road events. To address this, detecting anomalies in spatial-temporal (ST) LiDAR data is critical for ensuring safety. This paper proposes a novel low-complexity unsupervised framework named CNN-BiLSTM VAE for anomaly detection (AD) in non-image LiDAR data. The framework combines variational auto-encoder (VAE) reconstruction, CNN for spatial learning, and bidirectional LSTM for time-series learning in a mirror-to-mirror (M2M) architecture. Experimental results show that this method effectively detects anomalies in multidimensional ST LiDAR data, thereby maintaining robustness under various environmental conditions.

KW - Anomaly Detection

KW - Autonomous Driving

KW - LiDAR Data

KW - Sensors

KW - Variational Auto-Encoder

UR - http://www.scopus.com/inward/record.url?scp=85213004723&partnerID=8YFLogxK

U2 - 10.1109/VTC2024-Fall63153.2024.10757444

DO - 10.1109/VTC2024-Fall63153.2024.10757444

M3 - Conference contribution

AN - SCOPUS:85213004723

T3 - IEEE Vehicular Technology Conference

BT - 2024 IEEE 100th Vehicular Technology Conference, VTC 2024-Fall - Proceedings

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 100th IEEE Vehicular Technology Conference, VTC 2024-Fall

Y2 - 7 October 2024 through 10 October 2024

ER -

Anomaly Detection in Autonomous Vehicle's Lidar Sensor Data Using Variational Autoencoders

Abstract

Publication series

Conference

Bibliographical note

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this