TY - GEN
T1 - Brief industry paper
T2 - 27th IEEE Real-Time and Embedded Technology and Applications Symposium, RTAS 2021
AU - Wang, Ruixuan
AU - Kong, Fanxin
AU - Sudler, Hasshi
AU - Jiao, Xun
N1 - Funding Information:
This paper presents HDAD, an anomaly detection approach based on the emerging HDC. HDAD leverages the inherent correlation among existing sensors to detect any anomaly. We train HDAD using all normal samples and extract the representative patterns for normal samples. Then, for a given testing sample, HDAD first encodes it to an intermediate HV, then decodes it to a reconstructed feature vector, and finally checks the reconstruction error between testing sample and reconstruction result. We use three metrics for measuring the reconstruction errors: MSE, MAE, and cosine similarity. By checking the reconstruction error, HDAD is able to achieve 100% detection accuracy on a real-world vehicle sensors reading dataset. This paper presents the first effort in using HDC for anomaly detection and opens the door for this Fig. 2: Reconstruction result of detection for sensors data (MSE distance) Fig. 3: Reconstruction result of detection for sensors data (MAE distance) potential research direction. Our future work will consider using HDC for clustering and feature extraction, and use it for anomaly detection. Acknowledgments. This work was partially supported by NSF grant #2028889 and NSF #2028740. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation.
Publisher Copyright:
© 2021 IEEE.
PY - 2021/5
Y1 - 2021/5
N2 - As the connectivity of autonomous vehicles keeps growing, it is an accepted fact that they are even more vulnerable to malicious cyber-attacks. Recently, sensor spoofing has become an emerging attack that can compromise vehicle safety as vehicles are equipped with more sensors. Thus, it is critical to validate the sensor readings before utilizing them for future actions. In this paper, we develop HDAD, a hyperdimensional computing-based anomaly detection method. Hyperdimensional computing (HDC) is an emerging brain-inspired computing paradigm that mimics the brain cognition and leverages hyperdimensional vectors with fully distributed holographic representation and (pseudo)randomness. The key idea of HDAD is to use HDC to build encoder and decoder to reconstruct the sensor readings. The anomalous data typically have comparatively higher reconstruction errors than normal sensor readings. We explore three different metrics to measure the reconstruction error including mean squared error, mean absolute error, and cosine similarity. Using a real-world vehicle sensor reading dataset, we demonstrate the feasibility and efficacy of HDAD, opening the door for a new set of anomaly detection algorithm design.
AB - As the connectivity of autonomous vehicles keeps growing, it is an accepted fact that they are even more vulnerable to malicious cyber-attacks. Recently, sensor spoofing has become an emerging attack that can compromise vehicle safety as vehicles are equipped with more sensors. Thus, it is critical to validate the sensor readings before utilizing them for future actions. In this paper, we develop HDAD, a hyperdimensional computing-based anomaly detection method. Hyperdimensional computing (HDC) is an emerging brain-inspired computing paradigm that mimics the brain cognition and leverages hyperdimensional vectors with fully distributed holographic representation and (pseudo)randomness. The key idea of HDAD is to use HDC to build encoder and decoder to reconstruct the sensor readings. The anomalous data typically have comparatively higher reconstruction errors than normal sensor readings. We explore three different metrics to measure the reconstruction error including mean squared error, mean absolute error, and cosine similarity. Using a real-world vehicle sensor reading dataset, we demonstrate the feasibility and efficacy of HDAD, opening the door for a new set of anomaly detection algorithm design.
KW - anomaly detection
KW - brain inspired computing
KW - hyperdimensional computing
UR - http://www.scopus.com/inward/record.url?scp=85108328825&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85108328825&partnerID=8YFLogxK
U2 - 10.1109/RTAS52030.2021.00052
DO - 10.1109/RTAS52030.2021.00052
M3 - Conference contribution
AN - SCOPUS:85108328825
T3 - Proceedings of the IEEE Real-Time and Embedded Technology and Applications Symposium, RTAS
SP - 461
EP - 464
BT - Proceedings - 2021 IEEE 27th Real-Time and Embedded Technology and Applications Symposium, RTAS 2021
PB - Institute of Electrical and Electronics Engineers Inc.
Y2 - 18 May 2021 through 21 May 2021
ER -