Abstract:Connected and Autonomous Vehicles (CAVs) technology has the potential to transform the transportation system. Although these new technologies have many advantages, the implementation raises significant concerns regarding safety, security, and privacy. Anomalies in sensor data caused by errors or cyberattacks can cause severe accidents. To address the issue, this study proposed an innovative anomaly detection algorithm, namely the LSTM Autoencoder with Gaussian Mixture Model (LAGMM). This model supports anomalous CAV trajectory detection in the real-time leveraging communication capabilities of CAV sensors. The LSTM Autoencoder is applied to generate low-rank representations and reconstruct errors for each input data point, while the Gaussian Mixture Model (GMM) is employed for its strength in density estimation. The proposed model was jointly optimized for the LSTM Autoencoder and GMM simultaneously. The study utilizes realistic CAV data from a platooning experiment conducted for Cooperative Automated Research Mobility Applications (CARMAs). The experiment findings indicate that the proposed LAGMM approach enhances detection accuracy by 3% and precision by 6.4% compared to the existing state-of-the-art methods, suggesting a significant improvement in the field.

What problem does this paper attempt to address?

### What problem does this paper attempt to solve? This paper aims to solve the problem of trajectory anomaly detection in connected and autonomous vehicles (CAVs). Specifically, the authors focus on anomalies caused by sensor errors or cyber - attacks, which may lead to serious traffic accidents, traffic delays, and other safety and privacy issues. #### Main problem background 1. **Cybersecurity threats**: - CAVs rely on V2V (vehicle - to - vehicle) and V2I (vehicle - to - infrastructure) communication technologies, which bring many advantages while also bringing significant security, privacy, and reliability issues. - Various internal and external cyber - attacks (such as message forgery attacks, denial - of - service attacks, etc.) may affect the normal operation of CAVs, leading to serious consequences. 2. **Limitations of existing methods**: - Existing research mainly focuses on improving the security of communication channels or sensors to prevent certain types of attacks, but lacks effective countermeasures for attacks caused by physical access to CAVs themselves (such as controlling vehicle acceleration, braking, etc.). - Many existing anomaly detection algorithms ignore the temporal correlation of trajectory data, which may cause the model to be unable to identify complex time - dependent anomalies. #### Proposed solution To overcome the above challenges, this paper proposes a new anomaly detection algorithm based on LSTM autoencoder and Gaussian Mixture Model (GMM) - LAGMM (LSTM Autoencoder Gaussian Mixture Model). The main features of this model are as follows: - **LSTM autoencoder**: It is used to generate a low - rank representation and calculate the reconstruction error of each input data point. LSTM can capture the long - term dependencies in time - series data, so as to better handle the temporal dynamic characteristics of trajectory data. - **Gaussian Mixture Model (GMM)**: It is used for density estimation tasks. It models the class distribution by learning the parameters of different classes and classifies according to the likelihood of new data points belonging to each class. #### Model optimization and experimental verification - **Joint optimization**: The LAGMM model simultaneously optimizes the LSTM autoencoder and GMM to avoid local sub - optimal solutions and further reduce the reconstruction error. - **Experimental verification**: It is verified by using the real - world CAVs fleet experimental data from the CARMAs project. The results show that the LAGMM method improves the detection accuracy by 3% and the precision by 6.4% compared with the existing state - of - the - art methods. In conclusion, this research is committed to developing an efficient and real - time anomaly detection system to identify and mitigate potential threats in connected and autonomous vehicles, ensuring traffic safety and stable operation.

Anomaly Detection in Connected and Autonomous Vehicle Trajectories Using LSTM Autoencoder and Gaussian Mixture Model

Anomalous State Recognition of Lane-changing Behavior Using a Hybrid Autoencoder Architecture

Anomaly Candidate Identification and Starting Time Estimation of Vehicles from Traffic Videos.

A Hybrid Deep Learning Approach for Driver Anomalous Lane Changing Identification

Anomaly Detection in Automated Vehicles Using Multistage Attention-Based Convolutional Neural Network

Deep Learning-Based Anomaly Detection for Connected Autonomous Vehicles Using Spatiotemporal Information

Anomaly Detection for Automated Vehicles Integrating Continuous Wavelet Transform and Convolutional Neural Network

Mobile Edge Assisted Literal Multi-Dimensional Anomaly Detection of In-Vehicle Network Using LSTM.

Learning Representation for Anomaly Detection of Vehicle Trajectories

Anomaly Detection in Traffic Surveillance Videos with GAN-based Future Frame Prediction

Generative adversarial network for car following trajectory generation and anomaly detection

Driving Anomaly Detection Using Conditional Generative Adversarial Network

GAAD: GAN-enabled Autoencoder for Real-Time Sensor Anomaly Detection and Recovery in Autonomous Driving

Single and Mixed Sensory Anomaly Detection in Connected and Automated Vehicle Sensor Networks

Structural Attention-Based Recurrent Variational Autoencoder for Highway Vehicle Anomaly Detection

Anomaly Detection in Traffic Scenes via Spatial-aware Motion Reconstruction

A Novel Model for Ship Trajectory Anomaly Detection Based on Gaussian Mixture Variational Autoencoder

Anomaly diagnosis of connected autonomous vehicles: A survey

Evaluation of LSTM GAN for Trajectory Prediction in Connected and Autonomous Vehicles

Long Short-Term Memory-Based Human-Driven Vehicle Longitudinal Trajectory Prediction in a Connected and Autonomous Vehicle Environment

Autonomous Anomaly Detection System for Crime Monitoring and Alert Generation