Summary of Systematic Review: Anomaly Detection in Connected and Autonomous Vehicles, by J. R. V. Solaas et al.
Systematic Review: Anomaly Detection in Connected and Autonomous Vehicles
by J. R. V. Solaas, N. Tuptuk, E. Mariconti
First submitted to arxiv on: 4 May 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: None
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary This systematic review explores anomaly detection in connected and autonomous vehicles, focusing on AI algorithms like LSTM, CNN, and autoencoders, alongside one-class SVM. A comprehensive database search yielded 2160 articles, with 203 included after rigorous screening. Most models were trained using real-world data, with artificial anomalies injected to test performance. Evaluation metrics used include recall, accuracy, precision, F1-score, and false positive rate, with accuracy, precision, recall, and F1-score being the most frequently employed. The review highlights three key recommendations: incorporating multiple evaluation metrics for comprehensive assessment, sharing models publicly to facilitate collaboration, and developing benchmarking datasets with predefined anomalies or cyberattacks. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Anomaly detection in self-driving cars is crucial for ensuring their safety. This study looked at how AI algorithms are used to detect unusual events. They found that neural networks like LSTM and CNN were most commonly used, along with one-class SVM. To test these models, researchers added fake problems to the data they used to train them. The best way to evaluate these models is by using a combination of metrics, including accuracy and precision. The study suggests that making these models publicly available would help other researchers build on their work and improve the field. It also recommends creating special datasets for testing anomaly detection systems. Finally, it highlights the need for more research on how these systems perform in real-world scenarios. |
Keywords
» Artificial intelligence » Anomaly detection » Cnn » F1 score » Lstm » Precision » Recall
