Reproduce “Unsupervised Driver Behavior Profiling leveraging Recurrent Neural Networks” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Young Ah Choi, Kyung Ho Park, Eunji Park, Huy Kang Kim
The authors present an LSTM driver-profiling system that learns ordinary driving by predicting the next smartphone-sensor feature vector and treating prediction error as an anomaly score. Trained only on normal data, it separates normal driving from five of six tested aggressive behaviors reasonably well. Its average ROC-AUC is 0.8805, but acceleration is weaker; results cover two drivers and one car.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.