An Explainable Attention-Augmented LSTM Model for Robust Real-Time Detection of Driving Behavior Patterns and Road Anomalies in Smart Transportation Networks
Bibliographic record
Abstract
Accurate measurement and monitoring of vehicle dynamics and road surface conditions are critical for ensuring road safety and improving transportation infrastructure maintenance. This paper presents CV-Net (Connected Vehicle Network), a novel instrumentation-oriented measurement and monitoring framework that integrates smartphone-based multi-sensor instrumentation with an attention-augmented Long Short-Term Memory (LSTM) network for real-time detection of driving behavior patterns and road anomalies. Unlike conventional anomaly detection approaches that treat smartphones only as data sources, CV-Net treats the device as a calibrated measurement instrument, incorporating both calibrated and uncalibrated signals from accelerometers, gyroscopes, magnetometers, and GPS modules. A comprehensive calibration protocol and preprocessing pipeline are applied to mitigate sensor bias, drift, and noise, thereby improving measurement repeatability and reliability. The proposed method is validated on both public and real-world datasets, achieving measurement classification accuracies up to 99.18% for driving behavior and 99.94% for road anomaly detection, outperforming existing instrumentation-based monitoring approaches. A rigorous 5-fold cross-validation confirmed the model’s consistency with narrow 95% confidence intervals, and statistically significant results (p < 0.05) validated its measurement reliability and statistical significance. Computational efficiency is demonstrated through reduced inference latency and low MFLOPS, enabling deployment in resource-constrained measurement systems. Model interpretability is enhanced using Local Interpretable Model-Agnostic Explanations (LIME) and permutation importance, providing traceable measurement reasoning. The results establish CV-Net as a robust, efficient, and interpretable measurement and monitoring tool for intelligent transportation networks. Model Code: https://github.com/fariya17/CV_Net.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.002 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".