A Lightweight Hybrid Supervised-and-Q-Learning Model for Real-Time Autonomous Lane-Change Decisions
Bibliographic record
Abstract
Autonomous driving systems rely on the intelligent models for decision-making to deal with various traffic situations like lane changing and car following. Classical rule-based methods cannot cater for this need, while deep reinforcement (RL) approaches require inordinate training in the simulated environments before becoming applicable to real-world behaviors. To bridge this gap, this research proposes a hybrid framework that combines supervised learning with reinforcement learning (Q-learning) for improving lane-changing decisions. The proposed system is first a Deep Neural Networks, it predicts lane changing behavior using detailed structured vehicle data. The input data undergo preprocessing, including dealing with the missing values, standardizing features, and normalizing targets. The trained model then produces a continuous prediction score for possible “Move Left,” “Stay in Lane,” and “Move Right” actions, dynamically determined by percentile-based thresholds. A Q-learning module is then added to improve the decision-making process, which updates the lane changing actions based on a reward system. The epsilon-greedy approach, which strikes a balance between exploration and exploitation is used by the system to improve adaptability over time. Safe lane changes receive rewards, while unnecessary or risky lane changes receive penalties. Hybrid model shows the advanced lane selection method with high accuracy when compared to individual model.
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 | 0.001 | 0.001 |
| 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 | 0.001 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.004 | 0.001 |
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".