State Machine Approach for Lane Changing Driving Behavior Recognition
DOI: 10.3390/automation1010006
archive: archived pipeline: cataloged verified
Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)
Summary
This paper addresses the need for interpretable and personalized driving behavior recognition models to enhance Advanced Driving Assistance Systems (ADAS). Motivated by the high prevalence of human error in traffic accidents, the authors propose a state machine approach to recognize lane-changing maneuvers. Unlike complex machine learning algorithms such as Support Vector Machines or Hidden Markov Models, which often lack interpretability, this method offers a transparent framework where state transitions are defined by clear, optimized thresholds of environmental and operational variables. The methodology defines three driving states: Lane Keeping (LK), Lane Changing to the Left (LCL), and Lane Changing to the Right (LCR). The model utilizes ten input variables, including steering wheel angle, accelerator and brake pedal positions, indicator status, and Time To Collision (TTC) with surrounding vehicles in six directions. To determine the optimal transition thresholds between these states, the authors employed the Non-dominated Sorting Genetic Algorithm II (NSGA-II). This multi-objective optimization process aimed to maximize Accuracy (ACC) and Detection Rate (DR) while minimizing the False Alarm Rate (FAR). The experimental design involved collecting data from three participants using a SCANeR driving simulator. Each participant performed a 40-minute training drive and a 10-minute testing drive on a four-lane highway scenario, allowing for the development of personalized behavior models. The results demonstrate that the state machine approach effectively recognizes driving behaviors with high precision. When tested on their own data, the models achieved overall accuracy rates between 91.76% and 95.77%. The detection rates for lane keeping ranged from 91.70% to 97.37%, while false alarm rates remained low, particularly for lane-changing maneuvers (e.g., FAR for right lane changes was as low as 0.74%). The study also evaluated the model's generalizability by testing models trained on one driver against data from other drivers, showing consistent performance with overall accuracies generally above 91%. The algorithm proved computationally efficient, requiring only 26–68 seconds for training and 3–5 seconds for testing on a standard office PC. The significance of this work lies in providing a simple, reliable, and interpretable alternative to black-box AI models for ADAS development. By using a trainable state machine optimized via NSGA-II, the approach allows for the creation of personalized driving profiles that can predict individual driver intentions. This interpretability facilitates easier debugging and trust in automated systems, as the specific conditions triggering state changes are explicitly defined. The findings suggest that this method is a viable tool for improving driving safety by enabling systems to better anticipate and assist with human driving decisions.
Key finding
The optimized state machine model achieved high accuracy and detection rates in recognizing lane-changing maneuvers, demonstrating a close fit between calculated and actual driving behaviors across multiple participants.
Methodology
simulator
Sample size: 3
Provenance
The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed. Discovered via author_sweep_intake on 2026-05-28.
| Stage | Outcome | Tool | Model | Prompt | Attempts | Completed |
|---|---|---|---|---|---|---|
| discover | success | author_sweep | — | — | 2 | 2026-05-28 |
| archive | success | openalex | — | — | 9 | 2026-06-06 |
| extract | success | cached | — | — | 3 | 2026-06-10 |
| clean | success | clean | — | — | 1 | 2026-06-04 |
| chunk | success | chunk | — | — | 1 | 2026-06-04 |
| embed | success | embed | Qwen/Qwen3-Embedding-8B | — | 1 | 2026-06-04 |
| enrich | success | — | — | — | 1 | 2026-05-28 |
| promote | success | — | — | — | 1 | 2026-06-04 |
| summarize | success | llm | qwen3.6-27b-prismaquant | summ-v5 | 2 | 2026-06-10 |
| tag | success | vector_similarity | — | — | 15 | 2026-06-11 |
| verify | success | — | — | — | 2 | 2026-06-10 |
Summary generated by qwen3.6-27b-prismaquant on 2026-06-10; verification: verified.
Topics
Ranked by relevance to this paper. Hover a topic for its definition.
Information type
What kind of knowledge this paper contributes, grouped by family — independent of topic (what it is about) and method (how it was studied).
- Theoretical Contribution: computational model