A Study on a HMM-Based State Machine Approach for Lane Changing Behavior Recognition
DOI: 10.1109/access.2022.3224012
archive: archived pipeline: cataloged verified
Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)
Summary
This study addresses the critical need for accurate driving behavior recognition in Advanced Driving Assistance Systems (ADAS), motivated by the high prevalence of traffic accidents caused by human error, particularly during lane changes. The authors propose a novel machine learning model that combines an improved Hidden Markov Model (HMM) with a state machine approach to recognize three specific behaviors: lane change to the right (LCR), lane keeping (LK), and lane change to the left (LCL). The primary objective is to enhance estimation accuracy, detection rates, and minimize false alarm rates by optimizing the model’s input processing. The methodology integrates a state machine, which defines discrete states and transition conditions, with an improved HMM that provides stochastic estimations to determine those transitions. To refine the HMM’s input features, a prefilter is applied to quantize dynamic observation variables into segments. Two sets of environmental variables are tested: Model I uses distances, velocity deviations, and current lane; Model II uses Time to Collision (TTC) variables and current lane. The prefilter thresholds for these variables are optimized using the Non-Dominated Sorting Genetic-Algorithm-II (NSGA-II) to maximize accuracy and detection while minimizing false alarms. Data was collected from nine drivers performing 25-minute drives in a five-screen driving simulator configured to mimic a four-lane highway. The dataset was split 70:30 for training and testing, with behaviors labeled based on a 2.5-second duration threshold to ensure accuracy. The results demonstrate that applying the prefilter to TTC variables (Model II) yields superior estimation performance compared to using distance and velocity variables (Model I). The proposed HMM-based state machine model outperforms both an individual improved HMM and a previously developed Artificial Neural Network (ANN)-based state machine approach in terms of accuracy, detection rate, and false alarm rate. The optimization process successfully identified parameter values that generalized well across different drivers, indicating the model’s robustness. The significance of this work lies in its contribution to more reliable and interpretable ADAS components. By combining the stochastic capabilities of HMMs with the structural clarity of state machines, the approach offers a transparent alternative to black-box deep learning models. The findings suggest that TTC-based features, when properly prefiltered and optimized, are highly effective for real-time lane change recognition. This enhances the potential for ADAS to accurately predict driver intentions, thereby improving road safety and enabling more adaptive vehicle assistance systems.
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 | canonical_url | — | — | 17 | 2026-06-09 |
| extract | success | cached | — | — | 2 | 2026-06-09 |
| 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 | 1 | 2026-06-09 |
| tag | success | vector_similarity | — | — | 15 | 2026-06-11 |
| verify | success | — | — | — | 1 | 2026-06-09 |
Summary generated by qwen3.6-27b-prismaquant on 2026-06-09; 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