Developing a Portable Data Acquisition System to Study Road User Behavior
archive: archived pipeline: cataloged verified
Get this paper ↗ (full text — opens at the source; we link to it, we don't host it)
Summary
This report details the development of a Portable Data Acquisition System (PDAQS) designed to study road user behavior and improve traffic safety analysis. The research was motivated by the limitations of traditional data collection methods, such as driving simulators and stationary video cameras, which often fail to capture high-frequency, microscopic vehicle trajectories or operate effectively across diverse environments. By leveraging affordable instrumented vehicle technology, the authors aimed to create a portable system capable of capturing instantaneous vehicle information and driver behavior in natural driving conditions across urban and rural settings. The project involved assembling hardware and developing software to process sensor data. The hardware setup included a Velodyne UltraPuck LiDAR sensor for 360-degree environmental scanning, five Basler IP cameras for visual data, a Garmin GPS unit for location tracking, and a CSS CL2000 On-Board Diagnostics (OBD) sensor for vehicle speed. These components were synchronized and connected to a laptop computer. On the software side, the team developed a modularized framework to process LiDAR and video data. Initial attempts using Density-Based Spatial Clustering of Applications with Noise (DBSCAN) for object detection failed due to poor correlation with video data. Consequently, the team revised the design, implementing a scaled YoloV4 TensorFlow Lite model for 2-D image object detection and integrating the Open3D-ML framework with the PointPillars model for 3-D point cloud detection. This approach significantly reduced memory usage and improved inference speed. The primary results included a functional hardware assembly and two key software tools. First, an alignment tool was developed to manually calibrate raw data, allowing users to adjust point transformation parameters and synchronize LiDAR and camera timing in real time. This tool successfully overlaid LiDAR points onto detected objects in video frames. Second, the team achieved 3-D object detection using the PointPillars model, which identified objects such as cars and pedestrians within point cloud frames. While the detection models performed adequately, the system currently lacks the logic to tie video and LiDAR detections together for efficient object tracking across frames. The significance of this work lies in providing a flexible, portable alternative to fixed roadside sensors for capturing detailed traffic data. The developed PDAQS allows for the collection of high-frequency data on headways, speeds, and road user interactions. However, the authors conclude that further development is required before the system can be applied to real-time projects. Future work must focus on implementing algorithms or machine learning models to fuse 2-D and 3-D detections for robust object tracking, improving detection accuracy, and creating an intuitive graphical user interface for non-technical end users.
Key finding
The project successfully developed a portable data acquisition system and software framework that enables the synchronized detection of 2-D and 3-D road users using fused LiDAR and video data, though additional development is needed for continuous object tracking.
Methodology
lab_experiment
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 bulk_ingest_rosap on 2026-05-23 (6 acquisition events logged).
| Stage | Outcome | Tool | Model | Prompt | Attempts | Completed |
|---|---|---|---|---|---|---|
| discover | success | rosap | — | — | 2 | 2026-05-23 |
| archive | success | — | — | — | 1 | 2026-05-23 |
| extract | success | cached | — | — | 2 | 2026-06-10 |
| clean | success | — | — | — | 1 | 2026-06-01 |
| chunk | success | — | — | — | 1 | 2026-06-01 |
| embed | success | — | — | — | 1 | 2026-06-02 |
| enrich | success | — | — | — | 1 | 2026-05-23 |
| promote | success | — | — | — | 1 | 2026-05-23 |
| summarize | success | llm | qwen3.6-27b-prismaquant | summ-v5 | 3 | 2026-06-10 |
| tag | success | vector_similarity | — | — | 24 | 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).
- Methodological Resource: tool software, dataset resource