Microscopic Vehicle Trajectory Datasets from UAV-collected Video for Heterogeneous, Area-Based Urban Traffic

This paper offers openly available microscopic vehicle trajectory (MVT) datasets collected using unmanned aerial vehicles (UAVs) in heterogeneous, area-based urban traffic conditions. Traditional roadside video collection often fails in dense mixed traffic due to occlusion, limited viewing angles, and irregular vehicle movements. UAV-based recording provides a top-down perspective that reduces these issues and captures rich spatial and temporal dynamics. The datasets described here were extracted using the Data from Sky (DFS) platform and validated against manual counts, space mean speeds, and probe trajectories in earlier work. Each dataset contains time-stamped vehicle positions, speeds, longitudinal and lateral accelerations, and vehicle classifications at a resolution of 30 frames per second. Data were collected at six mid-block locations in the national capital region of India, covering diverse traffic compositions and density levels. Exploratory analyses highlight key behavioural patterns, including lane-keeping preferences, speed distributions, and lateral manoeuvres typical of heterogeneous and area-based traffic settings. These datasets are intended as a resource for the global research community to support simulation modelling, safety assessment, and behavioural studies under area-based traffic conditions. By making these empirical datasets openly available, this work offers researchers a unique opportunity to develop, test, and validate models that more accurately represent complex urban traffic environments.

💡 Research Summary

This paper presents and describes a collection of high-resolution Microscopic Vehicle Trajectory (MVT) datasets, openly released for the global research community. The data addresses a critical gap in traffic research by focusing on “heterogeneous, area-based” traffic conditions, typical of many developing countries like India, where diverse vehicle types (cars, motorcycles, auto-rickshaws, trucks) share the road without strict lane discipline.

The core innovation lies in the data collection and extraction methodology. Recognizing the limitations of traditional roadside cameras—such as occlusion, limited field of view, and perspective distortion in dense, mixed traffic—the authors employed Unmanned Aerial Vehicles (UAVs). UAVs provide a stable, top-down (“bird’s-eye”) view, minimizing occlusions and capturing a wide area, which is ideal for tracking the complex lateral movements and interactions characteristic of area-based traffic.

The raw UAV video footage was processed using the “Data from Sky” (DFS) platform, an advanced AI-based video analytics tool. A significant portion of the paper’s contribution is rooted in the prior validation of DFS for this specific context. Earlier work by the authors rigorously assessed DFS’s accuracy in heterogeneous settings by comparing its output against three ground truths: manual classified vehicle counts, space-mean speeds per vehicle category, and GPS trajectories from probe vehicles. This validation confirmed that DFS could achieve reliable accuracy (approaching 98-100% for key metrics) even in disordered traffic, thereby establishing credibility for the datasets derived from it.

The datasets themselves were collected at six mid-block locations in the National Capital Region of India, with varying road widths and traffic compositions. Each dataset, provided as a CSV file, contains detailed frame-by-frame data at 30 frames per second. For every vehicle, the data includes a unique ID, timestamp, local X and Y coordinates in meters, vehicle length and width, instantaneous speed (km/h), longitudinal and lateral acceleration (m/s²), and a classification into one of six types (Car, Motorized Two-Wheeler, Three-Wheeler, Light Commercial Vehicle, Bus, Heavy Commercial Vehicle).

The paper includes exploratory analysis, highlighting the datasets’ ability to reveal typical behavioral patterns in such environments, such as lane-keeping preferences, speed distributions, and frequent lateral maneuvers. By making these rich, validated empirical datasets openly available, the work provides an essential resource for researchers worldwide. It enables the development, testing, and calibration of more realistic traffic simulation models, facilitates advanced safety assessments using surrogate measures, and supports behavioral studies specifically tailored to the complexities of heterogeneous and area-based urban traffic systems, which have been historically underrepresented in existing trajectory data repositories.