SPT: Obtaining long trajectory data of disordered traffic using a swarm of unmanned aerial vehicles

• Developed a unique long-trajectory dataset for disordered traffic using a swarm of UAVs. • Used deep learning to build an automated tool for vehicle detection, classification, and tracking. • Revealed distinct behavioral patterns of different vehicle types under disordered traffic. Trajectory data is essential for developing algorithms and models that capture driver behavior at a microscopic level. The presence of trajectory datasets under disordered traffic conditions (a wide mix of vehicle types with the absence of lane discipline) is very limited. Existing datasets primarily cover shorter road sections, limiting the understanding of driving behaviour. This paper presents SPT (Tamil: சீரற்ற போக்குவரத்து தரவு/English: Disordered Traffic Data), which is the design and preliminary results of an experiment to create a long trajectory dataset for an urban arterial road in Chennai city, India, under disordered traffic conditions using a swarm of six Unmanned Aerial Vehicles (UAVs). The steps followed to obtain the detailed trajectory data from UAV traffic videos are 1) Georegistration, 2) Video Stabilization and Stitching, 3) Data Annotation, 4) Vehicle Detection, and 5) Vehicle Tracking. Finally, to remove the noises from the trajectory dataset, a Symmetric Exponential Moving Average (sEMA) filter technique was applied to smooth the positions of the vehicles and the first and second derivatives of the positions, i.e., speeds and accelerations were obtained using the central difference method. The resulting trajectory dataset was used to analyze the driver’s lateral position preferences, which highlights the unique behavior of different vehicle categories in disordered traffic conditions. The SPT is significantly larger than existing datasets in disordered traffic, enabling the calibration and validation of car-following, lane-changing models, and fully two-dimensional models.