IDS - JODHA: A Multimodal and Multi-Weather Traffic Dataset for Analyzing Agent Dynamics in Unstructured Environments

The Indian Driving Scenes - Jodhpur Dataset for Heterogeneous Environment Analysis (IDS-JODHA) is a multimodal, multi-weather dataset designed for analyzing agent dynamics within unstructured, heterogeneous Indian driving environments. The study of complex systems is often limited by a lack of empirical data from real-world, chaotic environments. Unstructured urban traffic provides a perfect natural laboratory for this inquiry, but its dynamics have remained difficult to quantify systematically. Collected on the streets of Jodhpur, India, this dataset captures the complete state of the traffic scene, including the dense interactions and motions of heterogeneous agents (vehicles, pedestrians), under both baseline (clear weather) and perturbed (rainy weather) conditions. It comprises over 87,000 high-resolution video frames and 16,000 synchronized LiDAR scans from a low-cost, portable sensor suite. By providing processed, spatially aligned, and temporally synchronized data streams, the dataset serves as the foundational resource for the empirical validation of traffic physics and self-organization theories, enabling researchers to move beyond simulation and probe the real-world interactions that govern one of the most complex collective systems.