Reconstruction of an Auto-Rickshaw Frontal Crash using FE Simulation with Validation using Captured Crash Video from India

The three-wheeled "Auto-Rickshaws" [Auto] plays a significant role in road transportation, especially in India. The crash safety and reconstruction studies have been widely used in four-wheelers, whereas the availability of such data for Auto was limited. In recent times, accident data processing from available videos is being utilized to observe the crash scenario. The crash parameters can be given as inputs to the crash analysis. This paper focuses on the process the real-world accident data and study crash characteristics. With limitation in the availability of detailed injuries post-crash, the study was restricted to reconstructing crash kinematics and estimating indicative injuries to the driver. The source of video data is videos of crash available in public domains like YouTube. PYTHON video processing tool has been used to process the set of real-world accident video data. Object detection, Pixel per meter computation and object tracking are the significant steps to process the accident data, from which the collision speed is obtained. The auto-rickshaw CAE model and driver dummy (Adult male 50 percentile) were used in LS DYNA to conduct crash analysis at obtained collision speed. The reconstructed crash with matching kinematics showed that the driver experienced a noticeable amount of impact forces near the neck joint and knees. This methodology is proposed as a step in the direction of understanding occupant safety in auto rickshaws.