Dataset on personal mobility vehicle's regular riding and fall events

Area of Science:

• Urban planning and transportation engineering
• Data science and machine learning
• Public safety and emergency response

Background:

• Personal mobility vehicles (PMVs) like electric scooters are increasingly common in urban areas.
• The integration of PMVs into smart cities presents challenges, including safety concerns and accident management.
• Rapid detection of accidents involving PMVs is critical for timely emergency services intervention.

Purpose of the Study:

• To present a novel dataset capturing sensor data from electric scooters during normal operation and various accident scenarios.
• To facilitate the development of advanced algorithms for automatic detection and classification of riding situations and accidents.
• To support research in enhancing the safety and integration of PMVs in smart city environments.

Main Methods:

• Collection of sensor data (e.g., accelerometer, gyroscope) from an electric scooter.
• Simulating and recording diverse accident types, including frontal collisions and lateral falls.
• Organizing and annotating the collected data for use in machine learning models.

Main Results:

• The dataset encompasses a variety of real-world riding conditions and accident scenarios.
• Provides a comprehensive resource for training and validating accident detection algorithms.
• Demonstrates the potential for sensor data to differentiate between normal and critical events.

Conclusions:

• The developed dataset is a valuable resource for advancing the field of automatic accident detection for personal mobility vehicles.
• Enabling faster response times can significantly improve safety outcomes for scooter and e-bike users.
• This work contributes to the broader goal of creating safer and more efficient smart city transportation ecosystems.