Advancing Sidewalk Connectivity Through AI
This project seeks to enhance pedestrian pathways, particularly in transit-served areas, by developing an AI-powered system to assess sidewalk conditions. The focus is on creating smooth, accessible pathways that prioritize the needs of pedestrians, especially the elderly and veterans. By leveraging AI and advanced sensor technology, the system will evaluate critical sidewalk features such as surface quality, curb accessibility, and obstructions, aiming to create safer and more inclusive walking environments.
In collaboration with the Utah Department of Transportation (UDOT), the research team will deploy advanced sensor technology equipped with high-resolution cameras and accelerometers to collect comprehensive sidewalk data. The system will navigate high-pedestrian-demand and transit-adjacent areas, ensuring a smooth and uninterrupted data collection process while measuring surface roughness and unevenness. The collected data will be processed using AI-driven deep learning and computer vision algorithms to rapidly identify and prioritize areas in need of repair or improvement, with a specific emphasis on conditions affecting the elderly and veterans.
The AI framework will analyze accelerometer readings to detect abnormal vibrations caused by rough or uneven surfaces. These readings will be paired with video data, which will be processed by a computer vision model to determine if the cause is related to infrastructure, such as cracks or misaligned concrete blocks. The AI will also be trained to differentiate between non-critical features like tram tracks and actual hazards, ensuring that the system effectively targets areas that pose tripping risks or other accessibility challenges.
By focusing on high-pedestrian areas, the project will prioritize repairs that directly impact the mobility of elderly pedestrians and veterans. The approach ensures that critical issues such as uneven surfaces or obstacles are addressed, improving safety and accessibility. The scalability of this solution allows it to be expanded beyond Utah, demonstrating the potential of AI to enhance pedestrian pathways in urban areas.
Project R9 will involve three partners. The Utah Department of Transportation (UDOT) will contribute as a key partner, providing real-world data and insights for testing the developed sidewalk evaluation method. Arizona State University (ASU) and Florida State University (FSU) will collaborate on AI and data processing techniques to ensure the effectiveness of the proposed system.
TBD