• Solutions


June 5, 2024

Automated pedestrian detection

Traffic Technology International | September 2023

Using artificial intelligence to improve pedestrian safety in high risk road environments, such as multilane, high-speed roadways.

Highways, bridges, and tunnels serve as critical infrastructure for the transportation of people and goods but are especially perilous environments for pedestrians. Indeed, most pedestrian fatalities in the United States occur on multilane, high-speed roadways even though the volume of pedestrians on these roadways is relatively low, according to a 2021 study by the Journal of Transport and Land Use (JTLU) which identified fatal pedestrian hot spot corridors. Overall, the number of pedestrian deaths on highways has risen sharply over the past decade. Data released by the National Highway Traffic Safety Administration (NHTSA) assessed that pedestrian fatalities on highways alone increased 60% between 2009 and 2018, averaging over 800 pedestrian deaths per year from 2015 to 2018. The NHTSA has identified several factors that contribute to the high risk of pedestrian accidents in tunnels specifically, including limited visibility, noise levels, and the fact that drivers are often not expecting to encounter pedestrians in tunnels. They have recommended several measures to enhance pedestrian safety: improved lighting, raising driver awareness, and vehicle-based pedestrian detection systems. Many traffic management agencies have invested in camera systems to monitor roadways, but pedestrians can be difficult to detect.

97% The accuracy rate of Citilog’s new Extended Pedestrian Detection solution

  1. Agencies are notified when pedestrians wander onto roadways where they are in danger
    spot manually while monitoring several locations until a more significant incident has occurred. By this time, it may be too late to prevent an accident resulting in injury or worse.
    Where Al comes in Citilog’s new Extended Pedestrian Detection system for highways, tunnels, and bridges is designed to address this problem. This Al solution leverages deep learning to detect pedestrians automatically and accurately on these multilane, high-speed roadways where they are at the greatest risk. Extended Pedestrian Detection is an enhancement to Citilog’s Automatic Incident Detection, which alerts traffic operators to other hazards including wrong way drivers, stopped vehicles, smoke in tunnels, debris, slow vehicles, and congestion.
    Pedestrians pose a particularly unique and deadly challenge on these roadways, which are not primarily designed for pedestrians and often lack sufficient protection for pedestrian safety, such as separated walkways or crossing opportunities. Thus, motorists who have left their vehicles and pedestrians from neighboring communities are at great peril every moment they are in or near lanes of traffic.
    Response times save lives In these scenarios, it is important that traffic operators respond quickly for the safety of those vulnerable pedestrians and drivers alike. Extended Pedestrian Detection has a detection accuracy rate of 97%, reducing false alarms by half, and alerting operators within seconds. The system can track pedestrian trajectory across all

  2. Pedestrians are detected and tracked across all lanes and shoulders with high accuracy
    monitored lanes of traffic and shoulders so operators can visualize their behavior and better understand the situation. It can also raise the level of alert in the case of multi-pedestrian detection. The camera agnostic technology can work with many existing monitoring systems that agencies may already have in place, significantly speeding deployment and lowering costs. The result is an effective automated monitoring system, allowing traffic operators to quickly trigger signage for driver awareness or lane closures, and to notify emergency responders to prevent accidents and save lives. Agencies will also have historical data to identify problem areas that can be improved with better design or safety measures. The Extended Pedestrian Detection system is a valuable tool for improving safety on our roads. It can help to reduce pedestrian accidents, making our roads safer for everyone.