Show-Me State Times

University of Missouri researchers are using advanced technology in their efforts to improve road safety, specifically targeting vulnerable road users like pedestrians and cyclists. Led by Associate Professor Yaw Adu-Gyamfi and graduate student Linlin Zhang from Mizzou’s College of Engineering, a recent study introduced a new method for examining the interaction between pedestrians, cyclists, and vehicles at traffic signals. This method combines light detection and ranging (lidar) with artificial intelligence (AI) to address transportation safety and mobility issues.

Lidar technology employs a camera and laser system to create a 3D view, enabling the measurement of distances and speeds of objects such as bicycles, cars, and people. “By having a better understanding of how pedestrians and cyclists interact with each other on the roads, this study will help us design advanced systems that will allow vehicles to better understand and avoid other road users. This is important especially as autonomous vehicles become more common,” said Adu-Gyamfi.

The lack of available industry data on interactions between cyclists, pedestrians, and vehicles at traffic signals is something this research aims to address. The technology can identify near-misses between vehicles and pedestrians, providing insights to prevent accidents. As it becomes more accessible, it could also monitor how vehicles and people approach intersections, sharing information with vehicles to enhance safety. “This approach would require working with car manufacturers to build the technology into vehicles,” noted Adu-Gyamfi. “In fact, some cars already connect with traffic systems using networks like cellular vehicle-to-everything (C-V2X).”

The system's data could improve transportation in additional ways, such as determining appropriate pedestrian signal times and monitoring cars for speeding in work zones. It also has the capability to detect pavement issues like pothole depths.

Researchers deployed a joint camera and lidar system at an intersection to track traffic flow for the project. They optimized the technology to function with a single lidar unit instead of two. Using a method called point cloud completion, they improved the visibility of pedestrians and other objects. “Instead of retraining a machine learning model to detect objects, we used a pre-trained one and created a new algorithm to estimate an object's height and width,” explained Adu-Gyamfi, adding that this approach allowed for more accurate object classification.

However, before this technology can be widely adopted on roads and highways, challenges with data processing, power supply, and weather conditions need to be addressed. The study, titled “Three-Dimensional Object Detection and High-Resolution Traffic Parameter Extraction Using Low-Resolution LiDAR Data,” was published in the Journal of Transportation Engineering and co-authored by Xiang Yu at Mizzou and Armstrong Aboah at North Dakota State University.