Pedestrian Intention Prediction

Pedestrian intention prediction aims to anticipate whether a pedestrian will cross a road, crucial for safe autonomous vehicle navigation. Current research focuses on developing robust models using diverse data sources (e.g., camera images, sensor data, pedestrian pose), employing architectures like graph neural networks, transformers, and spiking neural networks to process spatiotemporal information and improve prediction accuracy, particularly in challenging conditions like adverse weather or low light. This research area is vital for enhancing autonomous vehicle safety and reliability, contributing to the development of more trustworthy and effective intelligent transportation systems.