The student team from our school won the Transportation Forecasting Competition (TRANSFOR 24) international student competition and their faculty advisor was Professor Pu Ziyuan.
The student team utilized laser radar and video traffic fusion perception technology to extract accurate pedestrian and vehicle trajectories. Based on pedestrian trajectories and surrounding environmental conditions, they constructed a pedestrian intention prediction model based on Dynamic Bayesian Network (DBN) to predict the crossing intentions of pedestrians and vulnerable road users (VRUs). Finally, based on estimated Time to Collision (TTC) and pedestrian intentions, they evaluated the current traffic safety of the intersection. The scheme achieved a 90% accuracy rate in predicting pedestrian intentions and performed well, effectively and reasonably in evaluating the safety of the intersection, earning unanimous praise from judges such as Laura Chace and ultimately winning the championship.
