2018/06/25 by Nikita Jaipuria, Jaipuria, Nikita, Golnaz Habibi +3
Engineering · #Artificial Intelligence (cs.AI) #Autonomous Vehicle Technology and Safety #FOS: Computer and information sciences #Machine Learning (cs.LG) #Machine Learning (stat.ML) #Robotics (cs.RO) #Traffic Prediction and Management Techniques #Traffic and Road Safety
paper · pdf · doi:10.48550/arxiv.1806.09444
openalex publication_date 2018/06/25 · openalex created_date 2022/10/01 · openalex updated_date 2026/07/28
This paper presents a novel framework for accurate pedestrian intent\nprediction at intersections. Given some prior knowledge of the curbside\ngeometry, the presented framework can accurately predict pedestrian\ntrajectories, even in new intersections that it has not been trained on. This\nis achieved by making use of the contravariant components of trajectories in\nthe curbside coordinate system, which ensures that the transformation of\ntrajectories across intersections is affine, regardless of the curbside\ngeometry. Our method is based on the Augmented Semi Nonnegative Sparse Coding\n(ASNSC) formulation and we use that as a baseline to show improvement in\nprediction performance on real pedestrian datasets collected at two\nintersections in Cambridge, with distinctly different curbside and crosswalk\ngeometries. We demonstrate a 7.2% improvement in prediction accuracy in the\ncase of same train and test intersections. Furthermore, we show a comparable\nprediction performance of TASNSC when trained and tested in different\nintersections with the baseline, trained and tested on the same intersection.\n