Int2Planner: An Intention-based Multi-modal Motion Planner for Integrated Prediction and Planning

Xiaolei Chen, Junchi Yan, Wenlong Liao, Tao He, Pai Peng · 2025

Read Int2Planner: An Intention-based Multi-modal Motion Planner for Integrated Prediction and Planning by Xiaolei Chen, Junchi Yan, Wenlong Liao, Tao He, Pai Peng (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2501-12799.txt.

Description

Xiaolei Chen and co-authors unite motion prediction and planning for self-driving cars around route intention, in a 2025 paper. A car must guess what nearby road users will do and plan its path in the same breath, yet most systems split the two jobs, which hides their interaction. The route to the destination gives a steady clue about intent, and their planner, Int2Planner, builds on it.

Instead of fixed intention points, the system takes intention points from the route and grows a trajectory for each, giving several futures to choose among. On a private data set and a public benchmark, the method reached the best performance, and the ablations credit the route intention points. The team also ran the planner on real vehicles for hundreds of kilometres of city driving. It kept interacting with traffic throughout, which they offer as proof the design works outside the lab.

Catalog record

CREATORXiaolei Chen, Junchi Yan, Wenlong Liao, Tao He, Pai Peng
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE40 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2501-12799

Browse the full library · About the Tower of Babel