Skip to content

Latest commit

 

History

History
40 lines (30 loc) · 1.3 KB

README.md

File metadata and controls

40 lines (30 loc) · 1.3 KB

DPGP for Multi-Vehicle Interaction Scenario Extraction

The clustering results on NGSIM and Argoverse are coming soon.
This repo provides the python implementation of DPGP algorithm using Gaussian Process to represent multi-vehicle driving scenarios with Dirichlet Process adapting cluster numbers.
The python version code is implemented by Mengdi Xu, mengdixu@andrew.cmu.edu @SafeAI lab in CMU.
Initial MATLAB code implemented by Yaohui Guo and Vinay Varma Kalidindi.

Paper Reference:

Modeling Multi-Vehicle Interaction Scenarios Using Gaussian Random Field
https://arxiv.org/pdf/1906.10307.pdf

Improvement:

(a) fixed several bugs in the MATLAB version of code.
(b) The code structure is more clear and can easily be implemented for various applications.
Thanks members of SafeAI lab for discussion!

Input:

frames: list with element as object defined in frame.py

Output:

Mixture model as defined in mixtureModel.py

Implement:

Train DPGP: python main_argo.py
Visualization: python pattern_vis.py

Required python packages:

argoverse (for lane visualizaiton)
numpy == 1.16.4
scipy == 1.3.1
scikit-learn == 0.21.2
pandas == 0.25.0