PolarPoint-BEV

This is the official pytorch implementation of PolarPoint-BEV.

Setup

Download and setup CARLA 0.9.10.1 (from TCP)

mkdir carla
cd carla
wget https://carla-releases.s3.eu-west-3.amazonaws.com/Linux/CARLA_0.9.10.1.tar.gz
wget https://carla-releases.s3.eu-west-3.amazonaws.com/Linux/AdditionalMaps_0.9.10.1.tar.gz
tar -xf CARLA_0.9.10.1.tar.gz
tar -xf AdditionalMaps_0.9.10.1.tar.gz
rm CARLA_0.9.10.1.tar.gz
rm AdditionalMaps_0.9.10.1.tar.gz
cd ..

Clone this repo and build the environment

git clone https://github.com/lab-sun/PolarPoint-BEV.git
cd PolarPoint-BEV
conda env create -f environment.yml --name PolarPoint-BEV
conda activate PolarPoint-BEV

Dataset

Download the datasets and then extract it in the file of Data

The Control Prediction Module of the XPlan network is firstly pre-trained on the dataset from TCP

To train the XPlan network, please refers to trainset and valset.

Pretrained weights：

• Download the pretrained weights and then extract it in the file of weight
• The link for pretrained weights is weight.

Evaluation

The evaluation is performed in the Carla Simulator.

Step1: Launch the Carla server,

cd CARLA_ROOT
./CarlaUE4.sh --world-port=2000 -opengl

Set the parameters in the leaderboard/scripts/run_evaluation.sh.

Step2: Start the evaluation

sh leaderboard/scripts/run_evaluation.sh

Citation

If you found this code or dataset are useful in your research, please consider citing

@article{feng2024polarpoint,
  title={PolarPoint-BEV: Bird-eye-view Perception in Polar Points for Explainable End-to-end Autonomous Driving},
  author={Feng, Yuchao and Sun, Yuxiang},
  journal={IEEE Transactions on Intelligent Vehicles},
  year={2024},
  publisher={IEEE}
}