Please check our code in Existential Robotics Laboratory (ERL) GitHub Organization. This repository is a PyTorch implementation for paper Policy Learning for Active Target Tracking over Continuous SE(3) Trajectories in L4DC 2023. Authors: Pengzhi Yang, Shumon Koga, Arash Asgharivaskasi, Nikolay Atanasov. If you are using the code for research work, please cite:
@inproceedings{yang2023l4dc,
title={Policy Learning for Active Target Tracking over Continuous SE(3) Trajectories},
author={Yang, Pengzhi and Koga, Shumon and Asgharivaskasi, Arash and Atanasov, Nikolay},
booktitle={Learning for Dynamics and Control (L4DC)},
year={2023}
}
This paper proposes a novel model-based policy gradient algorithm for tracking dynamic targets using a mobile robot, equipped with an onboard sensor with limited field of view. The task is to obtain a continuous control policy for the mobile robot to collect sensor measurements that reduce uncertainty in the target states, measured by the target distribution entropy. We design a neural network control policy with the robot SE(3) pose and the mean vector and information matrix of the joint target distribution as inputs and attention layers to handle variable numbers of targets. We also derive the gradient of the target entropy with respect to the network parameters explicitly, allowing efficient model-based policy gradient optimization.
Clone the repository and cd into it,
conda create -n landmark_mapping python==3.8 -y
conda activate landmark_mapping
pip install -r requirements.txt
cd into the model_based_active_mapping directory and:
python run_model_based_training.py
python run_model_based_test.py
