This is the website for the survey paper “Robotic Learning for Informative Path Planning,” authored by Marija Popović, Joshua Ott, Julius Rückin, and Mykel Kochenderfer. The paper is currently under review at the Robotics and Autonomous Systems journal. The arXiv pre-print can be found here. This website provides a comprehensive catalog of papers reviewed in our survey with publicly available repositories to facilitate future studies in the field.

Abstract

Adaptive informative path planning (AIPP) is important to many robotics applications, enabling mobile robots to efficiently collect useful data about initially unknown environments. In addition, learning-based methods are increasingly used in robotics to enhance adaptability, versatility, and robustness across diverse and complex tasks. Our survey explores research on applying robotic learning to AIPP, bridging the gap between these two research fields. We begin by providing a mathematical problem definition for general AIPP problems. Next, we establish two complementary taxonomies of current work from the perspectives of (i) learning algorithms and (ii) robotic applications. We explore synergies, recent trends, and highlight the benefits of learning-based methods in AIPP frameworks. Finally, we discuss key challenges and promising future directions to enable more generally applicable and robust robotic data-gathering systems through learning. We provide a comprehensive catalog of papers reviewed in our survey, including publicly available repositories, to facilitate future studies in the field.

If you found this work useful for your own research, feel free to cite it.

@article{popovic2024robotic,
  title=,
  author={Popović, Marija and Ott, Joshua and R{\"u}ckin, Julius and Kochendorfer, Mykel J},
  journal={arXiv preprint arXiv:2404.06940},
  year={2024}
}

Survey Overview

Our review includes different aspects of learning (Sec. 4): supervised learning, reinforcement learning, imitation learning, and active learning. We focus on how these techniques can be used for AIPP in robotics. Furthermore, we discuss relevant application domains (Sec. 5), such as environmental monitoring, exploration and search, semantic scene understanding, and active simultaneous localization and mapping (SLAM).

Taxonomy of AIPP Applications

Our taxonomy considers four broad application areas: (i) environmental monitoring; (ii) exploration and search; (iii) semantic scene understanding; and (iv) active SLAM. We also include the utilized learning method: supervised learning; reinforcement learning; imitation learning; and/or active learning. We provide visual summary statistics for our survey according to the application area, learning method, and planning space considered by each paper.

Open-Source Code

Title	Authors	Venue, Year	Code, Latest Commit
CAtNIPP: Context-Aware Attention-based Network for Informative Path Planning	Cao et al.	CoRL, 2022	GitHub, Nov 09, 2022
Adaptive Informative Path Planning Using Deep Reinforcement Learning for UAV-based Active Sensing	Rückin et al.	ICRA, 2022	GitHub, Jul 14, 2022
Learning to Map for Active Semantic Goal Navigation	Georgakis et al.	ICLR, 2022	GitHub, Mar 22, 2022
Embodied Active Domain Adaptation for Semantic Segmentation via Informative Path Planning	Zurbrügg et al.	RA-L, 2022	GitHub, Dec 06, 2022
SC-Explorer: Incremental 3D Scene Completion for Safe and Efficient Exploration Mapping and Planning	Schmid et al.	arXiv, 2022	GitHub, Apr 30, 2024
An Informative Path Planning Framework for Active Learning in UAV-based semantic mapping	Rückin et al.	T-RO, 2023	GitHub, Jan 24, 2024
Fast- and Compute-efficient Sampling-based Local Exploration Planning via Distribution Learning	Schmid et al.	RA-L, 2022	GitHub, Nov 09, 2022
Learning to Learn How to Learn: Self-Adaptive Visual Navigation using Meta-Learning	Wortsman et al.	CVPR, 2019	GitHub, Sep 22, 2019
Autonomous Exploration Under Uncertainty via Deep Reinforcement Learning on Graphs	Chen et al.	IROS, 2020	GitHub, Jul 10, 2021
Deep Reinforcement Learning for Next-Best-View Planning in Agricultural Applications	Zeng et al.	ICRA, 2022	GitHub, Apr 05, 2022
Sequential Bayesian Optimization for Adaptive Informative Path Planning with Multimodal Sensing	Ott et al.	ICRA, 2023	GitHub, Jan 17, 2023
Embedded Stochastic Field Exploration with Micro Diving Agents using Bayesian Optimization guided tree-search and GMRFs	Duecker et al.	IROS, 2021	GitHub, Jul 26, 2021
An informative path planning framework for UAV-based terrain monitoring	Popović et al.	Autonomous Robots, 2020	GitHub, Jul 01, 2020
Deep Reinforcement Learning for Swarm Systems	Hüttenrauch et al.	JMLR, 2019	GitHub, May 15, 2020
Graph Neural Networks for Decentralized Multi-Robot Path Planning	Li et al.	IROS, 2020	GitHub, Jun 29, 2021
Learning Continuous Control Policies for Information-Theoretic Active Perception	Yang et al.	ICRA, 2023	GitHub, May 02, 2023
Learned Map Prediction for Enhanced Mobile Robot Exploration	Shrestha et al.	ICRA, 2019	GitHub, Aug 10, 2020
Online Exploration of Tunnel Networks Leveraging Topological CNN-based World Predictions	Saroya et al.	IROS, 2020	GitHub, Jul 28, 2020
Self-Learning Exploration and Mapping for Mobile Robots via Deep Reinforcement Learning	Chen et al.	AIAA, 2019	GitHub, Sep 15, 2020
Multi-UAV Adaptive Path Planning Using Deep Reinforcement Learning	Westheider et al.	IROS, 2023	GitHub, Apr 17, 2023
Multi-UAV Path Planning for Wireless Data Harvesting With Deep Reinforcement Learning	Bayerlein et al.	IEEE Open Journal of the Communications Society, 2021	GitHub, Jan 13, 2022
Bag of Views: An Appearance-based Approach to Next-Best-View Planning for 3D Reconstruction	Gazani et al.	RA-L, 2023	GitHub, Aug 09, 2023
Next-best-view regression using a 3D convolutional neural network	Vasquez-Gomez et al.	Machine Vision and Applications, 2021	GitHub, Feb 07, 2019
Uncertainty-driven Planner for Exploration and Navigation	Georgakis et al.	ICRA, 2022	GitHub, Aug 02, 2022
Pred-NBV: Prediction-guided Next-Best-View for 3D Object Reconstruction	Dhami et al.	IROS, 2023	GitHub, Mar 02, 2023
Occupancy Anticipation for Efficient Exploration and Navigation	Ramakrishnan et al.	ECCV, 2020	GitHub, Jul 01, 2021
SEER: Safe Efficient Exploration for Aerial Robots using Learning to Predict Information Gain	Tao et al.	ICRA, 2023	GitHub, Feb 20, 2024
Informative Path Planning for Active Learning in Aerial Semantic Mapping	Rückin et al.	IROS, 2022	GitHub, Feb 17, 2023
Semi-Supervised Active Learning for Semantic Segmentation in Unknown Environments Using Informative Path Planning	Rückin et al.	RA-L, 2024	GitHub, Jan 24, 2024
Data-Driven Planning via Imitation Learning	Choudhury et al.	IJRR, 2018	Bitbucket, Jun 13, 2017
NeU-NBV: Next Best View Planning Using Uncertainty Estimation in Image-Based Neural Rendering	Jin et al.	IROS, 2023	GitHub, Dec 16, 2023
ActiveNeRF: Learning where to See with Uncertainty Estimation	Pan et al.	ECCV, 2022	GitHub, Mar 28, 2023
NeurAR: Neural Uncertainty for Autonomous 3D Reconstruction with Implicit Neural Representations	Ran et al.	RA-L, 2023	GitHub, Mar 27, 2023
One-Shot View Planning for Fast and Complete Unknown Object Reconstruction	Pan et al.	arXiv, 2023	GitHub, Dec 25, 2023
How Many Views Are Needed to Reconstruct an Unknown Object Using NeRF?	Pan et al.	arXiv, 2024	GitHub, Jan 29, 2024
An information gain formulation for active volumetric 3D reconstruction	Isler et al.	ICRA, 2016	GitHub, Mar 13, 2018

Funding

This work was partially funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) under Germany’s Excellence Strategy - EXC 2070 – 390732324.