mlsanityData to 14 Oct 2026

Semantically-driven Deep Reinforcement Learning for Inspection Path Planning

arXiv 2505.14443 ↗First posted 20 May 2025cs.ROcited by 6

On the map: Deep reinforcement learning methods → RL for robot navigation and UAV control

Abstract (arXiv)

This paper introduces a novel semantics-aware inspection planning policy derived through deep reinforcement learning. Reflecting the fact that within autonomous informative path planning missions in unknown environments, it is often only a sparse set of objects of interest that need to be inspected, the method contributes an end-to-end policy that simultaneously performs semantic object visual inspection combined with collision-free navigation. Assuming access only to the instantaneous depth map, the associated segmentation image, the ego-centric local occupancy, and the history of past positions in the robot's neighborhood, the method demonstrates robust generalizability and successful crossing of the sim2real gap. Beyond simulations and extensive comparison studies, the approach is verified in experimental evaluations onboard a flying robot deployed in novel environments with previously unseen semantics and overall geometric configurations.

Key facts. Cited by 6 papers in the map (data through 14 Oct 2026). Placed in RL for robot navigation and UAV control, one of 15 directions in Deep reinforcement learning methods. Builds on 22 of its 37 references that are in the map. First posted on arXiv 20 May 2025.

Builds on

22 of 37 references.

References · 22

  1. 2018IMPALA: Scalable Distributed Deep-RL with Importance Weighted Actor-Learner Architecturescited by 883
  2. 2020Object Goal Navigation using Goal-Oriented Semantic Explorationcited by 429
  3. 2016Voxblox: Incremental 3D Euclidean Signed Distance Fields for On-Board MAV Planningcited by 150
  4. 2019Learning Exploration Policies for Navigationcited by 116
  5. 2020Sample Factory: Egocentric 3D Control from Pixels at 100000 FPS with Asynchronous Reinforcement Learningcited by 93
  6. 2020FUEL: Fast UAV Exploration using Incremental Frontier Structure and Hierarchical Planningcited by 48
  7. 2019An Efficient Sampling-based Method for Online Informative Path Planning in Unknown Environmentscited by 40
  8. 2017A Data-driven Model for Interaction-aware Pedestrian Motion Prediction in Object Cluttered Environmentscited by 24
  9. 2022Learning Active Camera for Multi-Object Navigationcited by 21
  10. 2025Aerial Gym Simulator: A Framework for Highly Parallelized Simulation of Aerial Robotscited by 15
  11. 2022Where to Look Next: Learning Viewpoint Recommendations for Informative Trajectory Planningcited by 14
  12. 2024Learning-based Methods for Adaptive Informative Path Planningcited by 10
10 more
  1. 2020Adaptive Informative Path Planning with Multimodal Sensingcited by 9
  2. 2024Neural Control Barrier Functions for Safe Navigationcited by 9
  3. 2021Zero-Shot Reinforcement Learning on Graphs for Autonomous Exploration Under Uncertaintycited by 7
  4. 2022SC-Explorer: Incremental 3D Scene Completion for Safe and Efficient Exploration Mapping and Planningcited by 7
  5. 2023Semantically-enhanced Deep Collision Prediction for Autonomous Navigation using Aerial Robotscited by 7
  6. 2021Enhancing Navigational Safety in Crowded Environments using Semantic-Deep-Reinforcement-Learning-based Navigationcited by 5
  7. 2023Semantics-aware Exploration and Inspection Path Planningcited by 4
  8. 2019Active Learning for UAV-based Semantic Mappingcited by 3
  9. 2022TIGRIS: An Informed Sampling-based Algorithm for Informative Path Planningcited by 3
  10. 2023Autonomous Exploration and General Visual Inspection of Ship Ballast Water Tanks using Aerial Robotscited by 2

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 1; 2026: 5.

6 papers.

  1. 2026Reinforcement Learning for Active Perception in Autonomous Navigationcited by 4
  2. 2025Quantum Grid Path Planning Using Parallel QAOA Circuits Based on Minimum Energy Principlecited by 0
  3. 2026Aerial Inspection Behaviors via RL-based Quadrotor Control for Under-canopy Forest Environmentscited by 0
  4. 2026Fast Human Attention Prediction for Fixation-guided Active Perception in Autonomous Navigationcited by 0
  5. 2026TAPAS: Throughput-adaptive Perception for Autonomous Systemscited by 0
  6. 2026CAVEAT: Recurrent Multimodal Diffusion Planning for Mapless Aerial Explorationcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025Sight Over Site: Perception-Aware Reinforcement Learning for Efficient Robotic Inspection
  2. 2025Curriculum-Based Multi-Tier Semantic Exploration via Deep Reinforcement Learning
  3. 2021Inverse reinforcement learning for autonomous navigation via differentiable semantic mapping and planning
  4. 2024SEEK: Semantic Reasoning for Object Goal Navigation in Real World Inspection Tasks
  5. 2020Learning Navigation Costs from Demonstration with Semantic Observations
  6. 2024Reinforcement Learning for Collision-free Flight Exploiting Deep Collision Encoding
  7. 2024Deep Reinforcement Learning with Dynamic Graphs for Adaptive Informative Path Planning
  8. 2026VLN on the Fly: An Onboard Vision-Language Navigation Stack for Aerial Robots