A Reinforcement Learning Approach to the View Planning Problem

Kaba, Mustafa Devrim; Uzunbaş, Mustafa Gökhan; Lim, Ser-Nam

doi:10.1109/cvpr.2017.541

Cited by 42 publications

(32 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The most widely used categorisation [25] classifies approaches as either scene-model-based or scene-model-free. Model-based approaches [1,2] require an a priori scene model and do not generalise well. Within the class of model-free approaches there are global, volumetric and surface representations.…”

Section: Related Workmentioning

confidence: 99%

Surface Edge Explorer (see): Planning Next Best Views Directly from 3D Observations

Border

Gammell

Newman

2018

2018 IEEE International Conference on Robotics and Automation (ICRA)

View full text Add to dashboard Cite

Surveying 3D scenes is a common task in robotics. Systems can do so autonomously by iteratively obtaining measurements. This process of planning observations to improve the model of a scene is called Next Best View (NBV) planning.NBV planning approaches often use either volumetric (e.g., voxel grids) or surface (e.g., triangulated meshes) representations. Volumetric approaches generalise well between scenes as they do not depend on surface geometry but do not scale to high-resolution models of large scenes. Surface representations can obtain high-resolution models at any scale but often require tuning of unintuitive parameters or multiple survey stages.This paper presents a scene-model-free NBV planning approach with a density representation. The Surface Edge Explorer (SEE) uses the density of current measurements to detect and explore observed surface boundaries. This approach is shown experimentally to provide better surface coverage in lower computation time than the evaluated state-of-the-art volumetric approaches while moving equivalent distances.

show abstract

Section: Related Workmentioning

confidence: 99%

Surface Edge Explorer (see): Planning Next Best Views Directly from 3D Observations

Border

Gammell

Newman

2018

2018 IEEE International Conference on Robotics and Automation (ICRA)

View full text Add to dashboard Cite

show abstract

“…In addition to lattice-like structures for view poses, we also evaluate a random sampling of view poses as done by [ 11 , 13 ]. Therefore, we use the previously defined limits

and

to construct a box.…”

Section: Proposed Architecture (Methods)mentioning

confidence: 99%

“…[ 9 ] proposed an evolutionary search algorithm and [ 10 ] applied linear programming to solve the VPP. As outlined by [ 11 ], these methods lack performance gains and efficiency over simple greedy algorithms.…”

Section: Introductionmentioning

confidence: 99%

A Reinforcement Learning Approach to View Planning for Automated Inspection Tasks

Landgraf

Meese

Pabst

et al. 2021

Sensors

View full text Add to dashboard Cite

Manual inspection of workpieces in highly flexible production facilities with small lot sizes is costly and less reliable compared to automated inspection systems. Reinforcement Learning (RL) offers promising, intelligent solutions for robotic inspection and manufacturing tasks. This paper presents an RL-based approach to determine a high-quality set of sensor view poses for arbitrary workpieces based on their 3D computer-aided design (CAD). The framework extends available open-source libraries and provides an interface to the Robot Operating System (ROS) for deploying any supported robot and sensor. The integration into commonly used OpenAI Gym and Baselines leads to an expandable and comparable benchmark for RL algorithms. We give a comprehensive overview of related work in the field of view planning and RL. A comparison of different RL algorithms provides a proof of concept for the framework’s functionality in experimental scenarios. The obtained results exhibit a coverage ratio of up to 0.8 illustrating its potential impact and expandability. The project will be made publicly available along with this article.

show abstract

“…With the rise of deep learning, active vision problems has also been tackled through learning-based approaches. The problem has been cast as a set covering optimization problem in which a reinforcement learning agent has to select the least amount of views to observe the area (Devrim Kaba et al, 2017). This approach assumes that the area is known in advance, and that an agent can be trained on this.…”

Section: Related Workmentioning

confidence: 99%

Active Vision for Robot Manipulators Using the Free Energy Principle

et al. 2021

View full text Add to dashboard Cite

Occlusions, restricted field of view and limited resolution all constrain a robot's ability to sense its environment from a single observation. In these cases, the robot first needs to actively query multiple observations and accumulate information before it can complete a task. In this paper, we cast this problem of active vision as active inference, which states that an intelligent agent maintains a generative model of its environment and acts in order to minimize its surprise, or expected free energy according to this model. We apply this to an object-reaching task for a 7-DOF robotic manipulator with an in-hand camera to scan the workspace. A novel generative model using deep neural networks is proposed that is able to fuse multiple views into an abstract representation and is trained from data by minimizing variational free energy. We validate our approach experimentally for a reaching task in simulation in which a robotic agent starts without any knowledge about its workspace. Each step, the next view pose is chosen by evaluating the expected free energy. We find that by minimizing the expected free energy, exploratory behavior emerges when the target object to reach is not in view, and the end effector is moved to the correct reach position once the target is located. Similar to an owl scavenging for prey, the robot naturally prefers higher ground for exploring, approaching its target once located.

show abstract

A Reinforcement Learning Approach to the View Planning Problem

Cited by 42 publications

References 29 publications

Surface Edge Explorer (see): Planning Next Best Views Directly from 3D Observations

Surface Edge Explorer (see): Planning Next Best Views Directly from 3D Observations

A Reinforcement Learning Approach to View Planning for Automated Inspection Tasks

Active Vision for Robot Manipulators Using the Free Energy Principle

Contact Info

Product

Resources

About