2026-09-19T11:44:10Zhttps://dspace.mit.edu/server/oai/requestoai:dspace.mit.edu:1721.1/1432482022-06-16T03:13:40Zcom_1721.1_7582com_1721.1_7581col_1721.1_131022
Kaelbling, Leslie Pack
Lozano-Pérez, Tomás
Goretkin, Gustavo Nunes
Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science
2022-06-15T13:07:10Z
2022-06-15T13:07:10Z
2022-02
2022-03-04T20:47:51.967Z
https://hdl.handle.net/1721.1/143248
The motion planning problem, of deciding how to move to achieve a goal, is ubiquitous in robotics. In many robotics applications, there is a map of the environment that is generally useful, but typically outdated as it does not include information about unknown obstacles, such as clutter. This thesis addresses the problem of planning for a robot with an onboard obstacle-detection sensor. The planning objective is to remain safe with respect to unknown obstacles by guaranteeing that the robot will not move into any region of the workspace before observing it.
Although much work has addressed a version of this problem in which the field of view of the sensor is a sphere around the robot, we address robots with a limited field of view, which may arise from sensor limitations or self-occlusions in the case of mobile manipulation robots. We provide a formal definition of the problem, which we call Visibility-Aware Motion Planning (VAMP), and several solution methods with different computational trade-offs. We demonstrate the behavior of these planning algorithms in illustrative planar domains. The key to an efficient solution is to aggressively prune paths, while ensuring that the overall search strategy is sound and complete.
We demonstrate that motion planning problems like VAMP benefit from a path-dependent formulation, in which the state at a search node is represented implicitly by the path to that node. The straightforward approach to computing the feasibility of a successor node in such a path-dependent formulation takes time linear in the path length to the node, in contrast to a (possibly very large) constant time for a more typical search formulation. For long-horizon plans, this linear-time computation for each node becomes prohibitive. To improve upon this, we introduce the use of a fully persistent spatial data structure (FPSDS). We apply a FPSDS to VAMP search, by using a nearest-neighbor data structure to perform bounding-volume queries. We demonstrate an asymptotic and practical improvement in the runtime of finding VAMP solutions in large domains. To the best of our knowledge, this is the first use of a fully persistent data structure for accelerating motion planning
Ph.D.
Massachusetts Institute of Technology
In Copyright - Educational Use Permitted
Copyright retained by author(s)
https://rightsstatements.org/page/InC-EDU/1.0/
Visibility-Aware Motion Planning
Thesis
application/pdf
Doctoral
Doctor of Philosophy
Publication
unknown
unknown
<Publication xmlns="https://www.openaire.eu/cerif-profile/1.1/" id="c0bab2f4-980a-4f8a-aa51-e0b13b97471b">
<Type xmlns="https://www.openaire.eu/cerif-profile/vocab/COAR_Publication_Types">http://purl.org/coar/resource_type/c_1843</Type>
<Title>Visibility-Aware Motion Planning</Title>
<PublishedIn>
<Publication>
</Publication>
</PublishedIn>
<PublicationDate>2022-02</PublicationDate>
<Authors>
<Author>
<DisplayName>Goretkin, Gustavo Nunes</DisplayName>
<Affiliation>
<OrgUnit>
</OrgUnit>
</Affiliation>
</Author>
</Authors>
<Editors>
</Editors>
<Publishers>
<Publisher>
<DisplayName>Massachusetts Institute of Technology</DisplayName>
<OrgUnit />
</Publisher>
</Publishers>
<License>https://rightsstatements.org/page/InC-EDU/1.0/</License>
<Abstract>The motion planning problem, of deciding how to move to achieve a goal, is ubiquitous in robotics. In many robotics applications, there is a map of the environment that is generally useful, but typically outdated as it does not include information about unknown obstacles, such as clutter. This thesis addresses the problem of planning for a robot with an onboard obstacle-detection sensor. The planning objective is to remain safe with respect to unknown obstacles by guaranteeing that the robot will not move into any region of the workspace before observing it.
Although much work has addressed a version of this problem in which the field of view of the sensor is a sphere around the robot, we address robots with a limited field of view, which may arise from sensor limitations or self-occlusions in the case of mobile manipulation robots. We provide a formal definition of the problem, which we call Visibility-Aware Motion Planning (VAMP), and several solution methods with different computational trade-offs. We demonstrate the behavior of these planning algorithms in illustrative planar domains. The key to an efficient solution is to aggressively prune paths, while ensuring that the overall search strategy is sound and complete.
We demonstrate that motion planning problems like VAMP benefit from a path-dependent formulation, in which the state at a search node is represented implicitly by the path to that node. The straightforward approach to computing the feasibility of a successor node in such a path-dependent formulation takes time linear in the path length to the node, in contrast to a (possibly very large) constant time for a more typical search formulation. For long-horizon plans, this linear-time computation for each node becomes prohibitive. To improve upon this, we introduce the use of a fully persistent spatial data structure (FPSDS). We apply a FPSDS to VAMP search, by using a nearest-neighbor data structure to perform bounding-volume queries. We demonstrate an asymptotic and practical improvement in the runtime of finding VAMP solutions in large domains. To the best of our knowledge, this is the first use of a fully persistent data structure for accelerating motion planning</Abstract>
<Access xmlns="http://purl.org/coar/access_right"
>
</Access>
</Publication>