Robotics: Science and Systems IX 2013
DOI: 10.15607/rss.2013.ix.004
|View full text |Cite
|
Sign up to set email alerts
|

Learning Semantic Maps from Natural Language Descriptions

Abstract: Abstract-This paper proposes an algorithm that enables robots to efficiently learn human-centric models of their environment from natural language descriptions. Typical semantic mapping approaches augment metric maps with higher-level properties of the robot's surroundings (e.g., place type, object locations), but do not use this information to improve the metric map. The novelty of our algorithm lies in fusing high-level knowledge, conveyed by speech, with metric information from the robot's low-level sensor … Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1
1

Citation Types

0
71
0

Year Published

2014
2014
2020
2020

Publication Types

Select...
5
1
1

Relationship

2
5

Authors

Journals

citations
Cited by 60 publications
(71 citation statements)
references
References 21 publications
0
71
0
Order By: Relevance
“…We describe an approach first presented by the authors [50] that enables robots to efficiently learn human-centric models of the environment from a narrated, guided tour ( Fig. 1) by fusing knowledge inferred from natural language descriptions with conventional low-level sensor data.…”
Section: Introductionmentioning
confidence: 99%
“…We describe an approach first presented by the authors [50] that enables robots to efficiently learn human-centric models of the environment from a narrated, guided tour ( Fig. 1) by fusing knowledge inferred from natural language descriptions with conventional low-level sensor data.…”
Section: Introductionmentioning
confidence: 99%
“…Boroditsky (2001) and Walter et al (2013)). The laserbot first runs Gmapping (also available as part of ROS) to produce an occupancy grid as output and then AMCL to allow it to localize using the occupancy grid (Fox et al, 1999).…”
Section: Methodsmentioning
confidence: 98%
“…SLAM systems are the state-of-the-art mapping systems for mobile robots, providing them with the capability to learn allocentric spatial representations. Symbol grounding within SLAM systems allows robots to name features of advanced spatial representations (Jung and Zelinsky, 2000;Schulz et al, 2011a;Walter et al, 2013). For spatial grounding, SLAM representations allow agents to name places other than the "here and now" and paths through the environment (Tellex et al, 2011).…”
Section: Design Choicesmentioning
confidence: 99%
See 2 more Smart Citations