Approximate Distributed Spatiotemporal Topic Models for Multi-Robot Terrain Characterization
Unsupervised learning techniques, such as Bayesian topic models, are capable of discovering latent structure directly from raw data. These unsupervised models can endow robots with the ability to learn from their observations without human supervision, and then use the learned models for tasks such...
| Main Authors: | , , , |
|---|---|
| Other Authors: | |
| Format: | Article |
| Language: | English |
| Published: |
Institute of Electrical and Electronics Engineers (IEEE)
2020
|
| Online Access: | https://hdl.handle.net/1721.1/125864 |
| _version_ | 1826190013154459648 |
|---|---|
| author | Doherty, Kevin Flaspohler, Genevieve Elaine Roy, Nicholas Girdhar, Yogesh |
| author2 | Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science |
| author_facet | Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science Doherty, Kevin Flaspohler, Genevieve Elaine Roy, Nicholas Girdhar, Yogesh |
| author_sort | Doherty, Kevin |
| collection | MIT |
| description | Unsupervised learning techniques, such as Bayesian topic models, are capable of discovering latent structure directly from raw data. These unsupervised models can endow robots with the ability to learn from their observations without human supervision, and then use the learned models for tasks such as autonomous exploration, adaptive sampling, or surveillance. This paper extends single-robot topic models to the domain of multiple robots. The main difficulty of this extension lies in achieving and maintaining global consensus among the unsupervised models learned locally by each robot. This is especially challenging for multi-robot teams operating in communication-constrained environments, such as marine robots. We present a novel approach for multi-robot distributed learning in which each robot maintains a local topic model to categorize its observations and model parameters are shared to achieve global consensus. We apply a combinatorial optimization procedure that combines local robot topic distributions into a globally consistent model based on topic similarity, which we find mitigates topic drift when compared to a baseline approach that matches topics naïvely, We evaluate our methods experimentally by demonstrating multi-robot underwater terrain characterization using simulated missions on real seabed imagery. Our proposed method achieves similar model quality under bandwidth-constraints to that achieved by models that continuously communicate, despite requiring less than one percent of the data transmission needed for continuous communication. |
| first_indexed | 2024-09-23T08:33:41Z |
| format | Article |
| id | mit-1721.1/125864 |
| institution | Massachusetts Institute of Technology |
| language | English |
| last_indexed | 2024-09-23T08:33:41Z |
| publishDate | 2020 |
| publisher | Institute of Electrical and Electronics Engineers (IEEE) |
| record_format | dspace |
| spelling | mit-1721.1/1258642022-09-23T12:55:57Z Approximate Distributed Spatiotemporal Topic Models for Multi-Robot Terrain Characterization Doherty, Kevin Flaspohler, Genevieve Elaine Roy, Nicholas Girdhar, Yogesh Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science Woods Hole Oceanographic Institution Unsupervised learning techniques, such as Bayesian topic models, are capable of discovering latent structure directly from raw data. These unsupervised models can endow robots with the ability to learn from their observations without human supervision, and then use the learned models for tasks such as autonomous exploration, adaptive sampling, or surveillance. This paper extends single-robot topic models to the domain of multiple robots. The main difficulty of this extension lies in achieving and maintaining global consensus among the unsupervised models learned locally by each robot. This is especially challenging for multi-robot teams operating in communication-constrained environments, such as marine robots. We present a novel approach for multi-robot distributed learning in which each robot maintains a local topic model to categorize its observations and model parameters are shared to achieve global consensus. We apply a combinatorial optimization procedure that combines local robot topic distributions into a globally consistent model based on topic similarity, which we find mitigates topic drift when compared to a baseline approach that matches topics naïvely, We evaluate our methods experimentally by demonstrating multi-robot underwater terrain characterization using simulated missions on real seabed imagery. Our proposed method achieves similar model quality under bandwidth-constraints to that achieved by models that continuously communicate, despite requiring less than one percent of the data transmission needed for continuous communication. National Science Foundation (Award 1734400) 2020-06-18T17:54:13Z 2020-06-18T17:54:13Z 2019-01 2018-10 2019-10-31T13:09:53Z Article http://purl.org/eprint/type/ConferencePaper 9781538680940 https://hdl.handle.net/1721.1/125864 Doherty, Kevin et al. "Approximate Distributed Spatiotemporal Topic Models for Multi-Robot Terrain Characterization." IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), October 2018, Institute of Electrical and Electronics Engineers, January 2019 © 2018 IEEE en http://dx.doi.org/10.1109/iros.2018.8594442 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) Creative Commons Attribution-Noncommercial-Share Alike http://creativecommons.org/licenses/by-nc-sa/4.0/ application/pdf Institute of Electrical and Electronics Engineers (IEEE) MIT web domain |
| spellingShingle | Doherty, Kevin Flaspohler, Genevieve Elaine Roy, Nicholas Girdhar, Yogesh Approximate Distributed Spatiotemporal Topic Models for Multi-Robot Terrain Characterization |
| title | Approximate Distributed Spatiotemporal Topic Models for Multi-Robot Terrain Characterization |
| title_full | Approximate Distributed Spatiotemporal Topic Models for Multi-Robot Terrain Characterization |
| title_fullStr | Approximate Distributed Spatiotemporal Topic Models for Multi-Robot Terrain Characterization |
| title_full_unstemmed | Approximate Distributed Spatiotemporal Topic Models for Multi-Robot Terrain Characterization |
| title_short | Approximate Distributed Spatiotemporal Topic Models for Multi-Robot Terrain Characterization |
| title_sort | approximate distributed spatiotemporal topic models for multi robot terrain characterization |
| url | https://hdl.handle.net/1721.1/125864 |
| work_keys_str_mv | AT dohertykevin approximatedistributedspatiotemporaltopicmodelsformultirobotterraincharacterization AT flaspohlergenevieveelaine approximatedistributedspatiotemporaltopicmodelsformultirobotterraincharacterization AT roynicholas approximatedistributedspatiotemporaltopicmodelsformultirobotterraincharacterization AT girdharyogesh approximatedistributedspatiotemporaltopicmodelsformultirobotterraincharacterization |