Papers
17
Total Citations
338
H-Index
9
About
Ben Talbot is a leading researcher in robot navigation and semantic mapping, whose work bridges the gap between symbolic human language and autonomous robotic exploration. His major contributions center on enabling robots to understand and navigate unfamiliar environments using only natural language descriptions or semantic cues—a paradigm shift from traditional metric mapping. His highly cited 2016 paper on place categorization and semantic mapping (146 citations) pioneered the use of convolutional networks for environment-agnostic semantic understanding on mobile robots. Talbot also introduced the novel concept of the "abstract map," allowing robots to "imagine" unseen spaces from symbolic descriptions and perform goal-directed exploration, as demonstrated in his 2015 and 2021 works. He co-developed OpenSeqSLAM2.0 (30 citations), a widely used toolbox for robust visual place recognition under changing conditions, and recently proposed Bayesian Controller Fusion (2023, 21 citations), a hybrid control strategy that integrates hand-crafted controllers with deep reinforcement learning. Talbot also leads the BenchBot initiative, a benchmarking suite for reproducible robotics research across simulation and reality. His work has earned over 300 total citations, with impact spanning domestic service robots, healthcare automation, and robust long-term autonomy.
Research Focus
Key Achievements
Top Papers
- 1Place categorization and semantic mapping on a mobile robot146 citations · 2016
- 2
- 3
- 4
- 5
- 6Find my office: Navigating real space from semantic descriptions18 citations · 2016
- 7Robot navigation in unseen spaces using an abstract map14 citations · 2021
- 8Place Categorization and Semantic Mapping on a Mobile Robot10 citations · 2015
- 9
- 10