Kevin Doherty

About

Kevin Doherty is a robotics researcher whose work sits at the intersection of simultaneous localization and mapping (SLAM), probabilistic inference, and robot perception. His research addresses fundamental challenges in enabling autonomous robots to build rich, reliable representations of their environments — particularly in the face of uncertainty, sparse data, and semantic complexity. Doherty's most influential contributions include pioneering work on multimodal semantic SLAM with probabilistic data association (82 citations), which leverages object detection to produce semantically meaningful maps for robot navigation. His widely read survey on advances in SLAM inference and representation (79 citations) has become an important reference for the broader robotics community. He has also made significant strides in 3D occupancy mapping, developing Bayesian generalized kernel inference methods that allow robots to predict occupancy in unobserved regions from sparse sensor data (50 and 37 citations respectively). More recently, Doherty has tackled sophisticated problems including pose ambiguity in object-based SLAM using multi-hypothesis approaches, spectral measurement sparsification for scalable lifelong SLAM, and self-training frameworks that improve object pose estimation in novel environments. Across his body of work, Doherty consistently advances the mathematical rigor and practical scalability of robot mapping systems, making him a noteworthy figure in modern autonomous navigation research.

Research Focus

Key Achievements

Top Papers

1. 1
   Multimodal Semantic SLAM with Probabilistic Data Association

   82 citations · 2019
2. 2
   Advances in Inference and Representation for Simultaneous Localization and Mapping

   79 citations · 2021
3. 3
   Learning-Aided 3-D Occupancy Mapping With Bayesian Generalized Kernel Inference

   50 citations · 2019
4. 4
   Bayesian generalized kernel inference for occupancy map prediction

   37 citations · 2017
5. 5
   Inference-Enabled Information-Theoretic Exploration of Continuous Action Spaces

   23 citations · 2017
6. 6
   Spectral Measurement Sparsification for Pose-Graph SLAM

   22 citations · 2022
7. 7
   A Multi-Hypothesis Approach to Pose Ambiguity in Object-Based SLAM

   19 citations · 2021
8. 8
   Discrete-Continuous Smoothing and Mapping

   16 citations · 2022
9. 9
   SLAM-Supported Self-Training for 6D Object Pose Estimation

   10 citations · 2022
10. 10
    Approximate Distributed Spatiotemporal Topic Models for Multi-Robot Terrain Characterization

    9 citations · 2018

Key Collaborators

Contact & Links