Peter Hall

About

Peter Hall is a researcher whose work sits at the intersection of robotics, computer vision, and statistical estimation, with a particular focus on how robots perceive and interact with their environments. His major contributions span two key areas: the mathematical estimation of convex shapes from noisy sensor data, and the development of multi-modal perception systems for robots. Hall’s foundational work on estimating convex sets from support function data (1997, 25 citations) and his extension to sets with corners (1999, 15 citations) provided rigorous statistical frameworks for recovering object shapes from laser-radar measurements, with direct applications in medical imaging and robotic vision. In robotics, Hall pioneered approaches that combine vision and touch for object recognition (2017, 23 citations), addressing the practical challenge of scarce tactile training data. He also developed the VGPN system (2018, 15 citations), which integrates voice commands with pointing gestures for more efficient robot navigation, reducing system overhead by eliminating the need for continuous gesture recognition. His work on probabilistic models for object segmentation (2011, 11 citations) and self-supervised visual classification (2011, 5 citations) further demonstrates his commitment to creating robust, real-world robotic perception systems that can learn and adapt in free environments.

Research Focus

Key Achievements

Top Papers

1. 1
   On the Estimation of a Convex Set from Noisy Data on its Support Function

   25 citations · 1997
2. 2
   Object recognition combining vision and touch

   23 citations · 2017
3. 3
   VGPN: Voice-Guided Pointing Robot Navigation for Humans

   15 citations · 2018
4. 4
   On the estimation of a convex set with corners

   15 citations · 1999
5. 5
   Probabilistic models for robot-based object segmentation

   11 citations · 2011
6. 6
   On the Estimation of a Convex Set From Noisy Data on Its Support Function

   9 citations · 1997
7. 7
   Visual object classification by robots, using on-line, self-supervised learning

   5 citations · 2011

Key Collaborators

Contact & Links