P
Paul Newman
Researcher at University of Oxford
Publications - 287
Citations - 21374
Paul Newman is an academic researcher from University of Oxford. The author has contributed to research in topics: Mobile robot & Radar. The author has an hindex of 59, co-authored 278 publications receiving 18608 citations. Previous affiliations of Paul Newman include University of Sydney & Carnegie Mellon University.
Papers
More filters
Journal ArticleDOI
A generative framework for fast urban labeling using spatial and temporal context
TL;DR: A multi-level classification framework for the semantic annotation of urban maps as provided by a mobile robot by framing the classification exercise probabilistically takes advantage of an information-theoretic bail-out policy when evaluating class-conditional likelihoods.
Proceedings ArticleDOI
Probably Unknown: Deep Inverse Sensor Modelling Radar
TL;DR: This work proposes to learn an Inverse Sensor Model (ISM) converting a raw radar scan to a grid map of occupancy probabilities using a deep neural network, selfsupervised using partial occupancy labels generated by lidar, allowing a robot to learn about world occupancy from past experience without human supervision.
Proceedings ArticleDOI
Fast Radar Motion Estimation with a Learnt Focus of Attention using Weak Supervision
TL;DR: This paper uses weak supervision to train a focus of attention policy which actively down-samples the measurement stream before data association steps are undertaken, and generates copious annotated measurements which can be used for training a learning algorithm in a weakly-supervised fashion.
Proceedings ArticleDOI
LAPS - localisation using appearance of prior structure: 6-DoF monocular camera localisation using prior pointclouds
Alexander D. Stewart,Paul Newman +1 more
TL;DR: Results are presented which demonstrate the applicability of the inherently cross-modal approach to the localisation of a camera against a lidar pointcloud using data gathered from a road vehicle.
Proceedings ArticleDOI
Closing loops without places
TL;DR: The concept of dynamic bagof-words is introduced, which is a novel form of query expansion based on finding cliques in the landmark co-visibility graph that can improve precision and recall for appearance-based localisation.