Showing papers by "Paul Newman published in 2023"
TL;DR: In this article , the authors propose a multi-session map management system which constructs the best maps for further localisation based on learned variance properties in an embedding space, and propose a new way to introspectively reject localisation queries that are likely to be incorrect.
Abstract: Localisation with Frequency-Modulated Continuous-Wave (FMCW) radar has gained increasing interest due to its inherent resistance to challenging environments. However, complex artefacts of the radar measurement process require appropriate uncertainty estimation to ensure the safe and reliable application of this promising sensor modality. In this work, we propose a multi-session map management system which constructs the best maps for further localisation based on learned variance properties in an embedding space. Using the same variance properties, we also propose a new way to introspectively reject localisation queries that are likely to be incorrect. For this, we apply robust noise-aware metric learning, which both leverages the short-timescale variability of radar data along a driven path (for data augmentation) and predicts the downstream uncertainty in metric-space-based place recognition. We prove the effectiveness of our method over extensive cross-validated tests of the Oxford Radar RobotCar and MulRan dataset. In this, we outperform the current state-of-the-art in radar place recognition and other uncertainty-aware methods when using only single nearest-neighbour queries. We also show consistent performance increases when rejecting queries based on uncertainty over a difficult test environment, which we did not observe for a competing uncertainty-aware place recognition system.
TL;DR: In this article , the authors propose a fast deployment pipeline for visually-servoed robots which does not assume anything about either the robot - e.g. sizes, colour or the presence of markers - or the deployment environment.
Abstract: This work proposes a fast deployment pipeline for visually-servoed robots which does not assume anything about either the robot - e.g. sizes, colour or the presence of markers - or the deployment environment. In this, accurate estimation of robot orientation is crucial for successful navigation in complex environments; manual labelling of angular values is, though, time-consuming and possibly hard to perform. For this reason, we propose a weakly supervised pipeline that can produce a vast amount of data in a small amount of time. We evaluate our approach on a dataset of remote camera images captured in various indoor environments demonstrating high tracking performances when integrated into a fully-autonomous pipeline with a simple controller. With this, we then analyse the data requirement of our approach, showing how it is possible to deploy a new robot in a new environment in less than 30.00 min.
TL;DR: In this article , the authors propose to fit distributions, such as histograms or Gaussians, to activations of neurons in a pre-trained feature extractor through which they pass the image(s) to represent.
Abstract: Selecting appropriate datasets is critical in modern computer vision. However, no general-purpose tools exist to evaluate the extent to which two datasets differ. For this, we propose representing images - and by extension datasets - using Distributions of Neuron Activations (DNAs). DNAs fit distributions, such as histograms or Gaussians, to activations of neurons in a pre-trained feature extractor through which we pass the image(s) to represent. This extractor is frozen for all datasets, and we rely on its generally expressive power in feature space. By comparing two DNAs, we can evaluate the extent to which two datasets differ with granular control over the comparison attributes of interest, providing the ability to customise the way distances are measured to suit the requirements of the task at hand. Furthermore, DNAs are compact, representing datasets of any size with less than 15 megabytes. We demonstrate the value of DNAs by evaluating their applicability on several tasks, including conditional dataset comparison, synthetic image evaluation, and transfer learning, and across diverse datasets, ranging from synthetic cat images to celebrity faces and urban driving scenes.
TL;DR: In this paper , a method for solving the localisation of a ground lidar using overhead imagery only is presented, which can be used as the map proxy for robot localisation.
Abstract: Abstract In this paper, we present a method for solving the localisation of a ground lidar using overhead imagery only. Public overhead imagery such as Google satellite images are readily available resources. They can be used as the map proxy for robot localisation, relaxing the requirement for a prior traversal for mapping as in traditional approaches. While prior approaches have focused on the metric localisation between range sensors and overhead imagery, our method is the first to learn both place recognition and metric localisation of a ground lidar using overhead imagery, and also outperforms prior methods on metric localisation with large initial pose offsets. To bridge the drastic domain gap between lidar data and overhead imagery, our method learns to transform an overhead image into a collection of 2D points, emulating the resulting point-cloud scanned by a lidar sensor situated near the centre of the overhead image. After both modalities are expressed as point sets, point-based machine learning methods for localisation are applied.