Ryan M. Eustice
YOU?
Author Swipe
View article: Correspondence-Free SE(3) Point Cloud Registration in RKHS via Unsupervised Equivariant Learning
Correspondence-Free SE(3) Point Cloud Registration in RKHS via Unsupervised Equivariant Learning Open
This paper introduces a robust unsupervised SE(3) point cloud registration method that operates without requiring point correspondences. The method frames point clouds as functions in a reproducing kernel Hilbert space (RKHS), leveraging S…
View article: RKHS-BA: A Robust Correspondence-Free Multi-View Registration Framework with Semantic Point Clouds
RKHS-BA: A Robust Correspondence-Free Multi-View Registration Framework with Semantic Point Clouds Open
This work reports a novel multi-frame Bundle Adjustment (BA) framework called RKHS-BA. It uses continuous landmark representations that encode RGB-D/LiDAR and semantic observations in a Reproducing Kernel Hilbert Space (RKHS). With a corre…
View article: A Robust Keyframe-Based Visual SLAM for RGB-D Cameras in Challenging Scenarios
A Robust Keyframe-Based Visual SLAM for RGB-D Cameras in Challenging Scenarios Open
The accuracy of RGB-D SLAM systems is sensitive to the image quality, and can be significantly compromised in adverse situations such as when input images are blurry, lacking in texture features, or overexposed. In this paper, based on Con…
View article: Multitask Learning for Scalable and Dense Multilayer Bayesian Map Inference
Multitask Learning for Scalable and Dense Multilayer Bayesian Map Inference Open
This article presents a novel and flexible multitask multilayer Bayesian\nmapping framework with readily extendable attribute layers. The proposed\nframework goes beyond modern metric-semantic maps to provide even richer\nenvironmental inf…
View article: Energy-Based Legged Robots Terrain Traversability Modeling via Deep Inverse Reinforcement Learning
Energy-Based Legged Robots Terrain Traversability Modeling via Deep Inverse Reinforcement Learning Open
This work reports on developing a deep inverse reinforcement learning method\nfor legged robots terrain traversability modeling that incorporates both\nexteroceptive and proprioceptive sensory data. Existing works use\nrobot-agnostic exter…
View article: Automatic Targetless Extrinsic Calibration of a 3D Lidar and Camera by Maximizing Mutual Information
Automatic Targetless Extrinsic Calibration of a 3D Lidar and Camera by Maximizing Mutual Information Open
This paper reports on a mutual information (MI) based algorithm for automatic extrinsic calibration of a 3D laser scanner and optical camera system. By using MI as the registration criterion, our method is able to work in situ without the …
View article: Multitask Learning for Scalable and Dense Multilayer Bayesian Map Inference
Multitask Learning for Scalable and Dense Multilayer Bayesian Map Inference Open
This article presents a novel and flexible multitask multilayer Bayesian mapping framework with readily extendable attribute layers. The proposed framework goes beyond modern metric-semantic maps to provide even richer environmental inform…
View article: A New Framework for Registration of Semantic Point Clouds from Stereo and RGB-D Cameras
A New Framework for Registration of Semantic Point Clouds from Stereo and RGB-D Cameras Open
This paper reports on a novel nonparametric rigid point cloud registration framework that jointly integrates geometric and semantic measurements such as color or semantic labels into the alignment process and does not require explicit data…
View article: Characterizing the Uncertainty of Jointly Distributed Poses in the Lie Algebra
Characterizing the Uncertainty of Jointly Distributed Poses in the Lie Algebra Open
An accurate characterization of pose uncertainty is essential for safe autonomous navigation. Early pose uncertainty characterization methods proposed by Smith, Self, and Cheeseman (SCC), used coordinate-based first-order methods to propag…
View article: Monocular Depth Prediction through Continuous 3D Loss
Monocular Depth Prediction through Continuous 3D Loss Open
This paper reports a new continuous 3D loss function for learning depth from monocular images. The dense depth prediction from a monocular image is supervised using sparse LIDAR points, which enables us to leverage available open source da…
View article: A Keyframe-based Continuous Visual SLAM for RGB-D Cameras via Nonparametric Joint Geometric and Appearance Representation
A Keyframe-based Continuous Visual SLAM for RGB-D Cameras via Nonparametric Joint Geometric and Appearance Representation Open
This paper reports on a robust RGB-D SLAM system that performs well in scarcely textured and structured environments. We present a novel keyframe-based continuous visual odometry that builds on the recently developed continuous sensor regi…
View article: Adaptive Continuous Visual Odometry from RGB-D Images
Adaptive Continuous Visual Odometry from RGB-D Images Open
In this paper, we extend the recently developed continuous visual odometry framework for RGB-D cameras to an adaptive framework via online hyperparameter learning. We focus on the case of isotropic kernels with a scalar as the length-scale…
View article: LiDARTag: A Real-Time Fiducial Tag using Point Clouds
LiDARTag: A Real-Time Fiducial Tag using Point Clouds Open
Image-based fiducial markers are widely used in robotics and computer vision problems such as object tracking in cluttered or textureless environments, camera (and multi-sensor) calibration tasks, or vision-based simultaneous localization …
View article: Continuous Direct Sparse Visual Odometry from RGB-D Images
Continuous Direct Sparse Visual Odometry from RGB-D Images Open
This paper reports on a novel formulation and evaluation of visual odometry from RGB-D images.Assuming a static scene, the developed theoretical framework generalizes the widely used direct energy formulation (photometric error minimizatio…
View article: Characterizing the Uncertainty of Jointly Distributed Poses in the Lie\n Algebra
Characterizing the Uncertainty of Jointly Distributed Poses in the Lie\n Algebra Open
An accurate characterization of pose uncertainty is essential for safe\nautonomous navigation. Early pose uncertainty characterization methods proposed\nby Smith, Self, and Cheeseman (SCC), used coordinate-based first-order methods\nto pro…
View article: Guaranteed Globally Optimal Planar Pose Graph and Landmark SLAM via Sparse-Bounded Sums-of-Squares Programming
Guaranteed Globally Optimal Planar Pose Graph and Landmark SLAM via Sparse-Bounded Sums-of-Squares Programming Open
Autonomous navigation requires an accurate model or map of the environment. While dramatic progress in the prior two decades has enabled large-scale SLAM, the majority of existing methods rely on non-linear optimization techniques to find …
View article: Contact-Aided Invariant Extended Kalman Filtering for Robot State Estimation
Contact-Aided Invariant Extended Kalman Filtering for Robot State Estimation Open
Legged robots require knowledge of pose and velocity in order to maintain stability and execute walking paths. Current solutions either rely on vision data, which is susceptible to environmental and lighting conditions, or fusion of kinema…
View article: Continuous Direct Sparse Visual Odometry from RGB-D Images
Continuous Direct Sparse Visual Odometry from RGB-D Images Open
This paper reports on a novel formulation and evaluation of visual odometry from RGB-D images. Assuming a static scene, the developed theoretical framework generalizes the widely used direct energy formulation (photometric error minimizati…
View article: Hybrid Contact Preintegration for Visual-Inertial-Contact State Estimation Using Factor Graphs
Hybrid Contact Preintegration for Visual-Inertial-Contact State Estimation Using Factor Graphs Open
The factor graph framework is a convenient modeling technique for robotic\nstate estimation where states are represented as nodes, and measurements are\nmodeled as factors. When designing a sensor fusion framework for legged robots,\none o…
View article: Contact-Aided Invariant Extended Kalman Filtering for Legged Robot State Estimation
Contact-Aided Invariant Extended Kalman Filtering for Legged Robot State Estimation Open
This paper derives a contact-aided inertial navigation observer for a 3D\nbipedal robot using the theory of invariant observer design. Aided inertial\nnavigation is fundamentally a nonlinear observer design problem; thus, current\nsolution…
View article: Legged Robot State-Estimation Through Combined Forward Kinematic and Preintegrated Contact Factors
Legged Robot State-Estimation Through Combined Forward Kinematic and Preintegrated Contact Factors Open
State-of-the-art robotic perception systems have achieved sufficiently good performance using Inertial Measurement Units (IMUs), cameras, and nonlinear optimization techniques, that they are now being deployed as technologies. However, man…
View article: Sparse Bayesian Inference for Dense Semantic Mapping
Sparse Bayesian Inference for Dense Semantic Mapping Open
Despite impressive advances in simultaneous localization and mapping, dense robotic mapping remains challenging due to its inherent nature of being a high-dimensional inference problem. In this paper, we propose a dense semantic robotic ma…
View article: Gaussian Processes Semantic Map Representation
Gaussian Processes Semantic Map Representation Open
In this paper, we develop a high-dimensional map building technique that incorporates raw pixelated semantic measurements into the map representation. The proposed technique uses Gaussian Processes (GPs) multi-class classification for map …
View article: Multipolicy Decision-Making for Autonomous Driving via Changepoint-based Behavior Prediction
Multipolicy Decision-Making for Autonomous Driving via Changepoint-based Behavior Prediction Open
To operate reliably in real-world traffic, an autonomous car must evaluate the consequences of its potential actions by anticipating the uncertain intentions of other traffic participants.This paper presents an integrated behavioral infere…