arXiv subjects

Saurabh Gupta

Publications and source records attributed to Saurabh Gupta.

At least 37 records · Page 2Linked to original sources

arXiv2025

Opening Articulated Structures in the Real World

What does it take to build mobile manipulation systems that can competently operate on previously unseen objects in previously unseen environments? This work answers this question using opening of articulated structures as a mobile manipulation testbed. Specifically, our focus is on the end-to-end performance on this task without any privileged information, i.e. the robot starts at a location with the novel target articulated object in view, and has to approach the object and successfully open it. We first develop a system for this task, and then conduct 100+ end-to-end system tests across 13 real world test sites. Our large-scale study reveals a number of surprising findings: a) modular systems outperform end-to-end learned systems for this task, even when the end-to-end learned systems are trained on 1000+ demonstrations, b) perception, and not precise end-effector control, is the primary bottleneck to task success, and c) state-of-the-art articulation parameter estimation models developed in isolation struggle when faced with robot-centric viewpoints. Overall, our findings highlight the limitations of developing components of the pipeline in isolation and underscore the need for system-level research, providing a pragmatic roadmap for building generalizable mobile manipulation systems. Videos, code, and models are available on the project website: https://arjung128.github.io/opening-articulated-structures/

cs.RO↗

arXiv2025

Learning Getting-Up Policies for Real-World Humanoid Robots

Automatic fall recovery is a crucial prerequisite before humanoid robots can be reliably deployed. Hand-designing controllers for getting up is difficult because of the varied configurations a humanoid can end up in after a fall and the challenging terrains humanoid robots are expected to operate on. This paper develops a learning framework to produce controllers that enable humanoid robots to get up from varying configurations on varying terrains. Unlike previous successful applications of learning to humanoid locomotion, the getting-up task involves complex contact patterns (which necessitates accurately modeling of the collision geometry) and sparser rewards. We address these challenges through a two-phase approach that induces a curriculum. The first stage focuses on discovering a good getting-up trajectory under minimal constraints on smoothness or speed / torque limits. The second stage then refines the discovered motions into deployable (i.e. smooth and slow) motions that are robust to variations in initial configuration and terrains. We find these innovations enable a real-world G1 humanoid robot to get up from two main situations that we considered: a) lying face up and b) lying face down, both tested on flat, deformable, slippery surfaces and slopes (e.g., sloppy grass and snowfield). This is one of the first successful demonstrations of learned getting-up policies for human-sized humanoid robots in the real world.

cs.RO↗

arXiv2025

How Do I Do That? Synthesizing 3D Hand Motion and Contacts for Everyday Interactions

We tackle the novel problem of predicting 3D hand motion and contact maps (or Interaction Trajectories) given a single RGB view, action text, and a 3D contact point on the object as input. Our approach consists of (1) Interaction Codebook: a VQVAE model to learn a latent codebook of hand poses and contact points, effectively tokenizing interaction trajectories, (2) Interaction Predictor: a transformer-decoder module to predict the interaction trajectory from test time inputs by using an indexer module to retrieve a latent affordance from the learned codebook. To train our model, we develop a data engine that extracts 3D hand poses and contact trajectories from the diverse HoloAssist dataset. We evaluate our model on a benchmark that is 2.5-10X larger than existing works, in terms of diversity of objects and interactions observed, and test for generalization of the model across object categories, action categories, tasks, and scenes. Experimental results show the effectiveness of our approach over transformer & diffusion baselines across all settings.

cs.CV↗

arXiv2025

3D Object Reconstruction with mmWave Radars

This paper presents RFconstruct, a framework that enables 3D shape reconstruction using commercial off-the-shelf (COTS) mmWave radars for self-driving scenarios. RFconstruct overcomes radar limitations of low angular resolution, specularity, and sparsity in radar point clouds through a holistic system design that addresses hardware, data processing, and machine learning challenges. The first step is fusing data captured by two radar devices that image orthogonal planes, then performing odometry-aware temporal fusion to generate denser 3D point clouds. RFconstruct then reconstructs 3D shapes of objects using a customized encoder-decoder model that does not require prior knowledge of the object's bound box. The shape reconstruction performance of RFconstruct is compared against 3D models extracted from a depth camera equipped with a LiDAR. We show that RFconstruct can accurately generate 3D shapes of cars, bikes, and pedestrians.

eess.IV↗

arXiv2025

PhysGen3D: Crafting a Miniature Interactive World from a Single Image

Envisioning physically plausible outcomes from a single image requires a deep understanding of the world's dynamics. To address this, we introduce PhysGen3D, a novel framework that transforms a single image into an amodal, camera-centric, interactive 3D scene. By combining advanced image-based geometric and semantic understanding with physics-based simulation, PhysGen3D creates an interactive 3D world from a static image, enabling us to "imagine" and simulate future scenarios based on user input. At its core, PhysGen3D estimates 3D shapes, poses, physical and lighting properties of objects, thereby capturing essential physical attributes that drive realistic object interactions. This framework allows users to specify precise initial conditions, such as object speed or material properties, for enhanced control over generated video outcomes. We evaluate PhysGen3D's performance against closed-source state-of-the-art (SOTA) image-to-video models, including Pika, Kling, and Gen-3, showing PhysGen3D's capacity to generate videos with realistic physics while offering greater flexibility and fine-grained control. Our results show that PhysGen3D achieves a unique balance of photorealism, physical plausibility, and user-driven interactivity, opening new possibilities for generating dynamic, physics-grounded video from an image.

cs.CV↗

arXiv2025

KISS-SLAM: A Simple, Robust, and Accurate 3D LiDAR SLAM System With Enhanced Generalization Capabilities

Robust and accurate localization and mapping of an environment using laser scanners, so-called LiDAR SLAM, is essential to many robotic applications. Early 3D LiDAR SLAM methods often exploited additional information from IMU or GNSS sensors to enhance localization accuracy and mitigate drift. Later, advanced systems further improved the estimation at the cost of a higher runtime and complexity. This paper explores the limits of what can be achieved with a LiDAR-only SLAM approach while following the "Keep It Small and Simple" (KISS) principle. By leveraging this minimalistic design principle, our system, KISS-SLAM, archives state-of-the-art performances in pose accuracy while requiring little to no parameter tuning for deployment across diverse environments, sensors, and motion profiles. We follow best practices in graph-based SLAM and build upon LiDAR odometry to compute the relative motion between scans and construct local maps of the environment. To correct drift, we match local maps and optimize the trajectory in a pose graph optimization step. The experimental results demonstrate that this design achieves competitive performance while reducing complexity and reliance on additional sensor modalities. By prioritizing simplicity, this work provides a new strong baseline for LiDAR-only SLAM and a high-performing starting point for future research. Further, our pipeline builds consistent maps that can be used directly for further downstream tasks like navigation. Our open-source system operates faster than the sensor frame rate in all presented datasets and is designed for real-world scenarios.

cs.RO↗

arXiv2025

Kinematic-ICP: Enhancing LiDAR Odometry with Kinematic Constraints for Wheeled Mobile Robots Moving on Planar Surfaces

LiDAR odometry is essential for many robotics applications, including 3D mapping, navigation, and simultaneous localization and mapping. LiDAR odometry systems are usually based on some form of point cloud registration to compute the ego-motion of a mobile robot. Yet, few of today's LiDAR odometry systems consider domain-specific knowledge or the kinematic model of the mobile platform during the point cloud alignment. In this paper, we present Kinematic-ICP, a LiDAR odometry system that focuses on wheeled mobile robots equipped with a 3D LiDAR and moving on a planar surface, which is a common assumption for warehouses, offices, hospitals, etc. Our approach introduces kinematic constraints within the optimization of a traditional point-to-point iterative closest point scheme. In this way, the resulting motion follows the kinematic constraints of the platform, effectively exploiting the robot's wheel odometry and the 3D LiDAR observations. We dynamically adjust the influence of LiDAR measurements and wheel odometry in our optimization scheme, allowing the system to handle degenerate scenarios such as feature-poor corridors. We evaluate our approach on robots operating in large-scale warehouse environments, but also outdoors. The experiments show that our approach achieves top performances and is more accurate than wheel odometry and common LiDAR odometry systems. Kinematic-ICP has been recently deployed in the Dexory fleet of robots operating in warehouses worldwide at their customers' sites, showing that our method can run in the real world alongside a complete navigation stack.

cs.RO↗

arXiv2024

Efficient LiDAR Bundle Adjustment for Multi-Scan Alignment Utilizing Continuous-Time Trajectories

Constructing precise global maps is a key task in robotics and is required for localization, surveying, monitoring, or constructing digital twins. To build accurate maps, data from mobile 3D LiDAR sensors is often used. Mapping requires correctly aligning the individual point clouds to each other to obtain a globally consistent map. In this paper, we investigate the problem of multi-scan alignment to obtain globally consistent point cloud maps. We propose a 3D LiDAR bundle adjustment approach to solve the global alignment problem and jointly optimize the available data. Utilizing a continuous-time trajectory allows us to consider the ego-motion of the LiDAR scanner while recording a single scan directly in the least squares adjustment. Furthermore, pruning the search space of correspondences and utilizing out-of-core circular buffer enables our approach to align thousands of point clouds efficiently. We successfully align point clouds recorded with a handheld LiDAR, as well as ones mounted on a vehicle, and are able to perform multi-session alignment.

cs.RO↗

arXiv2024

Learning Smooth Humanoid Locomotion through Lipschitz-Constrained Policies

Reinforcement learning combined with sim-to-real transfer offers a general framework for developing locomotion controllers for legged robots. To facilitate successful deployment in the real world, smoothing techniques, such as low-pass filters and smoothness rewards, are often employed to develop policies with smooth behaviors. However, because these techniques are non-differentiable and usually require tedious tuning of a large set of hyperparameters, they tend to require extensive manual tuning for each robotic platform. To address this challenge and establish a general technique for enforcing smooth behaviors, we propose a simple and effective method that imposes a Lipschitz constraint on a learned policy, which we refer to as Lipschitz-Constrained Policies (LCP). We show that the Lipschitz constraint can be implemented in the form of a gradient penalty, which provides a differentiable objective that can be easily incorporated with automatic differentiation frameworks. We demonstrate that LCP effectively replaces the need for smoothing rewards or low-pass filters and can be easily integrated into training frameworks for many distinct humanoid robots. We extensively evaluate LCP in both simulation and real-world humanoid robots, producing smooth and robust locomotion controllers. All simulation and deployment code, along with complete checkpoints, is available on our project page: https://lipschitz-constrained-policy.github.io.

cs.RO↗

arXiv2024

PhysGen: Rigid-Body Physics-Grounded Image-to-Video Generation

We present PhysGen, a novel image-to-video generation method that converts a single image and an input condition (e.g., force and torque applied to an object in the image) to produce a realistic, physically plausible, and temporally consistent video. Our key insight is to integrate model-based physical simulation with a data-driven video generation process, enabling plausible image-space dynamics. At the heart of our system are three core components: (i) an image understanding module that effectively captures the geometry, materials, and physical parameters of the image; (ii) an image-space dynamics simulation model that utilizes rigid-body physics and inferred parameters to simulate realistic behaviors; and (iii) an image-based rendering and refinement module that leverages generative video diffusion to produce realistic video footage featuring the simulated motion. The resulting videos are realistic in both physics and appearance and are even precisely controllable, showcasing superior results over existing data-driven image-to-video generation works through quantitative comparison and comprehensive user study. PhysGen's resulting videos can be used for various downstream applications, such as turning an image into a realistic animation or allowing users to interact with the image and create various dynamics. Project page: https://stevenlsw.github.io/physgen/

cs.CV↗

arXiv2024

3D Reconstruction of Objects in Hands without Real World 3D Supervision

Prior works for reconstructing hand-held objects from a single image train models on images paired with 3D shapes. Such data is challenging to gather in the real world at scale. Consequently, these approaches do not generalize well when presented with novel objects in in-the-wild settings. While 3D supervision is a major bottleneck, there is an abundance of a) in-the-wild raw video data showing hand-object interactions and b) synthetic 3D shape collections. In this paper, we propose modules to leverage 3D supervision from these sources to scale up the learning of models for reconstructing hand-held objects. Specifically, we extract multiview 2D mask supervision from videos and 3D shape priors from shape collections. We use these indirect 3D cues to train occupancy networks that predict the 3D shape of objects from a single RGB image. Our experiments in the challenging object generalization setting on in-the-wild MOW dataset show 11.6% relative improvement over models trained with 3D supervision on existing datasets.

cs.CV↗

arXiv2024

3D Hand Pose Estimation in Everyday Egocentric Images

3D hand pose estimation in everyday egocentric images is challenging for several reasons: poor visual signal (occlusion from the object of interaction, low resolution & motion blur), large perspective distortion (hands are close to the camera), and lack of 3D annotations outside of controlled settings. While existing methods often use hand crops as input to focus on fine-grained visual information to deal with poor visual signal, the challenges arising from perspective distortion and lack of 3D annotations in the wild have not been systematically studied. We focus on this gap and explore the impact of different practices, i.e. crops as input, incorporating camera information, auxiliary supervision, scaling up datasets. We provide several insights that are applicable to both convolutional and transformer models leading to better performance. Based on our findings, we also present WildHands, a system for 3D hand pose estimation in everyday egocentric images. Zero-shot evaluation on 4 diverse datasets (H2O, AssemblyHands, Epic-Kitchens, Ego-Exo4D) demonstrate the effectiveness of our approach across 2D and 3D metrics, where we beat past methods by 7.4% - 66%. In system level comparisons, WildHands achieves the best 3D hand pose on ARCTIC egocentric split, outperforms FrankMocap across all metrics and HaMeR on 3 out of 6 metrics while being 10x smaller and trained on 5x less data.

cs.CV↗

arXiv2024

Mitigating Perspective Distortion-induced Shape Ambiguity in Image Crops

Objects undergo varying amounts of perspective distortion as they move across a camera's field of view. Models for predicting 3D from a single image often work with crops around the object of interest and ignore the location of the object in the camera's field of view. We note that ignoring this location information further exaggerates the inherent ambiguity in making 3D inferences from 2D images and can prevent models from even fitting to the training data. To mitigate this ambiguity, we propose Intrinsics-Aware Positional Encoding (KPE), which incorporates information about the location of crops in the image and camera intrinsics. Experiments on three popular 3D-from-a-single-image benchmarks: depth prediction on NYU, 3D object detection on KITTI & nuScenes, and predicting 3D shapes of articulated objects on ARCTIC, show the benefits of KPE.

cs.CV↗

arXiv2024

Online Learning-Based Inertial Parameter Identification of Unknown Object for Model-Based Control of Wheeled Humanoids

Identifying the dynamic properties of manipulated objects is essential for safe and accurate robot control. Most methods rely on low noise force torque sensors, long exciting signals, and solving nonlinear optimization problems, making the estimation process slow. In this work, we propose a fast, online learning based inertial parameter estimation framework that enhances model based control. We aim to quickly and accurately estimate the parameters of an unknown object using only the robot's proprioception through end to end learning, which is applicable for real-time system. To effectively capture features in robot proprioception affected by object dynamics and address the challenge of obtaining ground truth inertial parameters in the real world, we developed a high fidelity simulation that uses more accurate robot dynamics through real-to-sim adaptation. Since our adaptation focuses solely on the robot, task-relevant data (e.g., holding an object) is not required from the real world, simplifying the data collection process. Moreover, we address both parametric and non-parametric modeling errors independently using Robot System Identification and Gaussian Processes. We validate our estimator to assess how quickly and accurately it can estimate physically feasible parameters of an manipulated object given a specific trajectory obtained from a wheeled humanoid robot. Our estimator achieves faster estimation speeds (around 0.1 seconds) while maintaining accuracy comparable to other methods. Additionally, our estimator further highlight its benefits in improving the performance of model based control by compensating object's dynamics and reinitializing new equilibrium point of wheeled humanoid

cs.RO↗

arXiv2024

Toward Control of Wheeled Humanoid Robots with Unknown Payloads: Equilibrium Point Estimation via Real-to-Sim Adaptation

Model-based controllers using a linearized model around the system's equilibrium point is a common approach in the control of a wheeled humanoid due to their less computational load and ease of stability analysis. However, controlling a wheeled humanoid robot while it lifts an unknown object presents significant challenges, primarily due to the lack of knowledge in object dynamics. This paper presents a framework designed for predicting the new equilibrium point explicitly to control a wheeled-legged robot with unknown dynamics. We estimated the total mass and center of mass of the system from its response to initially unknown dynamics, then calculated the new equilibrium point accordingly. To avoid using additional sensors (e.g., force torque sensor) and reduce the effort of obtaining expensive real data, a data-driven approach is utilized with a novel real-to-sim adaptation. A more accurate nonlinear dynamics model, offering a closer representation of real-world physics, is injected into a rigid-body simulation for real-to-sim adaptation. The nonlinear dynamics model parameters were optimized using Particle Swarm Optimization. The efficacy of this framework was validated on a physical wheeled inverted pendulum, a simplified model of a wheeled-legged robot. The experimental results indicate that employing a more precise analytical model with optimized parameters significantly reduces the gap between simulation and reality, thus improving the efficiency of a model-based controller in controlling a wheeled robot with unknown dynamics

cs.RO↗

arXiv2024

Benchmarks and Challenges in Pose Estimation for Egocentric Hand Interactions with Objects

We interact with the world with our hands and see it through our own (egocentric) perspective. A holistic 3Dunderstanding of such interactions from egocentric views is important for tasks in robotics, AR/VR, action recognition and motion generation. Accurately reconstructing such interactions in 3D is challenging due to heavy occlusion, viewpoint bias, camera distortion, and motion blur from the head movement. To this end, we designed the HANDS23 challenge based on the AssemblyHands and ARCTIC datasets with carefully designed training and testing splits. Based on the results of the top submitted methods and more recent baselines on the leaderboards, we perform a thorough analysis on 3D hand(-object) reconstruction tasks. Our analysis demonstrates the effectiveness of addressing distortion specific to egocentric cameras, adopting high-capacity transformers to learn complex hand-object interactions, and fusing predictions from different views. Our study further reveals challenging scenarios intractable with state-of-the-art methods, such as fast hand motion, object reconstruction from narrow egocentric views, and close contact between two hands and objects. Our efforts will enrich the community's knowledge foundation and facilitate future hand studies on egocentric hand-object interactions.

cs.CV↗

arXiv2024

Push Past Green: Learning to Look Behind Plant Foliage by Moving It

Autonomous agriculture applications (e.g., inspection, phenotyping, plucking fruits) require manipulating the plant foliage to look behind the leaves and the branches. Partial visibility, extreme clutter, thin structures, and unknown geometry and dynamics for plants make such manipulation challenging. We tackle these challenges through data-driven methods. We use self-supervision to train SRPNet, a neural network that predicts what space is revealed on execution of a candidate action on a given plant. We use SRPNet with the cross-entropy method to predict actions that are effective at revealing space beneath plant foliage. Furthermore, as SRPNet does not just predict how much space is revealed but also where it is revealed, we can execute a sequence of actions that incrementally reveal more and more space beneath the plant foliage. We experiment with a synthetic (vines) and a real plant (Dracaena) on a physical test-bed across 5 settings including 2 settings that test generalization to novel plant configurations. Our experiments reveal the effectiveness of our overall method, PPG, over a competitive hand-crafted exploration method, and the effectiveness of SRPNet over a hand-crafted dynamics model and relevant ablations.

cs.RO↗

arXiv2024

Diffusion Meets DAgger: Supercharging Eye-in-hand Imitation Learning

A common failure mode for policies trained with imitation is compounding execution errors at test time. When the learned policy encounters states that are not present in the expert demonstrations, the policy fails, leading to degenerate behavior. The Dataset Aggregation, or DAgger approach to this problem simply collects more data to cover these failure states. However, in practice, this is often prohibitively expensive. In this work, we propose Diffusion Meets DAgger (DMD), a method to reap the benefits of DAgger without the cost for eye-in-hand imitation learning problems. Instead of collecting new samples to cover out-of-distribution states, DMD uses recent advances in diffusion models to synthesize these samples. This leads to robust performance from few demonstrations. We compare DMD against behavior cloning baseline across four tasks: pushing, stacking, pouring, and shirt hanging. In pushing, DMD achieves 80% success rate with as few as 8 expert demonstrations, where naive behavior cloning reaches only 20%. In stacking, DMD succeeds on average 92% of the time across 5 cups, versus 40% for BC. When pouring coffee beans, DMD transfers to another cup successfully 80% of the time. Finally, DMD attains 90% success rate for hanging shirt on a clothing rack.

cs.RO↗