ICRA 2024 论文列表

Transparent objects are ubiquitous in industry, pharmaceuticals, and households. Grasping and manipulating these objects is a significant challenge for robots. Existing methods have difficulty reconstructing complete depth maps for challenging transparent objects, leaving holes in the depth reconstruction. Recent work has shown neural radiance fields (NeRFs) work well for depth perception in scenes with transparent objects, and these depth maps can be used to grasp transparent objects with high accuracy. NeRF-based depth reconstruction can still struggle with especially challenging transparent objects and lighting conditions. In this work, we propose Residual-NeRF, a method to improve depth perception and training speed for transparent objects. Robots often operate in the same area, such as a kitchen. By first learning a background NeRF of the scene without transparent objects to be manipulated, we reduce the ambiguity faced by learning the changes with the new object. We propose training two additional networks: a residual NeRF learns to infer residual RGB values and densities, and a Mixnet learns how to combine background and residual NeRFs. We contribute synthetic and real experiments that suggest Residual-NeRF improves depth perception of transparent objects. The results on synthetic data suggest Residual-NeRF outperforms the baselines with a 46.1 % lower RMSE and a 29.5 % lower MAE. Real-world qualitative experiments suggest Residual-NeRF leads to more robust depth maps with less noise and fewer holes. Website: https://residual-nerf.github.io

This paper presents a bionic foldable wing that imitates the hind wing of ladybirds. Based on the folding mechanism of the hind wing of ladybirds and the theory of origami, the motion model of the bionic foldable wing is established, yield the motion law of the crease angles and the variation relationship between the panels are obtained. Bionic foldable wings utilise shape memory alloy to drive wings to fold, and embedded torsion springs to release energy to realize the function of wing unfolding. In the experiments of the vehicle equipped with foldable wings, the lift and attitude torque of bionic foldable wings are measured by the F/T sensor. The experimental results indicated that its aerodynamic performance is basically close to that of our optimized non-foldable wings. Moreover, the vehicle with foldable wings has been able to overcome gravity to achieve flight, which provides a novel concept for the research on flapping wing.

State estimation is crucial for the performance and safety of numerous robotic applications. Among the suite of estimation techniques, particle filters have been identified as a powerful solution due to their non-parametric nature. Yet, in high-dimensional state spaces, these filters face challenges such as ’particle deprivation’ which hinders accurate representation of the true posterior distribution. This paper introduces a novel resampling-free particle filter designed to mitigate particle deprivation by forgoing the traditional resampling step. This ensures a broader and more diverse particle set, especially vital in high-dimensional scenarios. Theoretically, our proposed filter is shown to offer a near-accurate representation of the desired posterior distribution in high-dimensional contexts. Empirically, the effectiveness of our approach is underscored through a high-dimensional synthetic state estimation task and a 6D pose estimation derived from videos. We posit that as robotic systems evolve with greater degrees of freedom, particle filters tailored for high-dimensional state spaces will be indispensable.

In this work, we propose a novel approach to represent robot geometry as distance fields (RDF) that extends the principle of signed distance fields (SDFs) to articulated kinematic chains. Our method employs a combination of Bernstein polynomials to encode the signed distance for each robot link with high accuracy and efficiency while ensuring the mathematical continuity and differentiability of SDFs. We further leverage the kinematics chain of the robot to produce the SDF representation in joint space, allowing robust distance queries in arbitrary joint configurations. The proposed RDF representation is differentiable and smooth in both task and joint spaces, enabling its direct integration to optimization problems. Additionally, the 0-level set of the robot corresponds to the robot surface, which can be seamlessly integrated into whole-body manipulation tasks. We conduct various experiments in both simulations and with 7-axis Franka Emika robots, comparing against baseline methods, and demonstrating its effectiveness in collision avoidance and whole-body manipulation tasks. Project page: https://sites.google.com/view/lrdf/home

Ground testing is of paramount importance to verify and validate space operations and the associated control algorithms before on-orbit deployment. Although state-of-the-art facilities are capable of reproducing zero-G environment with high degree of fidelity, these infrastructures can be complemented with multi-rotors emulating free flying or free floating conditions, exploiting the similarities and analogies between both domains in terms of floating nature, attitude dynamics, and thrust-wrench relation through the mixer matrix. Furthermore, the effective workspace of the testbed can be extended to the dimensions of the flight area and the coverage of the positioning system. Therefore, this papers introduces a new way to recreate orbital motion within an indoor facility, considering the case study of trajectories derived from the Clohessy–Wiltshire equations. This advancement opens up avenues for replicating close-proximity operations between chaser and target satellites employing fully-actuated multi-rotors that allow decoupling translational and attitude dynamics.

The ability to manipulate objects in desired configurations is a fundamental requirement for robots to complete various practical applications. While certain goals can be achieved by picking and placing the objects of interest directly, object reorientation is needed for precise placement in most of the tasks. In such scenarios, the object must be reoriented and re-positioned into intermediate poses that facilitate accurate placement at the target pose. To this end, we propose a reorientation planning method, ReorientDiff, that utilizes a diffusion model-based approach. The proposed method employs both visual inputs from the scene, and goal-specific language prompts to plan intermediate reorientation poses. Specifically, the scene and language-task information are mapped into a joint scene-task representation feature space, which is subsequently leveraged to condition the diffusion model. The diffusion model samples intermediate poses based on the representation using classifier-free guidance and then uses gradients of learned feasibility-score models for implicit iterative pose-refinement. The proposed method is evaluated using a set of YCB-objects and a suction gripper, demonstrating a success rate of 95.2% in simulation. Overall, we present a promising approach to address the reorientation challenge in manipulation by learning a conditional distribution, which is an effective way to move towards generalizable object manipulation. More results can be found on our website: https://utkarshmishra04.github.io/ReorientDiff.

3D occupancy prediction holds significant promise in the fields of robot perception and autonomous driving, which quantifies 3D scenes into grid cells with semantic labels. Recent works mainly utilize complete occupancy labels in 3D voxel space for supervision. However, the expensive annotation process and sometimes ambiguous labels have severely constrained the usability and scalability of 3D occupancy models. To address this, we present RenderOcc, a novel paradigm for training 3D occupancy models only using 2D labels. Specifically, we extract a NeRF-style 3D volume representation from multi-view images, and employ volume rendering techniques to establish 2D renderings, thus enabling direct 3D supervision from 2D semantics and depth labels. Additionally, we introduce an Auxiliary Ray method to tackle the issue of sparse viewpoints in autonomous driving scenarios, which leverages sequential frames to construct comprehensive 2D rendering for each object. To our best knowledge, RenderOcc is the first attempt to train multi-view 3D occupancy models only using 2D labels, reducing the dependence on costly 3D occupancy annotations. Extensive experiments demonstrate that RenderOcc achieves comparable performance to models fully supervised with 3D labels, underscoring the significance of this approach in real-world applications. Our code is available at https://github.com/pmj110119/RenderOcc.

Until now, the potential of untethered magnetic robots (UMRs), propelled by external time-periodic magnetic fields, has been hindered by the limitations of wireless manipulation systems or noninvasive imaging techniques combined. The need for simultaneous actuation and noninvasive localization imposes a strict constraint on both functionalities. This study addresses this challenge by substantiating the feasibility through experimental validation, showcasing the direct teleoperation of UMRs within a fluid-filled lumen. This teleoperation capability is facilitated by a scalable X-ray-guided robotic platform, extendable to match the dimensions required for in vivo applications, marking a noteworthy advancement. Our methodology is demonstrated by teleoperating a 12-mm-long screw-shaped UMR (5 mm in diameter) within a bifurcated lumen, filled with blood. This navigation is achieved using controlled rotating magnetic fields, guided by real-time X-ray Fluoroscopy images. Incorporating a two-degree-of-freedom control system, we demonstrate the operator’s capability to use X-ray Fluoroscopy images to keep the UMR coupled with the external field during wireless teleoperations, resulting in a success rate of 76.6% when moving along the intended pathways, with a mean absolute position error of 1.6 ± 2.1 mm.

As the usage of bi-copters increases in military and civilian fields, the demand for reliable bi-copters is on the rise. This study focuses on controlling a bi-copter under rotor or servo stuck failure. A relaxed hover solution is derived for the bi-copter, by solving an optimization problem subject to rotor and servo stuck failures. The solution is used for designing a reduced attitude controller based on linear quadratic regulator (LQR). To ensure hover capability, we introduce a position controller based on a cascaded-PID. The numerical simulations are conducted to demonstrate that position control is possible, even with complete rotor or servo stuck failure, by driving the bi-copter into relaxed hover state through the abandonment of the yaw channel. Meanwhile, the FTC scheme is examined under constant wind disturbances and uncertainties in the rotational damping parameters.

Inspired by the necessity of morphological adaptation in animals, a growing body of work has attempted to expand robot training to encompass physical aspects of a robot’s design. However, reinforcement learning methods capable of optimizing the 3D morphology of a robot have been restricted to reorienting or resizing the limbs of a predetermined and static topological genus. Here we show policy gradients for designing freeform robots with arbitrary external and internal structure. This is achieved through actions that deposit or remove bundles of atomic building blocks to form higher-level nonparametric macrostructures such as appendages, organs and cavities. Although results are provided for open loop control only, we discuss how this method could be adapted for closed loop control and sim2real transfer to physical machines in future.

In light of the challenges and costs of real-world testing, autonomous vehicle developers often rely on testing in simulation for the creation of reliable systems. A key element of effective simulation is the incorporation of realistic traffic models that align with human knowledge, an aspect that has proven challenging due to the need to balance realism and diversity. Towards this end, in this work we develop a framework that employs reinforcement learning from human feedback (RLHF) to enhance the realism of existing traffic models. This work also identifies two main challenges: capturing the nuances of human preferences on realism and unifying diverse traffic simulation models. To tackle these issues, we propose using human feedback for alignment and employ RLHF due to its sample efficiency. We also introduce the first dataset for realism alignment in traffic modeling to support such research. Our framework, named TrafficRLHF, demonstrates its proficiency in generating realistic traffic scenarios that are well-aligned with human preferences through comprehensive evaluations on the nuScenes dataset.

Safe Reinforcement Learning (RL) plays an important role in applying RL algorithms to safety-critical real-world applications, addressing the trade-off between maximizing rewards and adhering to safety constraints. This work introduces a novel approach that combines RL with trajectory optimization to manage this trade-off effectively. Our approach embeds safety constraints within the action space of a modified Markov Decision Process (MDP). The RL agent produces a sequence of actions that are transformed into safe trajectories by a trajectory optimizer, thereby effectively ensuring safety and increasing training stability. This novel approach excels in its performance on challenging Safety Gym tasks, achieving significantly higher rewards and near-zero safety violations during inference. The method’s real-world applicability is demonstrated through a safe and effective deployment in a real robot task of box-pushing around obstacles. Further insights are available from the videos and appendix on our website: https://sites.google.com/view/safemdp.

Model-based approaches for planning and control for bipedal locomotion have a long history of success. It can provide stability and safety guarantees while being effective in accomplishing many locomotion tasks. Model-free reinforcement learning, on the other hand, has gained much popularity in recent years due to computational advancements. It can achieve high performance in specific tasks, but it lacks physical interpretability and flexibility in re-purposing the policy for a different set of tasks. For instance, we can initially train a neural network (NN) policy using velocity commands as inputs. However, to handle new task commands like desired hand or footstep locations at a desired walking velocity, we must retrain a new NN policy. In this work, we attempt to bridge the gap between these two bodies of work on a bipedal platform. We formulate a model-based reinforcement learning problem to learn a reduced-order model (ROM) within a model predictive control (MPC). Results show a 49% improvement in viable task region size and a 21% reduction in motor torque cost. All videos and code are available at https://sites.google.com/view/ymchen/research/rl-for-roms.

This work contributes a novel deep navigation policy that enables collision-free flight of aerial robots based on a modular approach exploiting deep collision encoding and reinforcement learning. The proposed solution builds upon a deep collision encoder that is trained on both simulated and real depth images using supervised learning such that it compresses the high-dimensional depth data to a low-dimensional latent space encoding collision information while accounting for the robot size. This compressed encoding is combined with an estimate of the robot’s odometry and the desired target location to train a deep reinforcement learning navigation policy that offers low-latency computation and robust sim2real performance. A set of simulation and experimental studies in diverse environments are conducted and demonstrate the efficiency of the emerged behavior and its resilience in real-life deployments.

In recent years, legged and wheeled-legged robots have gained prominence for tasks in environments predominantly created for humans across various domains. One significant challenge faced by many of these robots is their limited capability to navigate stairs, which hampers their functionality in multi-story environments. This study proposes a method aimed at addressing this limitation, employing reinforcement learning to develop a versatile controller applicable to a wide range of robots. In contrast to the conventional velocity-based controllers, our approach builds upon a position-based formulation of the RL task, which we show to be vital for stair climbing. Furthermore, the methodology leverages an asymmetric actor-critic structure, enabling the utilization of privileged information from simulated environments during training while eliminating the reliance on exteroceptive sensors during real-world deployment. Another key feature of the proposed approach is the incorporation of a boolean observation within the controller, enabling the activation or deactivation of a stair-climbing mode. We present our results on different quadrupeds and bipedal robots in simulation and showcase how our method allows the balancing robot Ascento to climb 15cm stairs in the real world, a task that was previously impossible for this robot.

Visual place recognition is a critical task in computer vision, especially for localization and navigation systems. Existing methods often rely on contrastive learning: image descriptors are trained to have small distance for similar images and larger distance for dissimilar ones in a latent space. However, this approach struggles to ensure accurate distance-based image similarity representation, particularly when training with binary pairwise labels, and complex re-ranking strategies are required. This work introduces a fresh perspective by framing place recognition as a regression problem, using camera field-of-view overlap as similarity ground truth for learning. By optimizing image descriptors to align directly with graded similarity labels, this approach enhances ranking capabilities without expensive re-ranking, offering data-efficient training and strong generalization across several benchmark datasets.

The disposal of waste electrical and electronic equipment (WEEE) presents a sustainability challenge, particularly for waste printed circuit boards (PCBs). PCBs are challenging to sort out from other waste materials in part because traditional industrial end-effectors struggle to reliably grip these irregularly shaped objects with unmodeled surface-mounted components. Vision-based separators, while effective for object categorization, face challenges with identifying precise grasp points on PCB surfaces. This paper studies regrasping control to enhance suction cup grasping performance on PCBs, addressing issues arising from uneven surfaces and intricate features that interfere with suction sealing. We categorize PCBs into two recycling levels – with large surface features intact or removed – and conduct experiments on both stationary and conveyor belt setups with realistic vision-based grasp planners. Results show that jumping regrasping improves pick-and-place success rate. Haptically driven jumping – using the Smart Suction Cup – is especially useful for unprocessed waste PCBs with large surface mount parts. The proposed method offers a promising solution to enhance the efficiency and reliability of robotic grasping in recycling applications.

In this paper, a region-determined localization method applied for unmanned ground vehicles (UGVs) is presented. The method aims to solve GNSS-denied localization problem using pole-like feature such as trees or street lights. The approach includes three parts: mapping, bounding, and localization. To map and reconstruct the environment, the hector mapping approach and circle-fitting method are adopted for the occupancy mapping and feature mapping. To bound out the available working region, we define the intersection area of features' enlarged radius and desired operating area as negative and positive virtual boundaries. While the robot is cruising, the likelihood detection method is adopted for obstacle searching and comparing. Using the detection's searching results as feedback reference, the Extended Kalman Filter (EKF) can modify the shifting between the GNSS signal and the true waypoints of the mowing robot. Three cruising demonstrations are presented to show the mapping and optimizing results. Different cases of demonstration represent different situations and potential issues.

Neural fields, coordinate-based neural networks, have recently gained popularity for implicitly representing a scene. In contrast to classical methods that are based on explicit representations such as point clouds, neural fields provide a continuous scene representation able to represent 3D geometry and appearance in a way which is compact and ideal for robotics applications. However, limited prior methods have investigated registering multiple neural fields by directly utilising these continuous implicit representations. In this paper, we present Reg-NF, a neural fields-based registration that optimises for the relative 6-DoF transformation between two arbitrary neural fields, even if those two fields have different scale factors. Key components of Reg-NF include a bidirectional registration loss, multi-view surface sampling, and utilisation of volumetric signed distance functions (SDFs). We showcase our approach on a new neural field dataset for evaluating registration problems. We provide an exhaustive set of experiments and ablation studies to identify the performance of our approach, while also discussing limitations to provide future direction to the research community on open challenges in utilizing neural fields in unconstrained environments.