RSS 2024 论文列表

Representation learning approaches for robotic manipulation have boomed in recent years. Due to the scarcity of in-domain robot data, prevailing methodologies tend to leverage large-scale human video datasets to extract generalizable features for visuomotor policy learning. Despite the progress achieved, prior endeavors disregard the interactive dynamics that capture behavior patterns and physical interaction during the manipulation process, resulting in an inadequate understanding of the relationship between objects and the environment. To this end, we propose a general pre-training pipeline that learns Manipulation by Predicting the Interaction (MPI) and enhances the visual representation. Given a pair of keyframes representing the initial and final states, along with language instructions, our algorithm predicts the transition frame and detects the interaction object, respectively. These two learning objectives achieve superior comprehension towards “how-to-interact” and “where-to-interact”. We conduct a comprehensive evaluation of four challenging robotic tasks. The experimental results demonstrate that MPI exhibits remarkable improvement by 10% to 64% compared with previous state-of-the-art in real-world robot platforms as well as simulation environments. Code and checkpoints are publicly shared at https://github.com/OpenDriveLab/MPI.

A significant challenge for real-world robotic manipulation is the effective 6DoF grasping of objects in cluttered scenes from any single viewpoint without needing additional scene exploration. This work re-interprets grasping as rendering and introduces NeuGraspNet, a novel method for 6DoF grasp detection that leverages advances in neural volumetric representations and surface rendering. We encode the interaction between a robot's end-effector and an object's surface by jointly learning to render the local object surface and learning grasping functions in a shared feature space. Our approach uses global (scene-level) features for grasp generation and local (grasp-level) neural surface features for grasp evaluation. This enables effective, fully implicit 6DoF grasp quality prediction, even in partially observed scenes. NeuGraspNet operates on random viewpoints, common in mobile manipulation scenarios, and outperforms existing implicit and semi-implicit grasping methods. We demonstrate the real-world applicability of the method with a mobile manipulator robot, grasping in open cluttered spaces. Project website at: https://sites.google.com/view/neugraspnet

Enabling robotic agents to perform complex long-horizon tasks has been a long-standing goal in robotics and artificial intelligence (AI). Despite the potential shown by large language models (LLMs), their planning capabilities remain limited to short-horizon tasks and they are unable to replace the symbolic planning approach. Symbolic planners, on the other hand, may encounter execution errors due to their common assumption of complete domain knowledge which is hard to manually prepare for an open-world setting. In this paper, we introduce a Language-Augmented Symbolic Planner (LASP) that integrates pre-trained LLMs to enable conventional symbolic planners to operate in an open-world environment where only incomplete knowledge of action preconditions, objects, and properties is initially available. In case of execution errors, LASP can utilize the LLM to diagnose the cause of the error based on the observation and interact with the environment to incrementally build up its knowledge base necessary for accomplishing the given tasks. Experiments demonstrate that LASP is proficient in solving planning problems in the open-world setting, performing well even in situations where there are multiple gaps in the knowledge.

Perceiving and understanding highly dynamic and changing environments is a crucial capability for robot autonomy. While large strides have been made towards developing dynamic SLAM approaches that estimate the robot pose accurately, a lesser emphasis has been put on the construction of dense spatio-temporal representations of the robot environment. A detailed understanding of the scene and its evolution through time is crucial for long-term robot autonomy and essential to tasks that require long-term reasoning, such as operating effectively in environments shared with humans and other agents and thus are subject to short and long-term dynamics. To address this challenge, this work defines the Spatio-temporal Metric-semantic SLAM (SMS) problem, and presents a framework to factorize and solve it efficiently. We show that the proposed factorization suggests a natural organization of a spatio-temporal perception system, where a fast process tracks short-term dynamics in an active temporal window, while a slower process reasons over long-term changes in the environment using a factor graph formulation. We provide an efficient implementation of the proposed spatio-temporal perception approach, that we call Khronos, and show that it unifies exiting interpretations of short-term and long-term dynamics and is able to construct a dense spatio-temporal map in real-time. We provide simulated and real results, showing that the spatio-temporal maps built by Khronos are an accurate reflection of a 3D scene over time and that Khronos outperforms baselines across multiple metrics. We further validate our approach on two heterogeneous robots in challenging, large-scale real-world environments.

We show that off-the-shelf text-based Transformers, with no additional training, can perform few-shot in-context visual imitation learning, mapping visual observations to action sequences that emulate the demonstrator's behaviour. We achieve this by transforming visual observations (inputs) and trajectories of actions (outputs) into sequences of tokens that a text-pretrained Transformer (GPT-4 Turbo) can ingest and generate, via a framework we call Keypoint Action Tokens (KAT). Despite being trained only on language, we show that these Transfermers excel at translating tokenised visual keypoint observations into action trajectories, performing on par or better than state-of-the-art imitation learning (diffusion policies) in the low-data regime on a suite of real-world, everyday tasks. Rather than operating in the language domain as is typical, KAT leverages text-based Transformers to operate in the vision and action domains to learn general patterns in demonstration data for highly efficient imitation learning, indicating promising new avenues for repurposing natural language models for embodied tasks. Videos can be found on our website https://www.robot-learning.uk/keypoint-action-tokens.

Surgical automation can improve the accessibility and consistency of life-saving procedures. Most surgeries require separating layers of tissue to access the surgical site, and suturing to re-attach incisions. These tasks involve deformable manipula- tion to safely identify and alter tissue attachment (boundary) topology. Due to poor visual acuity and frequent occlusions, surgeons tend to carefully manipulate the tissue in ways that enable inference of the tissue’s attachment points without causing unsafe tearing. In a similar fashion, we propose JIGGLE, a framework for estimation and interactive sensing of unknown boundary parameters in deformable surgical environments. This framework has two key components: (1) a probabilistic estimation to identify the current attachment points, achieved by integrating a differentiable soft-body simulator with an extended Kalman filter (EKF), and (2) an optimization-based active control pipeline that generates actions to maximize information gain of the tissue attachments, while simultaneously minimizing safety costs. The robustness of our estimation approach is demonstrated through experiments with real animal tissue, where we infer sutured attachment points using stereo endoscope observations. We also demonstrate the capabilities of our method in handling complex topological changes such as cutting and suturing.

Smooth, collision-free motion planning is a fundamental challenge in robotics with a wide range of applications such as automated manufacturing, search \& rescue, underwater exploration, etc. Graphs of Convex Sets (GCS) is a recent method for synthesizing smooth trajectories by decomposing the planning space into convex sets, forming a graph to encode the adjacency relationships within the decomposition, and then simultaneously searching this graph and optimizing parts of the trajectory to obtain the final trajectory. To do this, one must solve a Mixed Integer Convex Program (MICP) and to mitigate computational time, GCS proposes a convex relaxation that is empirically very tight. Despite this tight relaxation, motion planning with GCS for real-world robotics problems translates to solving the simultaneous batch optimization problem that may contain millions of constraints and therefore can be slow. This is further exacerbated by the fact that the size of the GCS problem is invariant to the planning query. Motivated by the observation that the trajectory solution lies only on a fraction of the set of convex sets, we present two implicit graph search methods for planning on the graph of convex sets called INSATxGCS (IxG) and IxG*. INterleaved Search And Trajectory optimization (INSAT) is a previously developed algorithm that alternates between searching on a graph and optimizing partial paths to find a smooth trajectory. By using an implicit graph search method INSAT on the graph of convex sets, we achieve faster planning while ensuring stronger guarantees on completeness and optimality. Moveover, introducing a search-based technique to plan on the graph of convex sets enables us to easily leverage well-established techniques such as search parallelization, lazy planning, anytime planning, and replanning as future work. Numerical comparisons against GCS demonstrate the superiority of IxG across several applications, including planning for an 18-degree-of-freedom multi-arm assembly scenario.

Despite the considerable potential of reinforcement learning (RL), robotics control tasks predominantly rely on imitation learning (IL) due to its better sample efficiency. However, it is costly to collect comprehensive expert demonstrations that enable IL to generalize to all possible scenarios, and any distribution shift would require recollecting data for finetuning. Therefore, RL is appealing if it can build upon IL as an efficient autonomous self-improvement procedure. We propose _imitation bootstrapped reinforcement learning_ (IBRL), a novel framework for sample-efficient RL with demonstrations that first trains an IL policy on the provided demonstrations and then uses it to propose alternative actions for both online exploration and bootstrapping target values. Compared to prior works that oversample the demonstrations or regularize RL with additional imitation loss, IBRL is able to utilize high quality actions from IL policies since the beginning of training, which greatly accelerates exploration and training efficiency. We evaluate IBRL on 6 simulation and 3 real-world tasks spanning various difficulty levels. IBRL significantly outperforms prior methods and the improvement is particularly more prominent in harder tasks.

Learning abstract state representations and knowledge is crucial for long-horizon robot planning. We present InterPreT, an LLM-powered framework for robots to learn symbolic predicates from language feedback of human non-experts during embodied interaction. The learned predicates provide relational abstractions of the environment state, facilitating the learning of symbolic operators that capture action preconditions and effects. By compiling the learned predicates and operators into a PDDL domain on-the-fly, InterPreT allows effective planning toward arbitrary in-domain goals using a PDDL planner. In both simulated and real-world robot manipulation domains, we demonstrate that InterPreT reliably uncovers the key predicates and operators governing the environment dynamics. Although learned from simple training tasks, these predicates and operators exhibit strong generalization to novel tasks with significantly higher complexity. In the most challenging generalization setting, InterPreT attains success rates of 73% in simulation and 40% in the real world, substantially outperforming baseline methods.

Humanoid robots hold great promise in assisting humans in diverse environments and tasks, due to their flexibility and adaptability leveraging human-like morphology. However, research in humanoid robots is often bottlenecked by the costly and fragile hardware setups. To accelerate algorithmic research in humanoid robots, we present a high-dimensional, simulated robot learning benchmark, HumanoidBench, featuring a humanoid robot equipped with dexterous hands and a variety of challenging whole-body manipulation and locomotion tasks. Our findings reveal that state-of-the-art reinforcement learning algorithms struggle with most tasks, whereas a hierarchical learning baseline achieves superior performance when supported by robust low-level policies, such as walking or reaching. With HumanoidBench, we provide the robotics community with a platform to identify the challenges arising when solving diverse tasks with humanoid robots, facilitating prompt verification of algorithms and ideas. The open-source code is available at https://humanoid- bench.github.io.

Humans inherently possess generalizable visual representations that empower them to efficiently explore and interact with the environments in manipulation tasks. We advocate that such a representation automatically arises from simultaneously learning about multiple simple perceptual skills that are critical for everyday scenarios (e.g., hand detection, state estimate, etc.) and is better suited for learning robot manipulation policies compared to current state-of-the-art visual representations purely based on self-supervised objectives. We formalize this idea through the lens of human-oriented multi-task fine-tuning on top of pre-trained visual encoders, where each task is a perceptual skill tied to human-environment interactions. We introduce Task Fusion Decoder as a plug-and-play embedding translator that utilizes the underlying relationships among these perceptual skills to guide the representation learning towards encoding meaningful structure for what’s important for all perceptual skills, ultimately empowering learning of downstream robotic manipulation tasks. Extensive experiments across a range of robotic tasks and embodiments, in both simulations and real-world environments, show that our Task Fusion Decoder consistently improves the representation of three state-of-the-art visual encoders including R3M, MVP, and EgoVLP, for downstream manipulation policy-learning. More demos, datasets, models, and code can be found at <

This paper addresses path set planning that yields important applications in robot manipulation and navigation such as path generation for deformable object keypoints and swarms. A path set refers to the collection of finite agent paths to represent the overall spatial path of a group of keypoints or a swarm, whose collective properties meet spatial and topological constraints. As opposed to planning a single path, simultaneously planning multiple paths with constraints poses nontrivial challenges in complex environments. This paper presents a systematic planning pipeline for homotopic path sets, a widely applicable path set class in robotics. An extended visibility check condition is first proposed to attain a sparse passage distribution amidst dense obstacles. Passage-aware optimal path planning compatible with sampling-based planners is then designed for single path planning with adjustable costs. Large accessible free space for path set accommodation can be achieved by the planned path while having a sufficiently short path length. After specifying the homotopic properties of path sets, path set generation based on deformable path transfer is proposed in an efficient centralized manner. The effectiveness of these methods is validated by extensive simulated and experimental results.

Recent open-vocabulary robot mapping methods enrich dense geometric maps with pre-trained visual-language features. While these maps allow for the prediction of point-wise saliency maps when queried for a certain language concept, large-scale environments and abstract queries beyond the object level still pose a considerable hurdle, ultimately limiting language-grounded robotic navigation. In this work, we present HOV-SG, a hierarchical open-vocabulary 3D scene graph mapping approach for language-grounded indoor robot navigation. Leveraging open-vocabulary vision foundation models, we first obtain state-of-the-art open-vocabulary segment-level maps in 3D and subsequently construct a 3D scene graph hierarchy consisting of floor, room, and object concepts, each enriched with open-vocabulary features. Our approach is able to represent multi-story buildings and allows robotic traversal of those using a cross-floor Voronoi graph. HOV-SG is evaluated on three distinct datasets and surpasses previous baselines in open-vocabulary semantic accuracy on the object, room, and floor level while producing a 75% reduction in representation size compared to dense open-vocabulary maps. In order to prove the efficacy and generalization capabilities of HOV-SG, we showcase successful long-horizon language-conditioned robot navigation within real-world multi-story environments. We provide code and trial video data at: https://hovsg.github.io.

Hybrid dynamical systems with nonlinear dynamics are one of the most general modeling tools for representing robotic systems, especially contact-rich systems. However, providing guarantees regarding the safety or performance of nonlinear hybrid systems remains a challenging problem because it requires simultaneous reasoning about continuous state evolution and discrete mode switching. In this work, we address this problem by extending classical Hamilton-Jacobi (HJ) reachability analysis, a formal verification method for continuous-time nonlinear dynamical systems, to hybrid dynamical systems. We characterize the reachable sets for hybrid systems through a generalized value function defined over discrete and continuous states of the hybrid system. We also provide a numerical algorithm to compute this value function and obtain the reachable set. Our framework can compute reachable sets for hybrid systems consisting of multiple discrete modes, each with its own set of nonlinear continuous dynamics, discrete transitions that can be directly commanded or forced by a discrete control input, while still accounting for control bounds and adversarial disturbances in the state evolution. Along with the reachable set, the proposed framework also provides an optimal continuous and discrete controller to ensure system safety. We demonstrate our framework in several simulation case studies, as well as on a real-world testbed to solve the optimal mode planning problem for a quadruped with multiple gaits.

In order to *generalize* to various tasks in the wild, robotic agents will need a suitable representation (i.e., vision network) that enables the robot to predict optimal actions given high dimensional vision inputs. However, learning such a representation requires an extreme amount of diverse training data, which is prohibitively expensive to collect on a real robot. How can we overcome this problem? Instead of collecting more robot data, this paper proposes using internet-scale, human videos to extract "affordances," both at the environment and agent level, and distill them into a pre-trained representation. We present a simple framework for pre-training representations on hand, object, and contact "affordance labels" that highlight relevant objects in images and how to interact with them. These affordances are automatically extracted from human video data (with the help of off-the-shelf computer vision modules) and used to fine-tune existing representations. Our approach can efficiently fine-tune *any* existing representation, and results in models with stronger downstream robotic performance across the board. We experimentally demonstrate (using 3000+ robot trials) that this affordance pre-training scheme boosts performance by a minimum of 15% on 5 real-world tasks, which consider three diverse robot morphologies (including a dexterous hand). Unlike prior works in the space, these representations improve performance across 3 different camera views. Quantitatively, we find that our approach leads to higher levels of generalization in out-of-distribution settings. Videos of our final policies and all code/weights/data can be found on our website: https://www.cs.cmu.edu/~data4robotics/hrp/

Although end-to-end robot learning has shown some success for robot manipulation, the learned policies are often not sufficiently robust to variations in object pose or geometry. To improve the policy generalization, we introduce spatially-grounded parameterized motion primitives in our method HACMan++. Specifically, we propose an action representation consisting of three components: "what" primitive type (such as grasp or push) to execute, "where" the primitive will be grounded (e.g. where the gripper will make contact with the world), and "how" the primitive motion is executed, such as parameters specifying the push direction or grasp orientation. These three components define a novel discrete-continuous action space for reinforcement learning. Our framework enables robot agents to learn to chain diverse motion primitives together and select appropriate primitive parameters to complete long-horizon manipulation tasks. By grounding the primitives on a spatial location in the environment, our method is able to effectively generalize across object shape and pose variations. Our approach significantly outperforms existing methods, particularly in complex scenarios demanding both high-level sequential reasoning and object generalization. With zero-shot sim-to-real transfer, our policy succeeds in challenging real-world manipulation tasks, with generalization to unseen objects. Videos can be found on the project website: https://sgmp-rss2024.github.io.

Autonomous mobile robots must maintain safety, but should not sacrifice performance, leading to the classical reach-avoid problem: find a trajectory that is guaranteed to reach a goal and avoid obstacles. This paper addresses the near danger case, also known as a narrow gap, where the agent starts near the goal, but must navigate through tight obstacles that block its path. The proposed method builds off the common approach of using a simplified planning model to generate plans, which are then tracked using a high-fidelity tracking model and controller. Existing approaches use reachability analysis to overapproximate the error between these models and ensure safety, but doing so introduces numerical approximation error conservativeness that prevents goal-reaching. The present work instead proposes a Piecewise Affine Reach-avoid Computation (PARC) method to tightly approximate the reachable set of the planning model. PARC significantly reduces conservativeness through a careful choice of the planning model and set representation, along with an effective approach to handling time-varying tracking errors. The utility of this method is demonstrated through extensive numerical experiments in which PARC outperforms state-of-the-art reach avoid methods in near-danger goal reaching. Furthermore, in a simulated demonstration, PARC enables the generation of provably-safe extreme vehicle dynamics drift parking maneuvers. A preliminary hardware demo on a TurtleBot3 also validates the method.

Abstract—This paper addresses the multi-faceted problem of robot grasping, where multiple criteria may conflict and differ in importance. We introduce a probabilistic framework, Grasp Ranking and Criteria Evaluation (GRaCE), which employs hierarchical rule-based logic and a rank-preserving utility function for grasps based on various criteria such as stability, kinematic constraints, and goal oriented functionalities. GRaCE's probabilistic nature means the framework handles uncertainty in a principled manner, i.e., the method is able to leverage the probability that a given criteria is satisfied. Additionally, we propose GRaCE-OPT, a hybrid optimization strategy that combines gradient-based and gradient-free methods to effectively navigate the complex, non-convex utility function. Experimental results in both simulated and real-world scenarios show that GRaCE requires fewer samples to achieve comparable or superior performance relative to existing methods. The modular architecture of GRaCE allows for easy customization and adaptation to specific application needs. Code and implementation details can be found online at https://github.com/clear-nus/GRaCE.

In deployment scenarios such as homes and warehouses, mobile robots are expected to autonomously navigate for extended periods, seamlessly executing tasks articulated in terms that are intuitively understandable by human operators. We present GO To Any Thing (GOAT), a universal navigation system capable of tackling these requirements with three key features: a) Multimodal: it can tackle goals specified via category labels, target images, and language descriptions, b) Lifelong: it benefits from its past experience in the same environment, and c) Platform Agnostic: it can be quickly deployed on robots with different embodiments. GOAT is made possible through a modular system design and a continually augmented instance-aware semantic memory that keeps track of the appearance of objects from different viewpoints in addition to category-level semantics. This enables GOAT to distinguish between different instances of the same category to enable navigation to targets specified by images and language descriptions. In experimental comparisons spanning over 90 hours in 9 different homes consisting of 675 goals selected across 200+ different object instances, we find GOAT achieves an overall success rate of 83%, surpassing previous methods and ablations by 32% (absolute improvement). GOAT improves with experience in the environment, from a 60% success rate at the first goal to a 90% success after exploration. In addition, we demonstrate that GOAT can readily be applied to downstream tasks such as pick and place and social navigation.