IROS 2023 论文列表

The safety-critical nature of vehicle steering is one of the main motivations for exploring the space of possible cyber-physical attacks against the steering systems of modern vehicles. This paper investigates the adversarial capabilities for destabilizing the interaction dynamics between human drivers and vehicle haptic shared control (HSC) steering systems. In contrast to the conventional robotics literature, where the main objective is to render the human-automation interaction dynamics stable by ensuring passivity, this paper takes the exact opposite route. In particular, to investigate the damaging capabilities of a successful cyber-physical attack, this paper demonstrates that an attacker who targets the HSC steering system can destabilize the interaction dynamics between the human driver and the vehicle HSC steering system through synthesis of time-varying impedance profiles. Specifically, it is shown that the adversary can utilize a properly designed non-passive and time-varying adversarial impedance target dynamics, which are fed with a linear combination of the human driver and the steering column torques. Using these target dynamics, it is possible for the adversary to generate in realtime a reference angular command for the driver input device and the directional control steering assembly of the vehicle. Furthermore, it is shown that the adversary can make the steering wheel and the vehicle steering column angular positions to follow the reference command generated by the time-varying impedance target dynamics using proper adaptive control strategies. Numerical simulations demonstrate the effectiveness of such time-varying impedance attacks, which result in a non-passive and inherently unstable interaction between the driver and the HSC steering system.

Rearranging objects is an essential skill for robots. To quickly teach robots new rearrangements tasks, we would like to generate training scenarios from high-level specifications that define the relative placement of objects for the task at hand. Ideally, to guide the robot's learning we also want to be able to rank these scenarios according to their difficulty. Prior work has shown how generating diverse scenario from specifications and providing the robot with easy-to-difficult samples can improve the learning. Yet, existing scenario generation methods typically cannot generate diverse scenarios while controlling their difficulty. We address this challenge by conditioning generative models on spatial logic specifications to generate spatially-structured scenarios that meet the specification and desired difficulty level. Our experiments showed that generative models are more effective and data-efficient than rejection sam-pling and that the spatially-structured scenarios can drastically improve training of downstream tasks by orders of magnitude.

Large Language Models (LLMs) trained using massive text datasets have recently shown promise in generating action plans for robotic agents from high-level text queries. However, these models typically do not consider the robot's environment, resulting in generated plans that may not actually be executable, due to ambiguities in the planned actions or environmental constraints. In this paper, we propose an approach to generate environmentally-aware action plans that agents are better able to execute. Our approach involves integrating environmental objects and object relations as additional inputs into LLM action plan generation to provide the system with an awareness of its surroundings, resulting in plans where each generated action is mapped to objects present in the scene. We also design a novel scoring function that, along with generating the action steps and associating them with objects, helps the system disambiguate among object instances and take into account their states. We evaluated our approach using the VirtualHome simulator and the ActivityPrograms knowledge base and found that action plans generated from our system had a 310% improvement in executability and a 147% improvement in correctness over prior work. The complete code and a demo of our method is publicly available at https://github.com/hri-ironlab/scene_aware_language_planner.

Despite their impressive performance in various surgical scene understanding tasks, deep learning-based methods are frequently hindered from deploying to real-world surgical applications for various causes. Particularly, data collection, annotation, and domain shift in-between sites and patients are the most common obstacles. In this work, we mitigate data-related issues by efficiently leveraging minimal source images to generate synthetic surgical instrument segmentation datasets and achieve outstanding generalization performance on unseen real domains. Specifically, in our framework, only one background tissue image and at most three images of each foreground instrument are taken as the seed images. These source images are extensively transformed and employed to build up the foreground and background image pools, from which randomly sampled tissue and instrument images are composed with multiple blending techniques to generate new surgical scene images. Besides, we introduce hybrid training-time augmentations to diversify the training data further. Extensive evaluation on three real-world datasets, i.e., Endo2017, Endo2018, and RoboTool, demonstrates that our one-to-many synthetic surgical instruments datasets generation and segmentation framework can achieve encouraging performance compared with training with real data. Notably, on the RoboTool dataset, where a more significant domain gap exists, our framework shows its superiority of generalization by a considerable margin. We expect that our inspiring results will attract research attention to improving model generalization with data synthesizing.

Acquiring dynamics is critical for robot learning and is fundamental to planning and control. This paper concerns two fundamental questions: How can we learn a model that covers massive, diverse robot dynamics? Can we construct a model that lifts the data-collection pain and domain expertise required for building specific robot models? We learn the dynamics involved in a dataset containing a large number of serial articulated robots and propose a new concept, “Gen2Real”, to transfer simulated, generalized models to physical, specific robots. We generate a large-scale dataset by randomizing dynamics parameters, topology configurations, and model dimensions, which, in sequence, correspond to different properties, connections, and numbers of robot links. A structure modified from the generative pre-trained transformer is applied to approximate the dynamics of massive heterogeneous robots. In Gen2Real, we transfer the pre-trained model to a target robot using distillation, for the sake of real-time computation. The results demonstrate the superiority of the proposed method in terms of its accuracy in learning a tremendous amount of robot dynamics and its generality to transfer to different robots.

Semantic segmentation of LiDAR point clouds can provide assistance for precise perception in autonomous driving, but traditional segmentation methods face challenges such as unbalanced class distribution and insufficient labeling. Generalized few-shot learning has been researched on image data, but these methods are difficult to apply directly to LiDAR point clouds. To tackle these challenges, we propose a generalized few-shot semantic segmentation method based on LiDAR point cloud data, enabling us to predict base and novel classes simultaneously. To improve the performance with limited novel class samples, we integrate semantic vectors and leverage the intrinsic relationship between base and novel class vectors to facilitate learning. We conduct comprehensive comparisons with other methods on the SemanticKITTI and constantly surpass them with higher mIoU, demonstrating the effectiveness of our method.

Camera-based tactile sensing is a low-cost, popular approach to obtain highly detailed contact geometry information. However, most existing camera-based tactile sensors are fingertip sensors, and longer fingers often require extraneous elements to obtain an extended sensing area similar to the full length of a human finger. Moreover, existing methods to estimate proprioceptive information such as total forces and torques applied on the finger from camera-based tactile sensors are not effective when the contact geometry is complex. We introduce GelSight Svelte, a curved, human finger-sized, single-camera tactile sensor that is capable of both tactile and proprioceptive sensing over a large area. GelSight Svelte uses curved mirrors to achieve the desired shape and sensing coverage. Proprioceptive information, such as the total bending and twisting torques applied on the finger, is reflected as deformations on the flexible backbone of GelSight Svelte, which are also captured by the camera. We train a convolutional neural network to estimate the bending and twisting torques from the captured images. We conduct gel deformation experiments at various locations of the finger to evaluate the tactile sensing capability and proprioceptive sensing accuracy. To demonstrate the capability and potential uses of GelSight Svelte, we conduct an object holding task with three different grasping modes that utilize different areas of the finger.

We study the multi-agent Bayesian optimization (BO) problem, where multiple agents maximize a black-box function via iterative queries. We focus on Entropy Search (ES), a sample-efficient BO algorithm that selects queries to maximize the mutual information about the maximum of the black-box function. One of the main challenges of ES is that calculating the mutual information requires computationallycostly approximation techniques. For multi-agent BO problems, the computational cost of ES is exponential in the number of agents. To address this challenge, we propose the Gaussian Max-value Entropy Search, a multi-agent BO algorithm with favorable sample and computational efficiency. The key to our idea is to use a normal distribution to approximate the function maximum and calculate its mutual information accordingly. The resulting approximation allows queries to be cast as the solution of a closed-form optimization problem which, in turn, can be solved via a modified gradient ascent algorithm and scaled to a large number of agents. We demonstrate the effectiveness of Gaussian max-value Entropy Search through numerical experiments on standard test functions and real-robot experiments on the source seeking problem. Results show that the proposed algorithm outperforms the multi-agent BO baselines in the numerical experiments and can stably seek the source with a limited number of noisy observations on real robots.

This paper introduces a game-theoretical approach to the multi-robot task allocation problem, where each robot is considered as self-interested and cannot share its personal utility functions. We consider the case where each robot can execute multiple tasks and each task requires only one robot. For real-world applications with mobile robots, we design a utility function that includes both assignment conflict penalties and path-dependent execution cost. For a robot to maximize its own utility, it needs to select a subset of conflict-free tasks that minimizes its total travel distance. Our approaches utilize a consensus communication scheme to share robots' task selection and the Speeding-Up and Slowing-Down (SUSD) strategy to search in a combinatorial action (task selection) space for a subset of tasks that can achieve a higher utility at each iteration. The SUSD strategy can perform a gradient-like search without calculating the derivatives, which allows robots to improve upon their current task selections. Simulation results show that robots using the proposed algorithms can successfully find Nash equilibria for effective coordination.

Language-Guided Robotic Manipulation (LGRM) is a challenging task as it requires a robot to understand human instructions to manipulate everyday objects. Recent approaches in LGRM rely on pre-trained Visual Grounding (VG) models to detect objects without adapting to manipulation environments. This results in a performance drop due to a substantial domain gap between the pre-training and real-world data. A straight-forward solution is to collect additional training data, but the cost of human-annotation is extortionate. In this paper, we propose Grounding Vision to Ceaselessly Created Instructions (GVCCI), a lifelong learning framework for LGRM, which continuously learns VG without human supervision. GVCCI iteratively generates synthetic instruction via object detection and trains the VG model with the generated data. We validate our framework in offline and online settings across diverse environments on different VG models. Experimental results show that accumulating synthetic data from GVCCI leads to a steady improvement in VG by up to 56.7% and improves resultant LGRM by up to 29.4%. Furthermore, the qualitative analysis shows that the unadapted VG model often fails to find correct objects due to a strong bias learned from the pre-training data. Finally, we introduce a novel VG dataset for LGRM, consisting of nearly 252k triplets of image-object-instruction from diverse manipulation environments.

Robotic navigation in unknown, cluttered environ-ments with limited sensing capabilities poses significant chal-lenges in robotics. Local trajectory optimization methods, such as Model Predictive Path Intergal (MPPI), are a promising solution to this challenge. However, global guidance is required to ensure effective navigation, especially when encountering challenging environmental conditions or navigating beyond the planning horizon. This study presents the GP-MPPI, an online learning-based control strategy that integrates MPPI with a local perception model based on Sparse Gaussian Process (SGP). The key idea is to leverage the learning capability of SGP to construct a variance (uncertainty) surface, which enables the robot to learn about the navigable space surrounding it, identify a set of suggested subgoals, and ultimately recommend the optimal subgoal that minimizes a predefined cost function to the local MPPI planner. Afterward, MPPI computes the optimal control sequence that satisfies the robot and collision avoidance constraints. Such an approach eliminates the necessity of a global map of the environment or an offline training process. We validate the efficiency and robustness of our proposed control strategy through both simulated and real-world experiments of 2D autonomous navigation tasks in complex unknown en-vironments, demonstrating its superiority in guiding the robot safely towards its desired goal while avoiding obstacles and escaping entrapment in local minima. The GPU implementation of GP-MPPI, including the supplementary video, is available at https://github.com/IhabMohamed/GP-MPPI.

In this work, we first formulate the problem of robotic water scooping using goal-conditioned reinforcement learning. This task is particularly challenging due to the complex dynamics of fluid and the need to achieve multi-modal goals. The policy is required to successfully reach both position goals and water amount goals, which leads to a large convoluted goal state space. To overcome these challenges, we introduce Goal Sampling Adaptation for Scooping (GOATS), a curriculum reinforcement learning method that can learn an effective and generalizable policy for robot scooping tasks. Specifically, we use a goal-factorized reward formulation and interpolate position goal distributions and amount goal distributions to create curriculum throughout the learning process. As a result, our proposed method can outperform the baselines in simulation and achieves 5.46% and 8.71% amount errors on bowl scooping and bucket scooping tasks, respectively, under 1000 variations of initial water states in the tank and a large goal state space. Besides being effective in simulation environments, our method can efficiently adapt to noisy real-robot water-scooping scenarios with diverse physical configurations and unseen settings, demonstrating superior efficacy and generalizability. The videos of this work are available on our project page: https://sites.google.com/view/goatscooping.

Inferring the posterior distribution in SLAM is critical for evaluating the uncertainty in localization and mapping, as well as supporting subsequent planning tasks aiming to reduce uncertainty for safe navigation. However, real-time full posterior inference techniques, such as Gaussian approximation and particle filters, either lack expressiveness for representing non-Gaussian posteriors or suffer from performance degeneracy when estimating high-dimensional posteriors. Inspired by the complementary strengths of Gaussian approximation and particle filters-scalability and non-Gaussian estimation, respectively-we blend these two approaches to infer marginal posteriors in SLAM. Specifically, Gaussian approximation provides robot pose distributions on which particle filters are conditioned to sample landmark marginals. In return, the maximum a posteriori point among these samples can be used to reset linearization points in the nonlinear optimization solver of the Gaussian approximation, facilitating the pursuit of global optima. We demonstrate the scalability, generalizability, and accuracy of our algorithm for real-time full posterior inference on realworld range-only SLAM and object-based bearing-only SLAM datasets.

In many applications of advanced robotic manipulation, six degrees of freedom (6DoF) object pose estimates are continuously required. In this work, we develop a multi-modality tracker that fuses information from visual appearance and geometry to estimate object poses. The algorithm extends our previous method ICG, which uses geometry, to additionally consider surface appearance. In general, object surfaces contain local characteristics from text, graphics, and patterns, as well as global differences from distinct materials and colors. To incorporate this visual information, two modalities are developed. For local characteristics, keypoint features are used to minimize distances between points from keyframes and the current image. For global differences, a novel region approach is developed that considers multiple regions on the object surface. In addition, it allows the modeling of external geometries. Experiments on the YCB-Video and OPT datasets demonstrate that our approach ICG+ performs best on both datasets, outperforming both conventional and deep learning-based methods. At the same time, the algorithm is highly efficient and runs at more than 300 Hz. The source code of our tracker is publicly available.

This paper develops a novel slip estimator using the invariant observer design theory and Disturbance Observer (DOB). The proposed state estimator for mobile robots is fully proprioceptive and combines data from an inertial measurement unit and body velocity within a Right Invariant Extended Kalman Filter (RI-EKF). By embedding the slip velocity into SE3 (3) matrix Lie group, the developed DOB-based RI-EKF provides real-time velocity and slip velocity estimates on different terrains. Experimental results using a Husky wheeled robot confirm the mathematical derivations and effectiveness of the proposed method in estimating the observable state variables. Open-source software is available for download and reproducing the presented results.

Learning from human feedback is an effective way to improve robotic learning in exploration-heavy tasks. Compared to the wide application of binary human feedback, scalar human feedback has been used less because it is believed to be noisy and unstable. In this paper, we compare scalar and binary feedback, and demonstrate that scalar feedback benefits learning when properly handled. We collected binary or scalar feedback respectively from two groups of crowdworkers on a robot task. We found that when considering how consistently a participant labeled the same data, scalar feedback led to less consistency than binary feedback; however, the difference vanishes if small mismatches are allowed. Additionally, scalar and binary feedback show no significant differences in their correlations with key Reinforcement Learning targets. We then introduce Stabilizing TEacher Assessment DYnamics (STEADY) to improve learning from scalar feedback. Based on the idea that scalar feedback is muti-distributional, STEADY reconstructs underlying positive and negative feedback distributions and re-scales scalar feedback based on feedback statistics. We show that models trained with scalar feedback + STEADY outperform baselines, including binary feedback and raw scalar feedback, in a robot reaching task with non-expert human feedback. Our results show that both binary feedback and scalar feedback are dynamic, and scalar feedback is a promising signal for use in interactive Reinforcement Learning.

The current learning pipelines for robotics manipulation infer movement primitives sequentially along the temporal-evolving axis, which can result in an accumulation of prediction errors and subsequently cause the visual observations to fall out of the training distribution. This paper proposes a novel hierarchical behavior cloning approach which tries to dissociate standard behaviour cloning (BC) pipeline to two stages. The intuition of this approach is to eliminate accumu-lation errors using a fixed spatial representation. At first stage, a high-level planner will be employed to translate the initial observation of the scene into task-specific spatial waypoints. Then, a low-level robotic path planner takes over the task of guiding the robot by executing a set of pre-defined elementary movements or actions known as primitives, with the goal of reaching the previously predicted waypoints. Our hierarchical keypoints-based paradigm aims to simplify existing temporal-evolving approach to a more simple way: directly spatialize the whole sequential primitives as a set of 8D waypoints only from the very first observation. Plentiful experiments demon-strate that our paradigm can achieve comparable results with Reinforcement Learning (RL) and outperforms existing offline BC approaches, with only a single-shot inference from the initial observation. Code and models are available at: https://github.com/KevinRiou22/spatial-fixing-il

We focus on robot navigation in crowded environments. To navigate safely and efficiently within crowds, robots need models for crowd motion prediction. Building such models is hard due to the high dimensionality of multiagent domains and the challenge of collecting or simulating interaction-rich crowd-robot demonstrations. While there has been important progress on models for offline pedestrian motion forecasting, transferring their performance on real robots is nontrivial due to close interaction settings and novelty effects on users. In this paper, we investigate the utility of a recent state-of-the-art motion prediction model (S-GAN) for crowd navigation tasks. We incorporate this model into a model predictive controller (MPC) and deploy it on a self-balancing robot which we subject to a diverse range of crowd behaviors in the lab. We demonstrate that while S-GAN motion prediction accuracy transfers to the real world, its value is not reflected on navigation performance, measured with respect to safety and efficiency; in fact, the MPC performs indistinguishably even when using a simple constant-velocity prediction model, suggesting that substantial model improvements might be needed to yield significant gains for crowd navigation tasks. Footage from our experiments can be found at https://youtu.be/mzFiXgSKsZ0.

This paper uses a data-driven approach to model a highly redundantly driven tensegrity manipulator's forward and inverse kinematics. The tensegrity manipulator is based on a class-1 tensegrity with 20 struts and bends by 40 pneumatic actuators whose internal pressures are independently controlled. Based on the data obtained through random trials with the robot, a VAE-based kinematics model is trained. The forward model, inverse model, and null space of kinematics are simultaneously acquired as subnetworks of the VAE-based kinematics model. Experiments confirmed that the subnetworks representing forward and inverse kinematics could be used for the end position estimation and control, respectively. In addition, the subnetwork representing null space can generate different target pressures that achieve the same end position, which was confirmed to mean variable stiffness properties similar to musculoskeletal robots.