ICRA 2024 论文列表

Motion scaling is essential to empower users to conduct precise manipulation during teleoperation for robot-assisted microsurgery (RAMS). A constant, small motion scaling ratio can enhance the precision of teleoperation but hinder the operator from quickly reaching distant targets. The concept of self-adaptive motion scaling has been proposed in previous work. However, previous frameworks required extensive manual tuning of core parameters, which significantly depends on prior knowledge and may potentially lead to non-optimal solutions. This paper presents a hybrid offline reinforcement learning and damping control approach to regulate the motion scaling ratio for different operations during offline training. This method can take user-specific characteristics into consideration and help them achieve better teleoperation performance. Comparisons are made with and without using the adaptive motion-scaling algorithm. Detailed user studies indicate that a suitable motion-scaling ratio can be obtained and adjusted online. The overall performance of the operators in terms of time cost for task completion is significantly improved, while the variance of average speed and the total distance for robot operation is reduced.

Model Predictive Control (MPC) relies heavily on the robot model for its control law. However, a gap always exists between the reduced-order control model with uncertainties and the real robot, which degrades its performance. To address this issue, we propose the controller of integrating a data-driven error model into traditional MPC for quadruped robots. Our approach leverages real-world data from sensors to compensate for defects in the control model. Specifically, we employ the Autoregressive Moving Average Vector (ARMAV) model to construct the state error model of the quadruped robot using data. The predicted state errors are then used to adjust the predicted future robot states generated by MPC. By such an approach, our proposed controller can provide more accurate inputs to the system, enabling it to achieve desired states even in the presence of model parameter inaccuracies or disturbances. The proposed controller exhibits the capability to partially eliminate the disparity between the model and the real-world robot, thereby enhancing the locomotion performance of quadruped robots. We validate our proposed method through simulations and real-world experimental trials on a large-size quadruped robot that involves carrying a 20 kg un-modeled payload (84% of body weight).

Autonomous Underwater Vehicles (AUVs) conduct missions underwater without the need for human intervention. A docking station (DS) can extend mission times of an AUV by providing a location for the AUV to recharge its batteries and receive updated mission information. Various methods for locating and tracking a DS exist, but most rely on expensive acoustic sensors, or are vision-based, which is significantly affected by water quality. In this paper, we present a vision-based method that utilizes adaptive color LED markers and dynamic color filtering to maximize landmark visibility in varying water conditions. Both AUV and DS utilize cameras to determine the water background color in order to calculate the desired marker color. No communication between AUV and DS is needed to determine marker color. Experiments conducted in a pool and lake show our method performs 10 times better than static color thresholding methods as background color varies. DS detection is possible at a range of 5 meters in clear water with minimal false positives.

Steel mill is one of the most extreme and hazardous working environments due to molten iron erupted from blast furnace. Current manual labor to remove lump iron near the outlet, which is essential to prevent lump iron from scattering or blocking of molten iron, is performed by equipped human workers using a long stick tool. Thus, implementation of robotic teleoperation system is in demand to ensure the safety of workers. However, the conventional command interface is not intuitive for tool manipulation (i.e. pivoting, sweeping). Besides, haptic interface, which is used to render interaction results efficiently, still limits performance due to narrow workspace and insufficient kinesthetic feedback output compared to requirements. This paper proposes a novel haptic command interface (POstick) specified to lump iron removal task with two types (KF and VF). Both POsticks have rod-shaped end tip which is identical to actual tool already used to accelerate training. POstick-KF has large workspace and high kinesthetic feedback output satisfying requirements. Further, POstick-VF has strength with unlimited workspace at the expense of the amount of haptic information from simple vibrotactile feedback. User study to compare the performance of POsticks and conventional interface reveals that POstick-KF and VF showed superior interaction and tracking ability, respectively. Moreover, these two properties are in trade-off relationship that cannot be compatible. Finally, we proposed a seamless and automatic conversion mechanism from POstick-VF to KF, and vice versa, to cover up inherent limits of haptic devices.

Robotic adaptation to unanticipated operating conditions is crucial to achieving persistence and robustness in complex real world settings. For a wide range of cutting-edge robotic systems, such as micro- and nano-scale robots, soft robots, medical robots, and bio-hybrid robots, it is infeasible to anticipate the operating environment a priori due to complexities that arise from numerous factors including imprecision in manufacturing, chemo-mechanical forces, and poorly understood contact mechanics. Drawing inspiration from data-driven modeling, geometric mechanics (or gauge theory), and adaptive control, we employ an adaptive system identification framework and demonstrate its efficacy in enhancing the performance of principally kinematic locomotors (those governed by Rayleigh dissipation or zero momentum conservation). We showcase the capability of the adaptive model to efficiently accommodate varying terrains and iteratively modified behaviors within a behavior optimization framework. This provides both the ability to improve fundamental behaviors and perform motion tracking to precision. Notably, we are capable of optimizing the gaits of the Purcell swimmer using approximately 10 cycles per link, which for the nine-link Purcell swimmer provides a factor of ten improvement in optimization speed over the state of the art. Beyond simply a computational speed up, this tenfold improvement may enable this method to be successfully deployed for in-situ behavior refinement, injury recovery, and terrain adaptation, particularly in domains where simulations provide poor guides for the real world.

The hybridisation of robot-assisted gait training and functional electrical stimulation (FES) can provide numerous physiological benefits to neurological patients. However, the design of an effective hybrid controller poses significant challenges. In this over-actuated system, it is extremely difficult to find the right balance between robotic assistance and FES that will provide personalised assistance, prevent muscle fatigue and encourage the patient’s active participation in order to accelerate recovery. In this paper, we present an adaptive hybrid robot-FES controller to do this and enable the triadic collaboration between the patient, the robot and FES. A patient-driven controller is designed where the voluntary movement of the patient is prioritised and assistance is provided using FES and the robot in a hierarchical order depending on the patient’s performance and their muscles’ fitness. The performance of this hybrid adaptive controller is tested in simulation and on one healthy subject. Our results indicate an increase in tracking performance with lower overall assistance, and less muscle fatigue when the hybrid adaptive controller is used, compared to its non adaptive equivalent. This suggests that our hybrid adaptive controller may be able to adapt to the behaviour of the user to provide assistance as needed and prevent the early termination of physical therapy due to muscle fatigue.

The hybrid nature of multi-contact robotic systems, due to making and breaking contact with the environment, creates significant challenges for high-quality control. Existing model-based methods typically rely on either good prior knowledge of the multi-contact model or require significant offline model tuning effort, thus resulting in low adaptability and robustness. In this paper, we propose a real-time adaptive multi-contact model predictive control framework, which enables online adaption of the hybrid multi-contact model and continuous improvement of the control performance for contact-rich tasks. This framework includes an adaption module, which continuously learns a residual of the hybrid model to minimize the gap between the prior model and reality, and a real-time multi-contact MPC controller. We demonstrated the effectiveness of the framework in synthetic examples, and applied it on hardware to solve contact-rich manipulation tasks, where a robot uses its end-effector to roll different unknown objects on a table to track given paths. The hardware experiments show that with a rough prior model, the multi-contact MPC controller adapts itself on-the-fly with an adaption rate around 20 Hz and successfully manipulates previously unknown objects with non-smooth surface geometries. Accompanying media can be found at: https://sites.google.com/view/adaptive-contact-implicit-mpc/home

Open-world robotic tasks such as autonomous driving pose significant challenges to robot control due to unknown and unpredictable events that disrupt task performance. Neural network-based reinforcement learning (RL) techniques (like DQN, PPO, SAC, etc.) struggle to adapt in large domains and suffer from catastrophic forgetting. Hybrid planning and RL approaches have shown some promise in handling environmental changes but lack efficiency in accommodation speed. To address this limitation, we propose an enhanced hybrid system with a nested hierarchical action abstraction that can utilize previously acquired skills to effectively tackle unexpected novelties. We show that it can adapt faster and generalize better compared to state-of-the-art RL and hybrid approaches, significantly improving robustness when multiple environmental changes occur at the same time.

Visual servoing (VS) is a widely used technique in industries where there are hundreds of robots, but it requires accurate camera calibration including camera intrinsic and extrinsic parameters. However, it is labour-intensive to calibrate robots one-by-one in practical use. In this paper, we propose a neural uncalibrated VS policy (NUVS) that can adapt to calibration disturbances with an adaption mechanism and a control-oriented guidance. It bridges the disturbance adaption of classical VS methods and the large convergence of learning-based VS methods. NUVS estimates the calibration embedding from past observations and servos to the desired pose under the supervision of a PBVS that can access the ground truth in simulation. With this adaption mechanism, NUVS outperforms the classical IBUVS algorithm when facing large initial camera pose offsets under the calibration disturbance. Supplementary material in: https://sites.google.com/view/neural-uncalibrated-vs

Multi-agent reinforcement learning has emerged as a promising approach for the control of multi-robot systems. Nevertheless, the low sample efficiency of MARL poses a significant obstacle to its broader application in robotics. While data augmentation appears to be a straightforward solution for improving sample efficiency, it usually incurs training instability, making the sample efficiency worse. Moreover, manually choosing suitable augmentations for a variety of tasks is a tedious and time-consuming process. To mitigate these challenges, our research theoretically analyzes the implications of data augmentation on MARL algorithms. Guided by these insights, we present AdaptAUG, an adaptive framework designed to selectively identify beneficial data augmentations, thereby achieving superior sample efficiency and overall performance in multi-robot tasks. Extensive experiments in both simulated and real-world multi-robot scenarios validate the effectiveness of our proposed framework.

Soft tissue tracking is crucial for computer-assisted interventions. Existing approaches mainly rely on extracting discriminative features from the template and videos to recover corresponding matches. However, it is difficult to adopt these techniques in surgical scenes, where tissues are changing in shape and appearance throughout the surgery. To address this problem, we exploit optical flow to naturally capture the pixel-wise tissue deformations and adaptively correct the tracked template. Specifically, we first implement an inter-frame matching mechanism to extract a coarse region of interest based on optical flow from consecutive frames. To accommodate appearance change and alleviate drift, we then propose an adaptive-template matching method, which updates the tracked template based on the reliability of the estimates. Our approach, Ada-Tracker, enjoys both short-term dynamics modeling by capturing local deformations and long-term dynamics modeling by introducing global temporal compensation. We evaluate our approach on the public SurgT benchmark, which is generated from Hamlyn, SCARED, and Kidney boundary datasets. The experimental results show that Ada-Tracker achieves superior accuracy and performs more robustly against prior works. Code is available at https://github.com/wrld/Ada-Tracker.

An open research question in robotics is how to combine the benefits of model-free reinforcement learning (RL)—known for its strong task performance and flexibility in optimizing general reward formulations—with the robustness and online replanning capabilities of model predictive control (MPC). This paper provides an answer by introducing a new framework called Actor-Critic Model Predictive Control. The key idea is to embed a differentiable MPC within an actor-critic RL framework. The proposed approach leverages the short-term predictive optimization capabilities of MPC with the exploratory and end-to-end training properties of RL. The resulting policy effectively manages both short-term decisions through the MPC-based actor and long-term prediction via the critic network, unifying the benefits of both model-based control and end-to-end learning. We validate our method in both simulation and the real world with a quadcopter platform across various high-level tasks. We show that the proposed architecture can achieve real-time control performance, learn complex behaviors via trial and error, and retain the predictive properties of the MPC to better handle out of distribution behaviour.

Mobile Manipulation (MoMa) systems incorporate the benefits of mobility and dexterity, due to the enlarged space in which they can move and interact with their environment. However, even when equipped with onboard sensors, e.g., an embodied camera, extracting task-relevant visual information in unstructured and cluttered environments, such as households, remains challenging. In this work, we introduce an active perception pipeline for mobile manipulators to generate motions that are informative toward manipulation tasks, such as grasping in unknown, cluttered scenes. Our proposed approach, ActPerMoMa, generates robot paths in a receding horizon fashion by sampling paths and computing path-wise utilities. These utilities trade-off maximizing the visual Information Gain (IG) for scene reconstruction and the task-oriented objective, e.g., grasp success, by maximizing grasp reachability. We show the efficacy of our method in simulated experiments with a dual-arm TIAGo++ MoMa robot performing mobile grasping in cluttered scenes with obstacles. We empirically analyze the contribution of various utilities and parameters, and compare against representative baselines both with and without active perception objectives. Finally, we demonstrate the transfer of our mobile grasping strategy to the real world, indicating a promising direction for active-perceptive MoMa.

While existing literature encompasses exoskeleton-assisted sit-stand tasks, the integration of energy recovery mechanisms remains unexplored. To push the boundaries further, this study introduces a portable pneumatic knee exoskeleton that operates in both quasi-passive and active modes, where active mode is utilized for aiding in standing up (power generation), thus the energy flows from the exoskeleton to the user, and quasi-passive mode for aiding in sitting down (power absorption), where the device absorbs and can store energy in the form of compressed air, leading to energy savings in active mode. The absorbed energy can be stored and later reused without compromising exoskeleton transparency in the meantime. In active mode, an air pump inflates the pneumatic artificial muscle (PAM), which stores the compressed air, that can then be released into a pneumatic cylinder to generate torque. All electronic and pneumatic components are integrated into the system, and the exoskeleton weighs 3.9 kg with a maximum torque of 20 Nm at the knee joint. The paper describes the mechatronic design, mathematical model and includes a pilot study with an able-bodied subject performing sit-to-stand tasks. The results show that the exoskeleton can recover energy while assisting the subject and reducing mean muscle activity by ~31%. Further results highlight air regeneration’s potential for energy saving in portable pneumatic exoskeletons, showing that the proposed device extends exoskeleton operation by ~27%.

Rather than having each newly deployed robot create its own map of its surroundings, the growing availability of SLAM-enabled devices provides the option of simply localizing in a map of another robot or device. In cases such as multi-robot or human-robot collaboration, localizing all agents in the same map is even necessary. However, localizing e.g. a ground robot in the map of a drone or head-mounted MR headset presents unique challenges due to viewpoint changes. This work investigates how active visual localization can be used to overcome such challenges of viewpoint changes. Specifically, we focus on the problem of selecting the optimal viewpoint at a given location. We compare existing approaches in the literature with additional proposed baselines and propose a novel data-driven approach. The result demonstrates the superior performance of our data-driven approach when compared to existing methods, both in controlled simulation experiments and real-world deployment.

Merging into dense highway traffic for an autonomous vehicle is a complex decision-making task, wherein the vehicle must identify a potential gap and coordinate with surrounding human drivers, each of whom may exhibit diverse driving behaviors. Many existing methods consider other drivers to be dynamic obstacles and, as a result, they are incapable of capturing the full intent of the human drivers through this passive planning. In this paper, we propose a novel dual control framework based on Model Predictive Path-Integral control to generate interactive trajectories. This framework incorporates a Bayesian inference approach to actively learn the agents’ parameters, i.e., other drivers’ model parameters. The proposed framework employs a sampling-based approach that is suitable for real-time implementation through the utilization of GPUs. We illustrate the effectiveness of our proposed methodology through comprehensive numerical simulations conducted in both high and low-fidelity simulation scenarios focusing on autonomous on-ramp merging.

Reactive planning enables the robots to deal with dynamic events in uncertain environments. However, existing methods heavily rely on the predefined hard-coded robot behaviors, e.g, a pre-coded temporal logic formula that specifies how robot should react. Little attention has been paid for autonomous generation of reactive tasks specifications during the runtime. As a first attempt towards this goal, this work develops a real-time decision-making and motion planning framework. It allows the robot to follow a global task planned offline while taking proactive decisions and generating temporal logic specifications for local reactive tasks when encountering dynamic events. Specifically, inspired by the causal knowledge graph, a proposition graph is developed, based on which the decision module encode the environment and the task as the Boolean logic and linear temporal logic (LTL), respectively. Based on the established proposition graph and perceived environment, the agent can autonomously generate an LTL formula to realize the local temporary task. A joint sampling algorithm is then developed, in which the automaton states of local and global task are jointly considered to generate a feasible planning that satisfies both global and local tasks. Experiments demonstrate the effectiveness of the proposed decision-making and motion planning.

Active object reconstruction using autonomous robots is gaining great interest. A primary goal in this task is to maximize the information of the object to be reconstructed, given limited on-board resources. Previous view planning methods exhibit inefficiency since they rely on an iterative paradigm based on explicit representations, consisting of (1) planning a path to the next-best view only; and (2) requiring a considerable number of less-gain views in terms of surface coverage. To address these limitations, we propose to integrate implicit representations into the One-Shot View Planning (OSVP). The key idea behind our approach is to use implicit representations to obtain the small missing surface areas instead of observing them with extra views. Therefore, we design a deep neural network, named OSVP, to directly predict a set of views given a dense point cloud refined from an initial sparse observation. To train our OSVP network, we generate supervision labels using dense point clouds refined by implicit representations and set covering optimization problems. Simulated experiments show that our method achieves sufficient reconstruction quality, outperforming several baselines under limited view and movement budgets. We further demonstrate the applicability of our approach in a real-world object reconstruction scenario.

Tactile perception is important for robotic systems that interact with the world through touch. Touch is an active sense in which tactile measurements depend on the contact properties of an interaction—e.g., velocity, force, acceleration— as well as properties of the sensor and object under test. These dependencies make training tactile perceptual models challenging. Additionally, the effects of limited sensor life and the near-field nature of tactile sensors preclude the practical collection of exhaustive data sets even for fairly simple objects. Active learning provides a mechanism for focusing on only the most informative aspects of an object during data collection. Here we employ an active learning approach that uses a data-driven model’s entropy as an uncertainty measure and explore relative to that entropy conditioned on the sensor state variables. Using a coverage-based ergodic controller, we train perceptual models in near-real time. We demonstrate our approach using a biomimentic sensor, exploring "tactile scenes" composed of shapes, textures, and objects. Each learned representation provides a perceptual sensor model for a particular tactile scene. Models trained on actively collected data outperform their randomly collected counterparts in real-time training tests. Additionally, we find that the resulting network entropy maps can be used to identify high salience portions of a tactile scene.