IROS 2023 论文列表

Soft robots are gaining popularity thanks to their intrinsic safety to contacts and adaptability. However, the potentially infinite number of Degrees of Freedom makes their modeling a daunting task, and in many cases only an approximated description is available. This challenge makes reinforcement learning (RL) based approaches inefficient when deployed on a realistic scenario, due to the large domain gap between models and the real platform. In this work, we demonstrate, for the first time, how Domain Randomization (DR) can solve this problem by enhancing RL policies for soft robots with: i) robustness w.r.t. unknown dynamics parameters; ii) reduced training times by exploiting drastically simpler dynamic models for learning; iii) better environment exploration, which can lead to exploitation of environmental constraints for optimal performance. Moreover, we introduce a novel algorithmic extension to previous adaptive domain randomization methods for the automatic inference of dynamics parameters for deformable objects. We provide an extensive evaluation in simulation on four different tasks and two soft robot designs, opening interesting perspectives for future research on Reinforcement Learning for closed-loop soft robot control.

Training with simulated data is a common approach in pose estimation research. However, a sim-to-real gap between clean simulated data and noisy real data will seriously weaken the generalization ability of the algorithm, especially for point clouds. To address this problem, this paper proposes a domain adaptive pose estimation network (DAPE-Net). For the feature extracted from the backbone, the network will conduct the real and simulation discrimination based on a feature discriminator, and complete the pose estimation by adversarial training. This makes the network pay more attention to the domain invariant features of simulation and real point clouds to complete domain adaptation. In our experiment, DAPE-Net improved the performance of pose estimation by 10%. To solve the problem that domain adaptation requires a small amount of real data, we propose a scheme that can semi-automatically collect real data in bin-picking scenarios for 6D pose estimation.

In this paper we investigate the effect of the unpredictability of surrounding cars on an ego-car performing a driving maneuver. We use Maximum Entropy Inverse reinforcement Learning to model reward functions for an ego-car conducting a lane change in a highway setting. We define a new feature based on the unpredictability of surrounding cars and use it in the reward function. We learn two reward functions from human data: a baseline and one that incorporates our defined unpredictability feature, then compare their performance with a quantitative and qualitative evaluation. Our evaluation demonstrates that incorporating the unpredictability feature leads to a better fit of human-generated test data. These results encourage further investigation of the effect of unpredictability on driving behavior.

It is crucial for hybrid unmanned aerial vehicles, such as lifting-wing multicopters, to plan a continuous, smooth, and collision-free trajectory to avoid obstacles. Unlike quad-copters, which typically work in indoor environments, lifting-wing multicopters typically fly at a high altitude with a high cruising speed, requiring higher maneuverability in the vertical direction. Inspired by birds, lifting-wing multicopters can take an inverted flight maneuver to gain more maneuverability than the corresponding multicopter owing to the additional lifting wing. In this paper, a rotation-aware collision-free motion planning strategy is proposed that takes aerodynamics into consideration and allows lifting-wing multicopters to fly at large rotation angles, even in inverted postures. Specifically, a collision-free state sequence is found using rotation-aware primitives by solving a graph search problem. The sequence is then refined with B-spline into smooth trajectories to be tracked by the differential flatness-based controller for lifting-wing multicopters. We analyze the proposed motion planning algorithm in different scenarios and demonstrate the feasibility of the generated trajectories in simulation and real-world experiments. Video: https://youtu.be/n87jK81zg_I

Deep learning methods are widely used in robotic applications. By learning from prior experience, the robot can abstract knowledge of the environment, and use this knowledge to accomplish different goals, such as object search, frontier exploration, or scene understanding, with a smaller amount of resources than might be needed without that knowledge. Most existing methods typically require a significant amount of sensing, which in turn has significant costs in terms of power consumption for acquisition and processing, and typically focus on models that are tuned for each specific goal, leading to the need to train, store and run each one separately. These issues are particularly important in a resource-constrained setting, such as with small-scale robots or during long-duration missions. We propose a single, multi-task deep learning architecture that takes advantage of the structure of the partial environment to predict different abstractions of the environment (thus reducing the need for rich sensing), and to leverage these predictions to simultaneously achieve different high-level goals (thus sharing computation between goals). As an example application of the proposed architecture, we consider the specific example of a robot equipped with a 2-D laser scanner and an object detector, tasked with searching for an object (such as an exit) in a residential building while constructing a topological map that can be used for future missions. The prior knowledge of the environment is encoded using a U-Net deep network architecture. In this context, our work leads to an object search algorithm that is complete, and that outperforms a more traditional frontier-based approach. The topological map we produce uses scene trees to qualitatively represent the environment as a graph at a fraction of the cost of existing SLAM-based solutions. Our results demonstrate that it is possible to extract multi-task semantic information that is useful for navigation and mapping directly fr... Show More

Customer satisfaction is not only determined by how workers treat customers but also by how they treat their coworkers. In this sense, we examined whether the dynamics of robot workers influence user satisfaction. To optimize the robot team's atmosphere for Korean culture, we adopted the Korean honorific language system to express hierarchy. We set four types of relationships between two robots: Case 1) both using non honorific language, indicating intimacy and a flat hierarchy; Case 2) both using honorific language, indicating non intimacy and a flat hierarchy; Case 3) delivery robot using non honorific language and monitoring robot using honorific language, indicating non intimacy and a tall hierarchy; Case 4) vice versa. People preferred a robot team with a flat hierarchy and both robots using honorific language, common in the service industry. Interviews revealed discomfort with tall hierarchies. Case 4 was least satisfactory in service, while Case 3 was least socially intelligent. People perceive a robot team mirroring human society as socially intelligent, but negative representations like Case 4 led to less satisfaction.

Operating in the near-vicinity of marine energy devices poses significant challenges to the control of underwater vehicles, predominantly due to the presence of large magnitude wave disturbances causing hazardous state perturbations. Approaches to tackle this problem have varied, but one promising solution is to adopt predictive control methods. Given the predictable nature of ocean waves, the potential exists to incorporate disturbance estimations directly within the plant model; this requires inclusion of a wave predictor to provide online preview information. To this end, this paper presents a Non-linear Model Predictive Controller with an integrated Deterministic Sea Wave Predictor for trajectory tracking of underwater vehicles. State information is obtained through an Extended Kalman Filter, forming a complete closed-loop strategy and facilitating online wave load estimations. The strategy is compared to a similar feed-forward disturbance mitigation scheme, showing mean performance improvements of 51% in positional error and 44.5% in attitude error. The preliminary results presented here provide strong evidence of the proposed method's high potential to effectively mitigate disturbances, facilitating accurate tracking performance even in the presence of high wave loading.

In this paper, we propose a Model Predictive Control (MPC) based distributed formation control method for a multi-robot system (MRS) that would move them among dynamic obstacles to a desired goal position. Specifically, after formulating the formation control, as a distributed version of MPC, we propose and evaluate three information-sharing schemes within the MRS; namely sharing (i) positions, (ii) complete predicted trajectories, and (iii) exponentially-sampled predicted trajectories. Using a simplified kinematic model for robots, we conducted systematic simulation experiments in (a) scenarios, where the robots are instructed to switch places, as one of the most challenging forms of formation changes, and in (b) scenarios where robots are instructed to reach a goal, within environments containing dynamic obstacles. In a set of systematic experiments conducted in simulation and with mini quadcopters, we have shown that sharing of exponentially-sampled trajectories (as opposed to positions, or complete trajectories) among the robots provides near-optimal paths while decreasing the required computation cost and communication bandwidth. Surprisingly, in the presence of noise, sharing exponentially-sampled trajectories among the robots decreased the variance in the final paths. The proposed method is demonstrated on a group of Crazyflie quadcopters.

Object detection plays an important role in computer vision tasks such as autonomous driving, robotics, etc. Typically, a detection model is firstly trained on collected data and then deployed in real world. However, the discrepancy exists between training (source) and testing (target) data, which degrades the detection model's performance in the real world. To mitigate the negative effects, Unsupervised Domain Adaptation (UDA) methods learn the features of a shared domain via a discriminator. However, existing discriminators consider only the in-distribution adversarial learning, which ignore the out-of-distribution data of individual domains. In this paper, we propose a disentangled discriminator to consider the in-distribution and outliers separately. It aligns the source and target data with split branches under a gated strategy. We combine the disentangled discriminator with a Teacher-Student (T-S) framework that trains the student using labeled source data and unlabeled target data under a self-training mechanism. Specifically, the teacher network, that is updated with the parameters of student network via the exponential moving average, predicts pseudo labels for unlabeled data. The quality of pseudo labels can be improved after alleviating the domain discrepancy thanks to the disentangled discriminator. Extensive experiments on benchmarks demonstrate the superiority of the proposed method. Specifically, we achieve 53.9% mAP on Foggy Cityscapes, which is 7.2% higher than the Oracle.

Discrete-time architectures have an advantage over their continuous counterparts as they can be directly executed on embedded hardware without the need for dis-cretization, and discretization can result in a loss of stability margin. This paper presents a discrete-time adaptive control architecture for uncertain scalar multi agent systems with coupled dynamics. Our strategy includes observer dynamics to handle the unmeasurable coupled dynamics and a user-assigned Laplacian matrix for coordination of the multiagent system. An illustrative example is presented to show the efficacy of the proposed control architectures.

Autonomous robots deployed in the real world will need control policies that rapidly adapt to environmental changes. To this end, we propose AutoRobotics-Zero (ARZ), a method based on AutoML-Zero that discovers zero-shot adaptable policies from scratch. In contrast to neural network adaption policies, where only model parameters are optimized, ARZ can build control algorithms with the full expressive power of a linear register machine. We evolve modular policies that tune their model parameters and alter their inference algorithm on-the-fly to adapt to sudden environmental changes. We demonstrate our method on a realistic simulated quadruped robot, for which we evolve safe control policies that avoid falling when individual limbs suddenly break. This is a challenging task in which two popular neural network baselines fail. Finally, we conduct a detailed analysis of our method on a novel and challenging non-stationary control task dubbed Cataclysmic Cartpole. Results confirm our findings that ARZ is significantly more robust to sudden environmental changes and can build simple, interpretable control policies.

Can knowing where you are assist in perceiving objects in your surroundings, especially under adverse weather and lighting conditions? In this work we investigate whether a prior map can be leveraged to aid in the detection of dynamic objects in a scene without the need for a 3D map or pixel-level map-query correspondences. We contribute an algorithm which refines an initial set of candidate object detections and produces a refined subset of highly accurate detections using a prior map. We begin by using visual place recognition (VPR) to retrieve a prior map image for a given query image, then use a binary classification neural network that compares the query and prior map image regions to validate the query detection. Once our classification network is trained, on approximately 1000 query-map image pairs, it is able to improve the performance of vehicle detection when combined with an existing off-the-shelf vehicle detector. We demonstrate our approach using standard datasets across two cities (Oxford and Zurich) under different settings of train-test separation of map-query traverse pairs. We further emphasize the performance gains of our approach against alternative design choices and show that VPR suffices for the task, eliminating the need for precise ground truth localization.

Automated Guided Vehicles (AGV) are omnipresent, and are able to carry out various kind of preprogrammed tasks. Unfortunately, a lot of manual configuration is still required in order to make these systems operational, and configuration needs to be re-done when the environment or task is changed. As an alternative to current inflexible methods, we employ a learning based method in order to perform directed exploration of a previously unseen environment. Instead of relying on handcrafted heuristic representations, the agent learns its own environmental representation through its embodiment. Our method offers loose coupling between the Reinforcement Learning (RL) agent, which is trained in simulation, and a separate, on real-world images trained task module. The uncertainty of the task module is used to direct the exploration behavior. As an example, we use a warehouse inventory task, and we show how directed exploration can improve the task performance through active data collection. We also propose a novel environment representation to efficiently tackle the sim2real gap in both sensing and actuation. We empirically evaluate the approach both in simulated environments and a real-world warehouse.

Thanks to the development of 2D keypoint detectors, monocular 3D human pose estimation (HPE) via 2D-to-3D uplifting approaches have achieved remarkable improvements. Still, monocular 3D HPE is a challenging problem due to the inherent depth ambiguities and occlusions. To handle this problem, many previous works exploit temporal information to mitigate such difficulties. However, there are many real-world applications where frame sequences are not accessible. This paper focuses on reconstructing a 3D pose from a single 2D keypoint detection. Rather than exploiting temporal information, we alleviate the depth ambiguity by generating multiple 3D pose candidates which can be mapped to an identical 2D keypoint. We build a novel diffusion-based framework to effectively sample diverse 3D poses from an off-the-shelf 2D detector. By considering the correlation between human joints by replacing the conventional denoising U-Net with graph convolutional network, our approach accomplishes further performance improvements. We evaluate our method on the widely adopted Human3.6M and HumanEva-I datasets. Comprehensive experiments are conducted to prove the efficacy of the proposed method, and they confirm that our model outperforms state-of-the-art multi-hypothesis 3D HPE methods.

Task and motion planning is one of the key problems in robotics today. It is often formulated as a discrete task allocation problem combined with continuous motion planning. Many existing approaches to TAMP involve explicit descriptions of task primitives that cause discrete changes in the kinematic relationship between the actor and the objects. In this work we propose an alternative, fully differentiable approach which supports a large number of TAMP problem instances. Rather than explicitly enumerating task primitives, actions are instead represented implicitly as part of the solution to a nonlinear optimization problem. We focus on decision making for robotic manipulators, specifically for pick and place tasks, and explore the efficacy of the model through a number of simulated experiments including multiple robots, objects and interactions with the environment. We also show several possible extensions.

Differentiable Simulations have recently proven useful for various robotic manipulation tasks, including cloth manipulation. In robotic cloth simulation, it is crucial to maintain intersection-free properties. We present DiffClothAI, a differentiable cloth simulation with intersection-free friction contact and two-way coupling with articulated rigid bodies. DiffClothAI integrates the Project Dynamics and Incremental Potential Contact coherently and proposes an effective method to derive gradients in the Cloth Simulation. It also establishes the differentiable coupling mechanism between articulated rigid bodies and cloth. We conduct a comprehensive evaluation of DiffClothAI's effectiveness and accuracy and perform a variety of experiments in downstream robotic manipulation tasks. Supplemental materials and videos are available on our project webpage at https://sites.google.com/view/diffsimcloth.

This paper presents a feedback-control framework for in-hand manipulation (IHM) with dexterous soft hands that enables the acquisition of manipulation skills in the real-world within minutes. We choose the deformation state of the soft hand as the control variable. To control for a desired deformation state, we use coarsley approximated Jacobians of the actuation-deformation dynamics. These Jacobian are obtained via explorative actions. This is enabled by the self-stabilizing properties of compliant hands, which allow us to use linear feedback control in the presence of complex contact dynamics. To evaluate the effectiveness of our approach, we show the generalization capabilities for a learned manipulation skill to variations in object size by 100 %, 360 degree changes in palm inclination and to disabling up to 50 % of the involved actuators. In addition, complex manipulations can be obtained by sequencing such feedback-skills.

Robotic dexterous grasping is a challenging problem due to the high degree of freedom (DoF) and complex contacts of multi-fingered robotic hands. Existing deep re-inforcement learning (DRL) based methods leverage human demonstrations to reduce sample complexity due to the high dimensional action space with dexterous grasping. However, less attention has been paid to hand-object interaction representations for high-level generalization. In this paper, we propose a novel geometric and spatial hand-object interaction representation, named DexRep, to capture object surface features and the spatial relations between hands and objects during grasping. DexRep comprises Occupancy Feature for rough shapes within sensing range by moving hands, Surface Feature for changing hand-object surface distances, and LocalGeo Feature for local geometric surface features most related to potential contacts. Based on the new representation, we propose a dexterous deep reinforcement learning method DexRepNet to learn a generalizable grasping policy. Experimental results show that our method outperforms baselines using existing representations for robotic grasping dramatically both in grasp success rate and convergence speed. It achieves a 93% grasping success rate on seen objects and higher than 80% grasping success rates on diverse objects of unseen categories in both simulation and real-world experiments.

We need to build robot systems that can operate in multiple environments, including underwater. For a robot to be waterproof, it needs to be covered all over its body and have a waterproof structure, but this is expensive to produce and consumes hardware resources such as weight. In this study, we propose a method of constructing a robot system in which paraffin, which can change its phase between solid and liquid at different temperatures, is cloaked on the robot's body surface, and grease is inserted into the robot's body to give it an acquired the waterproof shell for waterproofing purposes. We have studied an automated method of the waterproof shell application system by combining a small spider-shaped multi-legged robot that can measure the temperature of the environment as a robot to which we apply the waterproof shell and a life-size arm robot that performs the operation applying the waterproof shell to the small robot as a system. Through the addition of the waterproof shell to the spider-shaped robot and the removing the shell in unnecessary areas by controlling temperature, and the realization of operation in land and water environments, we used the proposed robot system to add acquired functions to operate underwater robots, thereby expanding the supported area of robots which have many DOFs and sensors from conventional ground to underwater.