arXiv subjects

Yue Wang

Publications and source records attributed to Yue Wang.

At least 757 records · Page 42Linked to original sources

arXiv2021

Efficient learning of goal-oriented push-grasping synergy in clutter

We focus on the task of goal-oriented grasping, in which a robot is supposed to grasp a pre-assigned goal object in clutter and needs some pre-grasp actions such as pushes to enable stable grasps. However, in this task, the robot gets positive rewards from environment only when successfully grasping the goal object. Besides, joint pushing and grasping elongates the action sequence, compounding the problem of reward delay. Thus, sample inefficiency remains a main challenge in this task. In this paper, a goal-conditioned hierarchical reinforcement learning formulation with high sample efficiency is proposed to learn a push-grasping policy for grasping a specific object in clutter. In our work, sample efficiency is improved by two means. First, we use a goal-conditioned mechanism by goal relabeling to enrich the replay buffer. Second, the pushing and grasping policies are respectively regarded as a generator and a discriminator and the pushing policy is trained with supervision of the grasping discriminator, thus densifying pushing rewards. To deal with the problem of distribution mismatch caused by different training settings of two policies, an alternating training stage is added to learn pushing and grasping in turn. A series of experiments carried out in simulation and real world indicate that our method can quickly learn effective pushing and grasping policies and outperforms existing methods in task completion rate and goal grasp success rate by less times of motion. Furthermore, we validate that our system can also adapt to goal-agnostic conditions with better performance. Note that our system can be transferred to the real world without any fine-tuning. Our code is available at https://github.com/xukechun/Efficient_goal-oriented_push-grasping_synergy.

cs.RO↗

arXiv2021

Improving Multi-Modal Learning with Uni-Modal Teachers

Learning multi-modal representations is an essential step towards real-world robotic applications, and various multi-modal fusion models have been developed for this purpose. However, we observe that existing models, whose objectives are mostly based on joint training, often suffer from learning inferior representations of each modality. We name this problem Modality Failure, and hypothesize that the imbalance of modalities and the implicit bias of common objectives in fusion method prevent encoders of each modality from sufficient feature learning. To this end, we propose a new multi-modal learning method, Uni-Modal Teacher, which combines the fusion objective and uni-modal distillation to tackle the modality failure problem. We show that our method not only drastically improves the representation of each modality, but also improves the overall multi-modal task performance. Our method can be effectively generalized to most multi-modal fusion approaches. We achieve more than 3% improvement on the VGGSound audio-visual classification task, as well as improving performance on the NYU depth V2 RGB-D image segmentation task.

cs.LG↗

arXiv2021

Improved Radar Localization on Lidar Maps Using Shared Embedding

We present a heterogeneous localization framework for solving radar global localization and pose tracking on pre-built lidar maps. To bridge the gap of sensing modalities, deep neural networks are constructed to create shared embedding space for radar scans and lidar maps. Herein learned feature embeddings are supportive for similarity measurement, thus improving map retrieval and data matching respectively. In RobotCar and MulRan datasets, we demonstrate the effectiveness of the proposed framework with the comparison to Scan Context and RaLL. In addition, the proposed pose tracking pipeline is with less neural networks compared to the original RaLL.

cs.RO↗

arXiv2021

Radar-to-Lidar: Heterogeneous Place Recognition via Joint Learning

Place recognition is critical for both offline mapping and online localization. However, current single-sensor based place recognition still remains challenging in adverse conditions. In this paper, a heterogeneous measurements based framework is proposed for long-term place recognition, which retrieves the query radar scans from the existing lidar maps. To achieve this, a deep neural network is built with joint training in the learning stage, and then in the testing stage, shared embeddings of radar and lidar are extracted for heterogeneous place recognition. To validate the effectiveness of the proposed method, we conduct tests and generalization experiments on the multi-session public datasets compared to other competitive methods. The experimental results indicate that our model is able to perform multiple place recognitions: lidar-to-lidar, radar-to-radar and radar-to-lidar, while the learned model is trained only once. We also release the source code publicly: https://github.com/ZJUYH/radar-to-lidar-place-recognition.

cs.CV↗

arXiv2021

Sufficient conditions for 2-dimensional global rigidity

The 2-dimensional global rigidity has been shown to be equivalent to 3-connectedness and redundant rigidity by a combination of two results due to Jackson and Jordán, and Connelly, respectively. By the characterization, a theorem of Lovász and Yemini implies that every $6$-connected graph is redundantly rigid, and thus globally rigid. The 6-connectedness is best possible, since there exist infinitely many 5-connected non-rigid graphs. Jackson, Servatius and Servatius used the idea of ``essential connectivity'' and proved that every 4-connected ``essentially 6-connected'' graph is redundantly rigid and thus global rigid. Since 3-connectedness is a necessary condition of global rigidity, it is interesting to study 3-connected graphs for redundant rigidity and thus globally rigidity. We utilize a different ``essential connectivity'', and prove that every 3-connected essentially 9-connected graph is redundantly rigid and thus globally rigid. The essential 9-connectedness is best possible. Under this essential connectivity, we also prove that every 4-connected essentially 6-connected graph is redundantly rigid and thus global rigid. Our proofs are based on discharging arguments.

math.CO↗

arXiv2021

Collaborative Recognition of Feasible Region with Aerial and Ground Robots through DPCN

Ground robots always get collision in that only if they get close to the obstacles, can they sense the danger and take actions, which is usually too late to avoid the crash, causing severe damage to the robots. To address this issue, we present collaboration of aerial and ground robots in recognition of feasible region. Taking the aerial robots' advantages of having large scale variance of view points of the same route which the ground robots is on, the collaboration work provides global information of road segmentation for the ground robot, thus enabling it to obtain feasible region and adjust its pose ahead of time. Under normal circumstance, the transformation between these two devices can be obtained by GPS yet with much error, directly causing inferior influence on recognition of feasible region. Thereby, we utilize the state-of-the-art research achievements in matching heterogeneous sensor measurements called deep phase correlation network(DPCN), which has excellent performance on heterogeneous mapping, to refine the transformation. The network is light-weighted and promising for better generalization. We use Aero-Ground dataset which consists of heterogeneous sensor images and aerial road segmentation images. The results show that our collaborative system has great accuracy, speed and stability.

cs.RO↗

arXiv2021

DNN-Chip Predictor: An Analytical Performance Predictor for DNN Accelerators with Various Dataflows and Hardware Architectures

The recent breakthroughs in deep neural networks (DNNs) have spurred a tremendously increased demand for DNN accelerators. However, designing DNN accelerators is non-trivial as it often takes months/years and requires cross-disciplinary knowledge. To enable fast and effective DNN accelerator development, we propose DNN-Chip Predictor, an analytical performance predictor which can accurately predict DNN accelerators' energy, throughput, and latency prior to their actual implementation. Our Predictor features two highlights: (1) its analytical performance formulation of DNN ASIC/FPGA accelerators facilitates fast design space exploration and optimization; and (2) it supports DNN accelerators with different algorithm-to-hardware mapping methods (i.e., dataflows) and hardware architectures. Experiment results based on 2 DNN models and 3 different ASIC/FPGA implementations show that our DNN-Chip Predictor's predicted performance differs from those of chip measurements of FPGA/ASIC implementation by no more than 17.66% when using different DNN models, hardware architectures, and dataflows. We will release code upon acceptance.

cs.LG↗

arXiv2021

1-Bit Compressive Sensing for Efficient Federated Learning Over the Air

For distributed learning among collaborative users, this paper develops and analyzes a communication-efficient scheme for federated learning (FL) over the air, which incorporates 1-bit compressive sensing (CS) into analog aggregation transmissions. To facilitate design parameter optimization, we theoretically analyze the efficacy of the proposed scheme by deriving a closed-form expression for the expected convergence rate of the FL over the air. Our theoretical results reveal the tradeoff between convergence performance and communication efficiency as a result of the aggregation errors caused by sparsification, dimension reduction, quantization, signal reconstruction and noise. Then, we formulate 1-bit CS based FL over the air as a joint optimization problem to mitigate the impact of these aggregation errors through joint optimal design of worker scheduling and power scaling policy. An enumeration-based method is proposed to solve this non-convex problem, which is optimal but becomes computationally infeasible as the number of devices increases. For scalable computing, we resort to the alternating direction method of multipliers (ADMM) technique to develop an efficient implementation that is suitable for large-scale networks. Simulation results show that our proposed 1-bit CS based FL over the air achieves comparable performance to the ideal case where conventional FL without compression and quantification is applied over error-free aggregation, at much reduced communication overhead and transmission latency.

cs.LG↗

arXiv2021

Toward Consistent Drift-free Visual Inertial Localization on Keyframe Based Map

Global localization is essential for robots to perform further tasks like navigation. In this paper, we propose a new framework to perform global localization based on a filter-based visual-inertial odometry framework MSCKF. To reduce the computation and memory consumption, we only maintain the keyframe poses of the map and employ Schmidt-EKF to update the state. This global localization framework is shown to be able to maintain the consistency of the state estimator. Furthermore, we introduce a re-linearization mechanism during the updating phase. This mechanism could ease the linearization error of observation function to make the state estimation more precise. The experiments show that this mechanism is crucial for large and challenging scenes. Simulations and experiments demonstrate the effectiveness and consistency of our global localization framework.

cs.RO↗

arXiv2021

Neural Motion Prediction for In-flight Uneven Object Catching

In-flight objects capture is extremely challenging. The robot is required to complete trajectory prediction, interception position calculation and motion planning in sequence within tens of milliseconds. As in-flight uneven objects are affected by various kinds of forces, motion prediction is difficult for a time-varying acceleration. In order to compensate the system's non-linearity, we introduce the Neural Acceleration Estimator (NAE) that estimates the varying acceleration by observing a small fragment of previous deflected trajectory. Moreover, end-to-end training with Differantiable Filter (NAE-DF) gives a supervision for measurement uncertainty and further improves the prediction accuracy. Experimental results show that motion prediction with NAE and NAE-DF is superior to other methods and has a good generalization performance on unseen objects. We test our methods on a robot, performing velocity control in real world and respectively achieve 83.3% and 86.7% success rate on a ploy urethane banana and a gourd. We also release an object in-flight dataset containing 1,500 trajectorys for uneven objects.

cs.RO↗

arXiv2021

Learn to Differ: Sim2Real Small Defection Segmentation Network

Recent studies on deep-learning-based small defection segmentation approaches are trained in specific settings and tend to be limited by fixed context. Throughout the training, the network inevitably learns the representation of the background of the training data before figuring out the defection. They underperform in the inference stage once the context changed and can only be solved by training in every new setting. This eventually leads to the limitation in practical robotic applications where contexts keep varying. To cope with this, instead of training a network context by context and hoping it to generalize, why not stop misleading it with any limited context and start training it with pure simulation? In this paper, we propose the network SSDS that learns a way of distinguishing small defections between two images regardless of the context, so that the network can be trained once for all. A small defection detection layer utilizing the pose sensitivity of phase correlation between images is introduced and is followed by an outlier masking layer. The network is trained on randomly generated simulated data with simple shapes and is generalized across the real world. Finally, SSDS is validated on real-world collected data and demonstrates the ability that even when trained in cheap simulation, SSDS can still find small defections in the real world showing the effectiveness and its potential for practical applications.

cs.CV↗

arXiv2021

RaLL: End-to-end Radar Localization on Lidar Map Using Differentiable Measurement Model

Compared to the onboard camera and laser scanner, radar sensor provides lighting and weather invariant sensing, which is naturally suitable for long-term localization under adverse conditions. However, radar data is sparse and noisy, resulting in challenges for radar mapping. On the other hand, the most popular available map currently is built by lidar. In this paper, we propose an end-to-end deep learning framework for Radar Localization on Lidar Map (RaLL) to bridge the gap, which not only achieves the robust radar localization but also exploits the mature lidar mapping technique, thus reducing the cost of radar mapping. We first embed both sensor modals into a common feature space by a neural network. Then multiple offsets are added to the map modal for exhaustive similarity evaluation against the current radar modal, yielding the regression of the current pose. Finally, we apply this differentiable measurement model to a Kalman Filter (KF) to learn the whole sequential localization process in an end-to-end manner. \textit{The whole learning system is differentiable with the network based measurement model at the front-end and KF at the back-end.} To validate the feasibility and effectiveness, we employ multi-session multi-scene datasets collected from the real world, and the results demonstrate that our proposed system achieves superior performance over $90km$ driving, even in generalization scenarios where the model training is in UK, while testing in South Korea. We also release the source code publicly.

cs.RO↗

arXiv2021

Dynamic Movement Primitive based Motion Retargeting for Dual-Arm Sign Language Motions

We aim to develop an efficient programming method for equipping service robots with the skill of performing sign language motions. This paper addresses the problem of transferring complex dual-arm sign language motions characterized by the coordination among arms and hands from human to robot, which is seldom considered in previous studies of motion retargeting techniques. In this paper, we propose a novel motion retargeting method that leverages graph optimization and Dynamic Movement Primitives (DMPs) for this problem. We employ DMPs in a leader-follower manner to parameterize the original trajectories while preserving motion rhythm and relative movements between human body parts, and adopt a three-step optimization procedure to find deformed trajectories for robot motion planning while ensuring feasibility for robot execution. Experimental results of several Chinese Sign Language (CSL) motions have been successfully performed on ABB's YuMi dual-arm collaborative robot (14-DOF) with two 6-DOF Inspire-Robotics' multi-fingered hands, a system with 26 DOFs in total.

cs.RO↗

arXiv2021

REDE: End-to-end Object 6D Pose Robust Estimation Using Differentiable Outliers Elimination

Object 6D pose estimation is a fundamental task in many applications. Conventional methods solve the task by detecting and matching the keypoints, then estimating the pose. Recent efforts bringing deep learning into the problem mainly overcome the vulnerability of conventional methods to environmental variation due to the hand-crafted feature design. However, these methods cannot achieve end-to-end learning and good interpretability at the same time. In this paper, we propose REDE, a novel end-to-end object pose estimator using RGB-D data, which utilizes network for keypoint regression, and a differentiable geometric pose estimator for pose error back-propagation. Besides, to achieve better robustness when outlier keypoint prediction occurs, we further propose a differentiable outliers elimination method that regresses the candidate result and the confidence simultaneously. Via confidence weighted aggregation of multiple candidates, we can reduce the effect from the outliers in the final estimation. Finally, following the conventional method, we apply a learnable refinement process to further improve the estimation. The experimental results on three benchmark datasets show that REDE slightly outperforms the state-of-the-art approaches and is more robust to object occlusion.

cs.CV↗

arXiv2021

Using Prior Knowledge to Guide BERT's Attention in Semantic Textual Matching Tasks

We study the problem of incorporating prior knowledge into a deep Transformer-based model,i.e.,Bidirectional Encoder Representations from Transformers (BERT), to enhance its performance on semantic textual matching tasks. By probing and analyzing what BERT has already known when solving this task, we obtain better understanding of what task-specific knowledge BERT needs the most and where it is most needed. The analysis further motivates us to take a different approach than most existing works. Instead of using prior knowledge to create a new training task for fine-tuning BERT, we directly inject knowledge into BERT's multi-head attention mechanism. This leads us to a simple yet effective approach that enjoys fast training stage as it saves the model from training on additional data or tasks other than the main task. Extensive experiments demonstrate that the proposed knowledge-enhanced BERT is able to consistently improve semantic textual matching performance over the original BERT model, and the performance benefit is most salient when training data is scarce.

cs.CL↗

arXiv2021

Deep Samplable Observation Model for Global Localization and Kidnapping

Global localization and kidnapping are two challenging problems in robot localization. The popular method, Monte Carlo Localization (MCL) addresses the problem by iteratively updating a set of particles with a "sampling-weighting" loop. Sampling is decisive to the performance of MCL [1]. However, traditional MCL can only sample from a uniform distribution over the state space. Although variants of MCL propose different sampling models, they fail to provide an accurate distribution or generalize across scenes. To better deal with these problems, we present a distribution proposal model, named Deep Samplable Observation Model (DSOM). DSOM takes a map and a 2D laser scan as inputs and outputs a conditional multimodal probability distribution of the pose, making the samples more focusing on the regions with higher likelihood. With such samples, the convergence is expected to be more effective and efficient. Considering that the learning-based sampling model may fail to capture the true pose sometimes, we furthermore propose the Adaptive Mixture MCL (AdaM MCL), which deploys a trusty mechanism to adaptively select updating mode for each particle to tolerate this situation. Equipped with DSOM, AdaM MCL can achieve more accurate estimation, faster convergence and better scalability compared to previous methods in both synthetic and real scenes. Even in real environments with long-term changing, AdaM MCL is able to localize the robot using DSOM trained only by simulation observations from a SLAM map or a blueprint map.

cs.RO↗

arXiv2021

Modeling and Leveraging Analytic Focus During Exploratory Visual Analysis

Visual analytics systems enable highly interactive exploratory data analysis. Across a range of fields, these technologies have been successfully employed to help users learn from complex data. However, these same exploratory visualization techniques make it easy for users to discover spurious findings. This paper proposes new methods to monitor a user's analytic focus during visual analysis of structured datasets and use it to surface relevant articles that contextualize the visualized findings. Motivated by interactive analyses of electronic health data, this paper introduces a formal model of analytic focus, a computational approach to dynamically update the focus model at the time of user interaction, and a prototype application that leverages this model to surface relevant medical publications to users during visual analysis of a large corpus of medical records. Evaluation results with 24 users show that the modeling approach has high levels of accuracy and is able to surface highly relevant medical abstracts.

cs.HC↗

arXiv2021

PREGAN: Pose Randomization and Estimation for Weakly Paired Image Style Translation

Utilizing the trained model under different conditions without data annotation is attractive for robot applications. Towards this goal, one class of methods is to translate the image style from another environment to the one on which models are trained. In this paper, we propose a weakly-paired setting for the style translation, where the content in the two images is aligned with errors in poses. These images could be acquired by different sensors in different conditions that share an overlapping region, e.g. with LiDAR or stereo cameras, from sunny days or foggy nights. We consider this setting to be more practical with: (i) easier labeling than the paired data; (ii) better interpretability and detail retrieval than the unpaired data. To translate across such images, we propose PREGAN to train a style translator by intentionally transforming the two images with a random pose, and to estimate the given random pose by differentiable non-trainable pose estimator given that the more aligned in style, the better the estimated result is. Such adversarial training enforces the network to learn the style translation, avoiding being entangled with other variations. Finally, PREGAN is validated on both simulated and real-world collected data to show the effectiveness. Results on down-stream tasks, classification, road segmentation, object detection, and feature matching show its potential for real applications. https://github.com/wrld/PRoGAN

cs.CV↗