基于深度强化学习的固体放射性废物抓取方法研究

针对固体放射性废物分拣作业中,放射性废物杂乱无序、远程遥操作抓取效率低、人工分拣危险性大等典型问题,提出一种基于深度强化学习的放射性固体废物抓取方法。该方法使用改进深度Q网络算法,通过获取的图像信息,使机器人与环境不断进行交互并获得回报奖励,回报奖励由机械臂动作执行结果和放射性区域内放射性活度的高低构成,根据◢Q◣值的大小得到机械臂的最佳抓取位置。用V-REP软件对UR5机械臂建立仿真模型,在仿真环境中完成不同类型固体放射性废物抓取的训练与测试。仿真结果表明,固体废物在松散放置时该方法可使机械臂抓取成功率大于90%,在紧密放置时抓取成功率大于65%,机械臂不会受到废物堆叠的影响,并且会优先抓取放射性区域内具有高放射性活度的物体。     

基于深度神经网络的多指灵巧手抓取手势优化

基于深度神经网络模型,提出了一种适用于多指灵巧手的抓取手势优化方法。首先,在仿真环境下构建了一个抓取数据集,并在此基础上训练了一个卷积神经网络,依据目标物体单目视觉信息和多指灵巧手抓取位形来预测抓取质量函数,由此可以将多指灵巧手的抓取规划问题转化为使抓取质量最大化的优化问题,进一步,基于深度神经网络中的反向传播和梯度上升算法实现多指灵巧手抓取手势的迭代与优化。在仿真环境中,比较该网络和仿真平台对同一抓取位形的抓取质量评估结果,再利用所提出的优化方法对随机搜索到的初始手势进行优化,比较优化前后手势的力封闭指标。最后,在实际机器人平台上验证本文方法的优化效果,结果表明,本文方法对未知物体的抓取成功率在80%以上,对于失败的抓取,优化后成功的比例达到90%。     

一种先验知识引导的基于二阶段渐进网络的自主抓取策略

针对未知不规则物体在堆叠场景下的抓取任务,提出一种基于二阶段渐进网络（two-stage progressive network,TSPN）的自主抓取方法．首先利用端对端策略获取全局可抓性分布,然后基于采样评估策略确定最优抓取配置．将以上2种策略融合,使得TSPN的结构更加精简,显著减少了需评估样本的数量,能够在保证泛化能力的同时提升抓取效率．为了加快抓取模型学习进程,引入一种先验知识引导的自监督学习策略,并利用220种不规则物体进行抓取学习．在仿真和真实环境下分别进行实验,结果表明该抓取模型适用于多物体、堆叠物体、未知不规则物体、物体位姿随机等多种抓取场景,其抓取准确率和探测速度较其他基准方法有明显提升．整个学习过程历时10天,结果表明使用先验知识引导的学习策略能显著加快学习进程．     

基于多重几何约束的未知物体抓取位姿估计

针对机器人在非结构化环境下面临的未知物体难以快速稳定抓取的问题,提出一种基于多重几何约束的未知物体抓取位姿估计方法.通过深度相机获取场景的几何点云信息,对点云进行预处理得到目标物体,利用简化的夹持器几何形状约束生成抓取位姿样本.然后,利用简化的力封闭约束对样本进行快速粗筛选.对抓取位姿的抓取几何轮廓进行力平衡约束分析,将稳定的位姿传送至机器人执行抓取.采用深度相机与6自由度机械臂组成实验平台,对不同姿态形状的物体进行抓取实验.实验结果表明,本文方法能够有效应对物体种类繁多、缺乏3维模型的情况,在单目标和多目标场景均具有良好的适用性.     

杂乱场景下小物体抓取检测研究

目的 杂乱场景下的物体抓取姿态检测是智能机器人的一项基本技能。尽管六自由度抓取学习取得了进展,但先前的方法在采样和学习中忽略了物体尺寸差异,导致在小物体上抓取表现较差。方法 提出了一种物体掩码辅助采样方法,在所有物体上采样相同的点以平衡抓取分布,解决了采样点分布不均匀问题。此外,学习时采用多尺度学习策略,在物体部分点云上使用多尺度圆柱分组以提升局部几何表示能力,解决了由物体尺度差异导致的学习抓取操作参数困难问题。通过设计一个端到端的抓取网络,嵌入了提出的采样和学习方法,能够有效提升物体抓取检测性能。结果 在大型基准数据集GraspNet-1Billion上进行评估,本文方法取得对比方法中的最优性能,其中在小物体上的抓取指标平均提升了7%,大量的真实机器人实验也表明该方法具有抓取未知物体的良好泛化性能。结论 本文聚焦于小物体上的抓取,提出了一种掩码辅助采样方法嵌入到提出的端到端学习网络中,并引入了多尺度分组学习策略提高物体的局部几何表示,能够有效提升在小尺寸物体上的抓取质量,并在所有物体上的抓取评估结果都超过了对比方法。     

基于多尺度特征融合的抓取位姿估计

抓取目标多样性、位姿随机性严重制约了机器人抓取的任务适应性,为提高机器人抓取成功率,提出一种融合多尺度特征的机器人抓取位姿估计方法.该方法以RGD信息为输入,采用ResNet-50主干网络,融合FPN(feature pyramid networks)获得多尺度特征作为抓取生成网络的输入,以生成抓取候选框;并将抓取方向...     

A Probabilistic Framework for Uncertainty-Aware High-Accuracy Precision Grasping of Unknown Objects

Grasping is a fundamental skill for robots which work for manipulation tasks. Grasping of unknown objects remains a big challenge. Precision grasping of unknown objects is even harder. Due to imperfection of sensor measurements and lack of prior knowledge of objects, robots have to handle the uncertainty effectively. In previous work (Chen and Wichert 2015), we use a probabilistic framework to tackle precision grasping of model-based objects. In this paper, we extend the probabilistic framework to tackle the problem of precision grasping of unknown objects. We first propose an object model called probabilistic signed distance function (p-SDF) to represent unknown object surface. p-SDF models measurement uncertainty explicitly and allows measurement from multiple sensors to be fused in real time. Based on the surface representation, we propose a model to evaluate the likelihood of grasp success for antipodal grasps. This model uses four heuristics to model the condition of force closure and perceptual uncertainty. A two step simulated annealing approach is further proposed to search and optimize a precision grasp. We use the object representation as a bridge to unify grasp synthesis and grasp execution. Our grasp execution is performed in a closed-loop, so that robots can actively reduce the uncertainty and react to external perturbations during a grasping process. We perform extensive grasping experiments using real world challenging objects and demonstrate that our method achieves high robustness and accuracy in grasping unknown objects.     

Deep learning-based method for vision-guided robotic grasping of unknown objects

Nowadays, robots are heavily used in factories for different tasks, most of them including grasping and manipulation of generic objects in unstructured scenarios. In order to better mimic a human operator involved in a grasping action, where he/she needs to identify the object and detect an optimal grasp by means of visual information, a widely adopted sensing solution is Artificial Vision. Nonetheless, state-of-art applications need long training and fine-tuning for manually build the object’s model that is used at run-time during the normal operations, which reduce the overall operational throughput of the robotic system. To overcome such limits, the paper presents a framework based on Deep Convolutional Neural Networks (DCNN) to predict both single and multiple grasp poses for multiple objects all at once, using a single RGB image as input. Thanks to a novel loss function, our framework is trained in an end-to-end fashion and matches state-of-art accuracy with a substantially smaller architecture, which gives unprecedented real-time performances during experimental tests, and makes the application reliable for working on real robots. The system has been implemented using the ROS framework and tested on a Baxter collaborative robot.     

Robust grasping strategy for assembling parts in various shapes

In a robotic cell, assembly robots have to grasp parts in various shapes robustly and accurately even under some uncertainties in the initial poses of the parts. For this purpose, it is necessary to develop a universal robotic hand and robust grasping strategies, i.e. finger motions that can achieve planned grasping robustly against the initial pose uncertainty of parts. In this paper, we propose a methodology to plan robust grasping strategies of a universal robotic hand for assembling parts in various shapes. In our approach, parts are aligned toward planned configurations during grasping actions, and the robustness of grasping strategies is analyzed and evaluated based on pushing operation analysis. As an application example, we plan robust grasping strategies for assembling a three-dimensional puzzle, and experimentally verify the robustness and effectiveness of the planned strategies for this assembly task.     

Improving monocular visual SLAM in dynamic environments: an optical-flow-based approach

Visual Simultaneous Localization and Mapping (visual SLAM) has attracted more and more researchers in recent decades and many state-of-the-art algorithms have been proposed with rather satisfactory performance in static scenarios. However, in dynamic scenarios, the performance of current visual SLAM algorithms degrades significantly due to the disturbance of the dynamic objects. To address this problem, we propose a novel method which uses optical flow to distinguish and eliminate the dynamic feature points from the extracted ones using RGB images as the only input. The static feature points are fed into the visual SLAM system for the camera pose estimation. We integrate our method with the original ORB-SLAM system and validate the proposed method with the challenging dynamic sequences from the TUM dataset and our recorded office dataset. The whole system can work in real time. Qualitative and quantitative evaluations demonstrate that our method significantly improves the performance of ORB-SLAM in dynamic scenarios.     

一种基于抓取簇和碰撞体素的工业零件抓取姿态检测算法

针对工业上常见的散乱堆叠零件的抓取问题,提出一种基于抓取簇和碰撞体素的抓取姿态检测算法。所提出的抓取簇是定义在零件上的连续抓取姿态集合,解决了传统方法中因采用离散固定抓取点而导致可抓取点丢失、筛选效率低的问题。先对料箱和场景点云进行体素化;然后把包含料箱或点云的体素标记为碰撞体素,并把与碰撞体素相邻的体素标记为风险体素,从而建立体素化的碰撞模型;接下来,根据抓取簇的几何性质计算出候选的抓取姿态及其对应的抓取路径;最后,通过检测抓取路径所经过的体素类型来完成快速的碰撞检测,从而筛选出最优抓取姿态。基于所提算法搭建了完整的Bin-Picking系统,并对多种实际工业场景中常见的零件进行仿真实验和实际抓取实验,结果表明：该算法能够快速、准确地检测出安全的抓取姿态,实际抓取的平均成功率达92.2%,料箱平均清空率达87.2%,较传统方法有明显提升,且抓取过程均未发生碰撞,可满足实际工业应用的要求。     

Design and analysis of a flexible robotic hand with soft fingers and a changeable palm

This study describes the design of a novel flexible robotic hand that can adapt its configurations to different grasping demands. Firstly, a mathematical model, based on the Yeoh strain energy function and virtual work principle, is established to investigate deformation properties of the designed soft finger. To achieve a flexible grasping capability, a changeable palm is presented with its variable configurations in terms of target objects with different sizes and shapes. A kinematic model of the flexible robotic hand is established, and then the numerical simulations based on the Monte-Carlo method and Matlab is applied to analyse the workspace of the hand and address the parameter optimisation problem of the rigid-flexible coupled system. Furthermore, an optimised grasping strategy on the basis of the principle of optimal efficiency is proposed to obtain an optimal grasping pose for the target object. Finally, a prototype is developed and tested in a laboratory to demonstrate the feasibility and effectiveness of our proposed hand. The results of practical experiments show that the robotic hand cannot only stably grasp objects with different sizes and shapes but also flexibly manipulate soft and fragile ones.     

基于点云采样权重估计的未知物体抓取位姿生成方法

针对机械臂在非结构环境中对未知物体抓取位姿生成困难及抓取稳定性差的问题,提出一种基于点云采样权重估计的抓取位姿生成方法.首先通过移动深度相机的方式拼接得到较完整的物体点云信息,并对物体的几何特性进行分析,有效避开物体不宜抓取的位置进行抓取位姿样本生成;然后结合几何约束条件实现抓取位姿搜索,并利用力封闭条件对样本稳定性进行评估;最后为了对实际的抓取位姿进行评价,根据其稳定性、夹取深度、夹取角度等设定抓取可行性指标,据此在工作空间输出最佳抓取位姿并完成指定的抓取任务.实验结果表明,采用所提方法能够高效生成大量且稳定的抓取位姿,并在仿真环境中有效实现机械臂对单个或多个随机摆放的未知物体的抓取任务.     

Attribute filter based infrared and visible image fusion

Infrared and visible image fusion is an effective image processing technique to obtain more comprehensive information, which can help people better understand various scenarios. In this paper, a novel infrared and visible image fusion method is proposed which fully considers the attributes of objects in source images. Benefiting from the attribute and the edge-preserving filters, the prominent objects in the infrared source image are effectively extracted. Then, the weight-based Laplacian pyramid fusion strategy is adopted to get more natural fusion results. The experimental results on the public image fusion datasets and a new infrared–visible video fusion dataset show that the proposed method achieves state-of-the-art fusion performance in terms of both visual and objective evaluations. The proposed algorithm is also implemented in an infrared–visible dual sensor system, which demonstrated the practicability of our fusion method.     

Automatically composing and parameterizing skills by evolving Finite State Automata

We propose a robotics algorithm that is able to simultaneously combine, adapt and create actions to solve a task. The actions are combined in a Finite State Automaton whose structure is determined by a novel evolutionary algorithm. The actions parameters, or new actions, are evolved alongside the FSA topology. Actions can be combined together in a hierarchical fashion. This approach relies on skills that with which the robot is already provided, like grasping or motion planning. Therefore software reuse is an important advantage of our proposed approach. We conducted several experiments both in simulation and on a real mobile manipulator PR2 robot, where skills of increasing complexity are evolved. Our results show that (i) an FSA generated in simulation can be directly applied to a real robot without modifications and (ii) the evolved FSA is robust to the noise and the uncertainty arising from real-world sensors.     

异质信息网络中基于元路径的社团发现算法研究

实际的网络化数据往往包含多种类型的对象和关系,采用异质信息网络可以更好地对其建模,因此异质信息网络分析逐渐成为数据挖掘的研究热点。虽然同质信息网络中的社团发现已经被深入研究,但是异质信息网络中的社团发现还很少被研究。该文研究异质信息网络中的社团发现问题,提出了一个新的社团发现算法框架HCD(heterogeneous community detection)。该框架由两部分组成: 基于单条元路径的社团发现算法HCD_sgl和融合多条元路径的社团发现算法HCD_all。HCD_sgl首先确定在给定元路径下所有节点的初始标签,再利用改进的标签传递算法进行最终的社团发现;HCD_all是在HCD_sgl的基础上将基于多条元路径的社团发现结果进行融合。通过在真实数据集和人工数据集上的实验验证了HCD算法的有效性。     

Multi-class grasping classifiers using EEG data and a common spatial pattern filter

This paper describes the classification of various human actions from brain activity. In particular, we focus on grasping movements and estimate grasping patterns from electroencephalogram (EEG) data. EEG data is converted to grasping features by using a common spatial pattern filter (CSP filter), and the features are subsequently classified into grasping categories by using the k-nearest neighbor method. We tested the pipeline of feature extraction and classification on the EEG dataset. The EEG data were acquired while participants grasped an object according to the Cutkosky’s grasping taxonomy, in which grasping movements are categorized into nine power-type grasping patterns and seven precision-type grasping patterns. The best classification rate for 9-class power-type grasping patterns was 48% and for 7-class precision-type grasping patterns was 40%.     

融入注意力机制的多模特征机械臂抓取位姿检测

针对机械臂抓取检测任务中对未知物体抓取位姿检测精度低、耗时长等问题,提出一种融入注意力机制多模特征抓取位姿检测网络.首先,设计多模态特征融合模块,在融合多模态特征同时对其赋权加强;然后,针对较浅层残差网络提取重点特征能力较弱的问题,引入卷积注意力模块,进一步提升网络特征提取能力;最后,通过全连接层对提取特征直接进行回归拟合,得到最优抓取检测位姿.实验结果表明,在Cornell公开抓取数据集上,所提出算法的图像拆分检测精度为98.9%,对象拆分检测精度为98.7%,检测速度为51FPS,对10类物体的100次真实抓取实验中,成功率为95%.     

Shape Morphing-Based Control of Robotic Visual Servoing

We present an approach for controlling robotic interactions with objects, using synthetic images generated by morphing shapes. In particular, we attempt the problem of positioning an eye-in-hand robotic system with respect to objects in the workspace for grasping and manipulation. In our formulation, the grasp position (and consequently the approach trajectory of the manipulator), varies with each object. The proposed solution to the problem consists of two parts. First, based on a model-based object recognition framework, images of the objects taken at the desired grasp pose are stored in a database. The recognition and identification of the grasp position for an unknown input object (selected from the family of recognizable objects) occurs by morphing its contour to the templates in the database and using the virtual energy spent during the morph as a dissimilarity measure. In the second step, the images synthesized during the morph are used to guide the eye-in-hand system and execute the grasp. The proposed method requires minimal calibration of the system. Furthermore, it conjoins techniques from shape recognition, computer graphics, and vision-based robot control in a unified engineering amework. Potential applications range from recognition and positioning with respect to partially-occluded or deformable objects to planning robotic grasping based on human demonstration.     

空间众包环境下的3类对象在线任务分配

随着移动互联网技术与O2O（offline-to-online）商业模式的发展,各类空间众包平台变得日益流行,如滴滴出行、百度外卖等空间众包平台更与人们日常生活密不可分.在空间众包研究中,任务分配问题更是其核心问题之一,该问题旨在研究如何将实时出现的空间众包任务分配给适宜的众包工人.但大部分现有研究所基于的假设过强,存在两类不足:（1）现有工作通常假设基于静态场景,即全部众包任务和众包工人的时空信息在任务分配前已完整获知.但众包任务与众包工人在实际应用中动态出现,且需实时地对其进行任务分配,因此现存研究结果在实际应用中缺乏可行性;（2）现有研究均假设仅有两类众包参与对象,即众包任务与众包工人,而忽略了第三方众包工作地点对任务分配的影响.综上所述,为弥补上述不足,本文提出了一类新型动态任务分配问题,即空间众包环境下的三类对象在线任务分配.该问题不但囊括了任务分配中的三类研究对象,即众包任务、众包工人和众包工作地点,而且关注动态环境.本文进而设计了随机阈值算法,并给出了该算法在最差情况下的竞争比分析.特别的是,本文还采用在线学习方法进一步优化了随机阈值算法,提出自适应随机阈值算法,并证明该优化策略可逼近随机阈值算法使用不同阈值所能达到的最佳效果.最终,本文通过在真实数据集和具有不同分布人造数据集上进行的大量实验验证了算法的效果与性能.