Autonomous online generation of a motor representation of the workspace for intelligent whole-body reaching

We describe a learning strategy that allows a humanoid robot to autonomously build a representation of its workspace: we call this representation Reachable Space Map. Interestingly, the robot can use this map to: (i) estimate the Reachability of a visually detected object (i.e. judge whether the object can be reached for, and how well, according to some performance metric) and (ii) modify its body posture or its position with respect to the object to achieve better reaching. The robot learns this map incrementally during the execution of goal-directed reaching movements; reaching control employs kinematic models that are updated online as well. Our solution is innovative with respect to previous works in three aspects: the robot workspace is described using a gaze-centered motor representation, the map is built incrementally during the execution of goal-directed actions, learning is autonomous and online. We implement our strategy on the 48-DOFs humanoid robot Kobian and we show how the Reachable Space Map can support intelligent reaching behavior with the whole-body (i.e. head, eyes, arm, waist, legs).     

Multi-Modal Interaction of Human and Home Robot in the Context of Room Map Generation

In robotics, the idea of human and robot interaction is receiving a lot of attention lately. In this paper, we describe a multi-modal system for generating a map of the environment through interaction of a human and home robot. This system enables people to teach a newcomer robot different attributes of objects and places in the room through speech commands and hand gestures. The robot learns about size, position, and topological relations between objects, and produces a map of the room based on knowledge learned through communication with the human. The developed system consists of several sections including: natural language processing, posture recognition, object localization and map generation. This system combines multiple sources of information and model matching to detect and track a human hand so that the user can point toward an object of interest and guide the robot to either go near it or to locate that object's position in the room. The positions of objects in the room are located by monocular camera vision and depth from focus method.     

Velocity anisotropy of an industrial robot

Industrial robots are part of production systems and it is important to place them into the system according to their properties and behaviour. The information, obtained from the technical sheets of robots, about workspace (its dimensions and shape) is insufficient for designing the production system. The information about mobility is missing. To represent the behaviour of the robot in the workspace, velocity anisotropy of the robot is introduced and defined as the length of the shortest velocity ellipsoid axes, which can be constructed for any position of robot in its tool centre point. The position of a tool centre point is equivalent to the point in the workspace. A graphical representation of the 3D workspace with included velocity anisotropy is then performed and an example for a design of a robotised welding production system is given. In this example the benefits of the graphical representation of the workspace with included velocity anisotropy are presented and discussed.     

A 3D shape segmentation approach for robot grasping by parts

Neuro-psychological findings have shown that human perception of objects is based on part decomposition. Most objects are made of multiple parts which are likely to be the entities actually involved in grasp affordances. Therefore, automatic object recognition and robot grasping should take advantage from 3D shape segmentation. This paper presents an approach toward planning robot grasps across similar objects by part correspondence. The novelty of the method lies in the topological decomposition of objects that enables high-level semantic grasp planning.In particular, given a 3D model of an object, the representation is initially segmented by computing its Reeb graph. Then, automatic object recognition and part annotation are performed by applying a shape retrieval algorithm. After the recognition phase, queries are accepted for planning grasps on individual parts of the object. Finally, a robot grasp planner is invoked for finding stable grasps on the selected part of the object. Grasps are evaluated according to a widely used quality measure. Experiments performed in a simulated environment on a reasonably large dataset show the potential of topological segmentation to highlight candidate parts suitable for grasping.     

Workspace mapping based on multisensor information fusion using heterogeneous onboard sensors

In recent years, multiple robot systems that perform team operations have been developed. These robot systems are expected to execute complicated tasks smoothly in a given congested workspace. In this article, we propose a workspace mapping algorithm using ultrasonic stereo sonar and an image sensor in order to operate the mobile robots among obstacles. This workspace mapping algorithm involves two steps: (1) the position detection of obstacles using ultrasonic stereo sonar, and (2) the shape detection of obstacles using an image sensor. While each robot moves around in the given workspace, the two steps of the mapping algorithm are repeated and sensor data are collected. The robot measures the distance and the direction of obstacles using ultrasonic stereo sonar. The shape of obstacles is also captured using an onboard image sensor. A workspace map is created based on the sensor data accumulated from the proposed method, and successful results are also obtained through experiments.     

On-Line Robotic Interception Planning Using a Rendezvous-Guidance Technique

A novel method for online, robotic interception of moving objects using visual feedback is proposed in this paper. No prior knowledge of the motion of the object is assumed. Since such objects might depart quickly from the workspace of the robot, fast interception is a critical issue. Thus, a novel time-optimal rendezvous-guidance technique that takes the dynamic limitations of the robot into account has been developed. In the proposed methodology, first, a parallel-navigation rule, originally introduced in the missile-guidance literature, is applied to generate a set of instantaneous task-space velocity commands, which, if executed, would keep the end-effector on a collision course with the object. Subsequently, a rendezvous-guidance method is utilized to reduce the original command set to one with velocity-matching capability. Finally, the fastest velocity command in the reduced set is chosen such that the dynamic limitations of the actuators of the robot are not violated. The proposed algorithm results in a fast and robust interception as shown by several simulation examples in 2D and 3D workspace.     

利用线画图象二维串的图象理解方法

本文提出一种基于符号运算和面向规则线画图象的自动图象理解方法，方法用代数符号表达空间物体，用两个符号串描述物体在特定的投影空间上的相互关系，上方法所构造的代数系统规定了包括微分，腐蚀等等代数操作定义了若干条运算规则。利用这些操作和规则，可以实现对图象中物体的空间位置，运动趋势，自身大小的变化等等的自动分析。     

Position estimation of a mobile robot using images of a moving target in intelligent space with distributed sensors

Latest advances in hardware technology and state-of-the-art of mobile robots and artificial intelligence research can be employed to develop autonomous and distributed monitoring systems. A mobile service robot requires the perception of its present position to co-exist with humans and support humans effectively in populated environments. To realize this, a robot needs to keep track of relevant changes in the environment. This paper proposes localization of a mobile robot using images recognized by distributed intelligent networked devices in intelligent space (ISpace) in order to achieve these goals. This scheme combines data from the observed position, using dead-reckoning sensors, and the estimated position, using images of moving objects, such as a walking human captured by a camera system, to determine the location of a mobile robot. The moving object is assumed to be a point-object and projected onto an image plane to form a geometrical constraint equation that provides position data of the object based on the kinematics of the ISpace. Using the a priori known path of a moving object and a perspective camera model, the geometric constraint equations that represent the relation between image frame coordinates for a moving object and the estimated robot's position are derived. The proposed method utilizes the error between the observed and estimated image coordinates to localize the mobile robot, and the Kalman filtering scheme is used for the estimation of the mobile robot location. The proposed approach is applied for a mobile robot in ISpace to show the reduction of uncertainty in determining the location of a mobile robot, and its performance is verified by computer simulation and experiment.     

Probabilistic models for robot-based object segmentation

This paper introduces a novel probabilistic method for robot based object segmentation. The method integrates knowledge of the robot’s motion to determine the shape and location of objects. This allows a robot with no prior knowledge of its workspace to isolate objects against their surroundings by moving them and observing their visual feedback. The main contribution of the paper is to improve upon current methods by allowing object segmentation in changing environments and moving backgrounds. The approach allows optimal values for the algorithm parameters to be estimated. Empirical studies against alternatives demonstrate clear improvements in both planar and three dimensional motion.     

3D Collision-Free Motion Based on Collision Index

A collision-free motion planning method for mobile robots moving in 3-dimensional workspace is proposed in this article. To simplify the mathematical representation and reduce the computation complexity for collision detection, objects in the workspace are modeled as ellipsoids. By means of applying a series of coordinate and scaling transformations between the robot and the obstacles in the workspace, intersection check is reduced to test whether the point representing the robot falls outside or inside the transformed ellipsoids representing the obstacles. Therefore, the requirement of the computation time for collision detection is reduced drastically in comparison with the computational geometry method, which computes a distance function of the robot segments and the obstacles. As a measurement of the possible occurrence of collision, the collision index, which is defined by projecting conceptually an ellipsoid onto a 3-dimensional Gaussian distribution contour, plays a significant role in planning the collision-free path. The method based on reinforcement learning search using the defined collision index for collision-free motion is proposed. A simulation example is given in this article to demonstrate the efficiency of the proposed method. The result shows that the mobile robot can pass through the blocking obstacles and reach the desired final position successfully after several trials.     

A multiple object geometric deformable model for image segmentation

Deformable models are widely used for image segmentation, most commonly to find single objects within an image. Although several methods have been proposed to segment multiple objects using deformable models, substantial limitations in their utility remain. This paper presents a multiple object segmentation method using a novel and efficient object representation for both two and three dimensions. The new framework guarantees object relationships and topology, prevents overlaps and gaps, enables boundary-specific speeds, and has a computationally efficient evolution scheme that is largely independent of the number of objects. Maintaining object relationships and straightforward use of object-specific and boundary-specific smoothing and advection forces enables the segmentation of objects with multiple compartments, a critical capability in the parcellation of organs in medical imaging. Comparing the new framework with previous approaches shows its superior performance and scalability.     

一种基于GOR+GPU算法的机器人视觉导航方法

提出一种一般物体识别(GOR)方法.借鉴词袋(BOW)的统计模型,利用SIFT(尺度不变特征变换)检测算子进行特征向量描述.为了增加信息的冗余度,利用物体部件空间关系的统计信息来描述一幅图片中所有特征点的空间(相对距离和角度)关系,增广了原BOW模型中的特征向量.运用无监督判别分类器支持向量机(SVM)来实现分类识别.与此同时,采用GPU加速技术来实现SIFT特征提取与描述,以保证其实时性.然后,存手绘地图辅助导航的基础上,将该方法成功地应用到室内移动机器人导航上.实验结果表明,基于该方法的机器人导航技术具有较强的鲁棒性和有效性.     

基于颜色和形状特征的彩色图像显示与检索技术

提出了一种有效用于抽取特征、索引和检索彩色图像的技术途径,通过提取图像的颜色不变量,建立相应的色度直方图（hue histogram）来表示图像的颜色分布特征,为了描述图像中对象的位置及方向特征,首先计算图像的色度轮廓并对其进行Radon变换,然后计算相应的“空间直方图”,由此,得到了一种基于图像的颜色分布特征和形状特征的新的图像表示方法,为了计量图像的全局相似度,基于“累积距离”和“基于向量的距离”定义了两种图像的距离度量,并分别讨论了距离度量的选取与归一化、不同子特征的组合等CBIR所涉及的关键问题,实验结果表明,文中提出的方法能获得满意的检索性能,其检索结果能较好地接近于人的视觉感和结果。     

Feature-based tactile object recognition

Tactile sensing offers powerful capabilities for robotic perception. Through the use of array-force sensors, precisely located surface information about objects in the workspace is available wherever the robot arm may reach. In order to use this information to identify objects and their placement, interpretation processes should employ proprioceptive information and should use tactile image features which reflect object characteristics. A technique is described for the generation of constraints on object identity and placement such that information from multiple sensor contacts may cooperate towards interpretation.     

Qualitative representation of spatial knowledge in two-dimensional space

Various relation-based systems, concerned with the qualitative representation and processing of spatial knowledge, have been developed in numerous application domains. In this article, we identify the common concepts underlying qualitative spatial knowledge representation, we compare the representational properties of the different systems, and we outline the computational tasks involved in relation-based spatial information processing. We also describesymbolic spatial indexes, relation-based structures that combine several ideas in spatial knowledge representation. A symbolic spatial index is an array that preserves only a set of spatial relations among distinct objects in an image, called the modeling space; the index array discards information, such as shape and size of objects, and irrelevant spatial relations. The construction of a symbolic spatial index from an input image can be thought of as a transformation that keeps only a set of representative points needed to define the relations of the modeling space. By keeping the relative arrangements of the representative points in symbolic spatial indexes and discarding all other points, we maintain enough information to answer queries regarding the spatial relations of the modeling space without the need to access the initial image or an object database. Symbolic spatial indexes can be used to solve problems involving route planning, composition of spatial relations, and update operations.     

遥感影像中模糊对象表达

遥感影像中的空间对象及其边界具有模糊性。针对遥感影像中模糊对象的表现特征,借助云理论和云模型来对模糊对象进行表达研究。基于影像的灰度和梯度等数字特征,构建基于数字图像的对象云,用对象云表达模糊对象,核外半云环表达其边界,从而用模糊数学的方法来合理表达遥感影像模糊对象,最后以遥感影像为例进行验证。该方法不仅丰富和完善了云理论,同时也为模糊对象表达,遥感影像的理解和分析提供了一种新的思路。     

A method of real-time recognition of moving objects and its application

A simple method of real-time recognition of moving objects, as well as its theoretical basis are described. This method utilizes a TV camera to pick up image signals from an object moving in the horizontal scan direction. The shape and position of the object are detected at every vertical scanning period by using the information from a few window-like image portions each consisting of four sub-areas. This method is succesfully applied as an electronic eye to a bolting robot used for concrete pile/pole manufacture.