Vision Navigation Based PID Control for Line Tracking Robot

In a controlled indoor environment,line tracking has become the most practical and reliable navigation strategy for autonomous mobile robots.A line tracking robot is a self-mobile machine that can recognize and track a painted line on thefloor.In general,the path is set and can be visible,such as a black line on a white surface with high contrasting colors.The robot’s path is marked by a distinct line or track,which the robot follows to move.Several scientific contributions from the disciplines of vision and control have been made to mobile robot vision-based navigation.Localization,automated map generation,autonomous navigation and path tracking is all becoming more frequent in vision applications.A visual navigation line tracking robot should detect the line with a camera using an image processing technique.The paper focuses on combining computer vision techniques with a proportional-integral-derivative(PID)control-ler for automatic steering and speed control.A prototype line tracking robot is used to evaluate the proposed control strategy.

Seam Tracking and Visual Control for Robotic Arc Welding Based on Structured Light Stereovision

A real-time arc welding robot visual control system based on a local network with a multi-level hierarchy is developed in this paper. It consists of an intelligence and human-machine interface level, a motion planning level, a motion control level and a servo control level. The last three levels form a local real-time open robot controller, which realizes motion planning and motion control of a robot. A camera calibration method based on the relative movement of the end-effector connected to a robot is proposed and a method for tracking weld seam based on the structured light stereovision is provided. Combining the parameters of the cameras and laser plane, three groups of position values in Cartesian space are obtained for each feature point in a stripe projected on the weld seam. The accurate three-dimensional position of the edge points in the weld seam can be calculated from the obtained parameters with an information fusion algorithm. By calculating the weld seam parameter from position and image data, the movement parameters of the robot used for tracking can be determined. A swing welding experiment of type V groove weld is successfully conducted, the results of which show that the system has high resolution seam tracking in real-time, and works stably and efficiently.

Design of Robot Welding Seam Tracking System with Structured Light Vision

Robot welding is an important developing direction of welding automation and intelligentization, and automatic seam tracking technology is one of principal research domains. Nowadays, seam tracking system with structured light vision becomes a hot research. Structured light vision seam tracking products abroad are generally very expensive and can only be applied on special occasions. In China, the research of structured light vision seam tracking system is still just on the stage of experiments. A robot real-time seam tracking system with line structured light vision is designed. The hardware system is set up, a filtering method for line structure seam image is improved, and compared with common filtering, it has better effect and characteristic of real time. Two methods, fast template matching and fast Hough transform, to recognize the image coordinates of seam center are improved. Two new image recognition methods, structure element matching and comer detecting, are proposed. The comparison of seam image recognition shows that fast template matching and comer detecting are more precise and stable than the other two methods, and comer detecting is the best in real time. A simultaneous calibration for camera parameters and robot hand-eye is also proposed, and calculation shows that the calibration is effective and feasible. The robot seam tracking tests for linear and folded lap-joint are performed, which are based on the above four image recognition methods, and the results indicate that four image recognition methods are all applicable to real-time seam tracking, and the whole system sufficed for the requirements of real-time seam tracking. Automatic seam tracking with line structured light vision is feasible and has good versatility.

3D Object Visual Tracking for the 220 kV/330 kV High-Voltage Live-Line Insulator Cleaning Robot

The 3D object visual tracking problem is studied for the robot vision system of the 220kV/330kV high-voltage live-line insulator cleaning robot. The SUSAN Edge based Scale Invariant Feature (SESIF) algorithm based 3D objects visual tracking is achieved in three stages: the first frame stage,tracking stage,and recovering stage. An SESIF based objects recognition algorithm is proposed to find initial location at both the first frame stage and recovering stage. An SESIF and Lie group based visual tracking algorithm is used to track 3D object. Experiments verify the algorithm's robustness. This algorithm will be used in the second generation of the 220kV/330kV high-voltage live-line insulator cleaning robot.

Computer Vision-based Navigation and Predefined Track Following Control of a Small Robotic Airship

为小机器的飞艇，飞艇应该能够跟随一条预定义的轨道，这被要求。在这份报纸，计算机为机器的飞艇的基于视觉的航行和最佳的模糊控制策略被建议。第一，视觉航行基于环境的自然里程碑被介绍。例如，当飞艇在一个城市上正在飞时，大楼能被用作其几何性质从数字地图或一个地理信息系统(GIS ) 被知道的视觉烽火。然后，几何方法论被采用关于飞艇的取向和位置提取信息。以便在一个预定义的磁道上保留飞艇，一个模糊飞行控制系统被设计，它把那些数据用作它的输入。并且基因算法(气体) ，一个通用全球优化方法，被利用优化模糊控制器的会员功能。最后，航行和控制策略被验证。

Hybrid Visual Servoing Control for Robotic Arc Welding Based on Structured Light Vision

A novel hybrid visual servoing control method based on structured light vision is pro-posed for robotic arc welding with a general six degrees of freedom robot. It consists of a positioncontrol inner-loop in Cartesian space and two outer-loops. One is position-based visual control inCartesian space for moving in the direction of weld seam, i.e., weld seam tracking, another is image-based visual control in image space for adjustment to eliminate the errors in the process of tracking.A new Jacobian matrix from image space of the feature point on structured light stripe to Cartesianspace is provided for dierential movement of the end-e?ector. The control system model is simplifiedand its stability is discussed. An experiment of arc welding protected by gas CO2 for verifying iswell conducted.

基于视觉跟踪实时引导的传送带跟随方法

针对滚筒输送线场景下的机器人随线质检问题,提出一种基于视觉跟踪实时引导的传送带跟随方法,即通过定位相机跟踪待测目标的运动,实时引导机器人调整动作跟随运动目标并检测。首先,通过顶部固定相机获取包含待测目标的实时视频,构建多目标跟踪(MOT)算法跟踪待测目标的位置与旋转角;然后,通过图像信息解算目标的6D位姿,进一步解算对应的机器人工作点坐标;最后,根据实时工作点坐标在线调整机器人动作,保持机器人末端与工作点重合,实现跟随检测。实验结果表明,所提方法可实现滚筒输送线场景下的机器人随线检测;相较于现有的基于目标检测的跟随方法,不存在跟随偏差累积问题。

基于计算机视觉的采摘机器人作业优化研究

为进一步提高我国采摘机器人智能化作业水平及采摘效率,基于计算机视觉应用技术展开优化研究。以采摘机器人结构组成为设计基点,运用视觉控制核心理念建立采摘机器人视觉识别处理与控制数学模型,实施相适应的采摘路径规划和系统采摘状态输出设计。同时,进行视觉采摘验证试验,结果表明:基于计算机视觉的采摘机器人系统优化正确可行,整机综合采摘效率可达94.61%,系统识别准确率与控制精度分别相对提升了6.12%和5.25%,机器人采摘成功率可提升至95.98%。因此,将计算机视觉处理技术有效应用至采摘机器人设计改进,对于类似农业采摘与收获装备开发研究可提供创新及借鉴思路,推广价值良好。

混合现实场景下结合SSVEP与眼动追踪的脑控机械臂系统

针对混合现实场景下脑控机械臂系统交互性差、指令集小的问题,设计了一种结合稳态视觉诱发电位(SSVEP)和眼动追踪技术的混合现实脑控机械臂系统。该系统通过眼动追踪技术实现目标区域的初选,而SSVEP信号则被用于在初选区域内识别最终的目标指令。在不增加刺激类别数量的前提下扩大了指令集,并根据受试者的视线停留区域实现异步控制。离线实验结果表明,在使用相同刺激类别数量的情况下,增加视觉刺激目标数量不会对分类准确率产生显著影响。通过在线实验验证了系统的适用性,相较于使用单一SSVEP范式的机械臂控制系统,所提出的系统具有更好的交互性和更大的指令集。

基于PSO-BP神经网络的分拣机器人视觉反馈跟踪

针对分拣机器人视觉反馈跟踪精度差、耗时较长的问题,研究基于粒子群算法-反向传播(particle swarm optimization-back propagation,PSO-BP)神经网络的分拣机器人视觉反馈跟踪方法,以提升视觉反馈跟踪效果。依据分拣机器人的视觉反馈信息,建立分拣机器人运动学模型,并求解分拣机器人机械臂输出位置和输入位置的误差函数;利用PSO算法优化BP神经网络的权值与偏置;在权值与偏置优化后的BP神经网络内,输入误差函数,预测分拣机器人视觉反馈跟踪控制量;利用预测视觉反馈跟踪控制量,在线调整增量式比例-积分-微分(proportional-integral-derivative,PID)的参数,输出高精度的分拣机器人视觉反馈跟踪控制量,实现分拣机器人视觉反馈跟踪。实验结果表明,该方法可有效视觉反馈跟踪分拣机器人机械臂的关节角;存在干扰情况下,在运行时间为10 s左右时,阶跃响应趋于稳定;有干扰情况下,视觉反馈跟踪的平均误差为0.09 cm,耗时平均值为0.10 ms;无干扰情况下,平均误差为0.03 cm,耗时平均值为0.04 ms。

双闭环PID应用下投篮机器人路径跟踪控制研究

投篮机器人路径图像噪声会影响机器人路径跟踪控制效果,导致投篮机器人偏离正常轨迹,为此提出双闭环PID应用下投篮机器人路径跟踪控制方法。分析投篮机器人架构,利用OpenMv视觉识别方法采集投篮机器人路径图像,结合小波阈值算法对采集到的图像实施去噪处理,从而获取清晰的投篮机器人路径图像。根据图像去噪处理结果,将曲线拟合法与特征点检测法相结合,实现投篮机器人路径识别。根据投篮机器人路径识别结果确定路径实际偏移量,结合推动器实现位置、移动速度建立的双闭环PID控制器,从而实现投篮机器人路径跟踪控制。实验结果表明,该方法的投篮机器人路径图像去噪熊效果好、跟踪控制效果好,充分验证了该方法的有效性。

面向溺水救援机器人平稳跟踪的模糊比例微分控制视觉伺服方案

针对溺水救援机器人的平稳跟踪需求,提出一种基于模糊比例微分控制的视觉伺服方案,通过场馆顶部安装的俯视摄像头获取机器人的坐标位置,使用虚拟导航线设计动态行走路径,计算偏移方向、偏移角并制定转弯规则;利用三角隶属度函数制定模糊比例、微分控制规则表,根据2种规则表中的比例、微分系数分别控制机器人的偏移角、距离偏移量及其变化率,实现机器人的跟踪与避障的平稳性调整;机器人采用核相关滤波跟踪算法对溺水人员进行跟踪,在溺水人员位置变化时调整运动方向,依靠实时、稳定的声呐信息设置偏移角,实现避障功能;对所提出的方案在先锋P3DX型机器人上进行实验验证。结果表明:机器人能及时应对目标变化,没有跟错、跟丢现象;在不同工况时的横滚角均小于1°,避障率达到100%,能够快速、稳定、准确地到达目标位置,达到溺水救援机器人的前期应用要求。

最远距离命中移动靶心的自主连续体机器人——世界纪录介绍

针对连续体机器人利用视觉感知实现对外部目标瞄准跟踪控制问题,大连理工大学彭海军教授团队设计研发了一款基于视觉伺服的自主连续体机器人.经世界记录认证(WRCA)官方现场严格测试,该机器人被认定为“最远距离命中移动靶心的自主连续体机器人”,最远有效跟踪距离为30.129米.

应用多传感器激光视觉的焊接机器人变质心补偿控制

制造、安装、机械传动和温度变化等诸多方面的误差较大,产生补偿机器人变质心偏移度,任务执行途中失去期望位姿,为了提高焊接机器人工作效率,提出应用多传感器激光视觉的焊接机器人变质心补偿控制方法。建立焊接机器人变质心模型,应用激光传感器扫描焊接机器人实际轨迹和位姿的几何特征,获取激光视觉图像标定管道焊缝,结合PLC控制器,调整焊接机器人轨迹和位姿,消除惯性力对变质心位移的干扰,实现焊接机器人变质心补偿控制。实验结果表明,在X-Y坐标面、Z-X坐标面和Z-Y坐标面上的变质心补偿轨迹均与期望轨迹误差5 cm左右,高度重合;控制后,焊接机器人惯性位移在0.1 cm之内,近似于零,惯性力不再引导焊接机器人移动,可以稳定停靠在目标位置,变质心控制性能好。

基于视觉信息的移动机器人自主导航位置跟踪方法

机器人实时位置节点与预设导航轨迹样点存在较大偏差,会导致导航系统无法准确跟踪机器人运动行为。为了解决上述问题,设计基于视觉信息的移动机器人自主导航位置跟踪方法。基于视觉信息模型提取并连接导航点,以便对导航视觉信息实施保存。针对机器人自主导航运动行为进行逆解运算,并根据同步位置跟踪器建模条件。计算跟踪误差,完成基于视觉信息的移动机器人自主导航位置跟踪方法的设计。实验结果表明,该方法可将实时位置节点与预设导航轨迹样点间的坐标偏差控制在1%之内,不会因偏离问题而使导航系统无法准确跟踪机器人运动行为。

基于多目视觉识别技术的电芯上料设备研究

液冷储能柜及风冷储能柜作为储能设备及储能控制系统的核心部件,其关键设备液冷电池模块(PACK)及风冷电池模块的生产过程涉及多种厂家各规格及各种存储、运输形式的电芯物料,电芯上料设备自动、通用问题成为了亟须解决的问题。针对液冷电池模块(PACK)及风冷电池模块生产线中电芯来料形式不统一的问题,基于视觉识别技术,将其与工业机器人集成,设计对应的电芯上料设备技术方案来解决生产线中电芯上料的准确性和快速性问题。通过多目视觉检测系统在电芯搬运的工业机器人上的应用,实现视觉识别分析夹爪状态并自调整,适应各规格及各种摆放方式的电芯,解决了工业机器人在抓取过程中存在的定位不准确的问题,极大地适应了电池电芯出厂包装中电芯分布零散,分布层次较多的实际状况,解决了现有行业生产过程中存在的设备抓取的效率低、定位准确度差、实际生产过程稳定性差等问题,大大提高了生产效率。

水下捕捞机器人视觉系统发展趋势分析

为突破作业环境、劳动力等因素对渔业发展的制约,解决当前水下捕捞能见度低、机器人识别分辨率低等问题,促进渔业向现代化转型升级,从学术角度梳理、分析国内外水下机器人视觉系统关键技术发展趋势,为相关研究提供参考。基于CiteSpace软件,分别选取中国知网(China National Knowledge Infrastucture,CNKI)的数据库和Web of Science的核心数据库的文献样本,从发文量分布、关键词共现图谱、关键词时间线图谱等方面完成可视化分析,并对水下捕捞机器人作业环境、视觉系统技术进行阐述。发文量和关键词时间线图谱表明,国内外水下机器人视觉系统关键技术的研发均经历初始阶段、探索阶段和成长阶段。关键词共现图谱分析表明,人机交互、识别算法设计等方面是水下机器人视觉系统相关研究的热点。未来,水下捕捞机器人视觉系统发展应聚焦智能化、自主化,重点解决目标检测优化、水生物数据库建立、数字化系统构建等多方面难点问题。

基于踪片Tracklet关联的视觉目标跟踪:现状与展望

近年来,由于计算机视觉技术的发展和计算机硬件性能的提高,基于视觉的目标跟踪方法得到了飞速的发展.其中,基于踪片(Tracklet)关联的目标跟踪方法因为具有对目标遮挡的强鲁棒性、算法运行的快速性等优点得到了广泛关注,本文对这类方法的最新研究进展进行了综述.首先,简明地介绍了视觉目标跟踪的基本知识、研究意义和研究现状.然后,通过感兴趣目标检测、跟踪特征提取、踪片生成、踪片关联与补全四个步骤,系统详尽地介绍了基于踪片关联的目标跟踪方法,分析了近年来提出的一些踪片关联方法的优缺点.最后,本文指出了该研究问题的发展方向,一方面要提出更先进的目标跟踪模型,另一方面要采用平行视觉方法进行虚实互动的模型学习与评估.

Adaptive tracking control for uncertain dynamic nonholonomic mobile robots based on visual servoing

The trajectory tracking control problem of dynamic nonholonomic wheeled mobile robots is considered via visual servoing feedback. A kinematic controller is firstly presented for the kinematic model, and then, an adaptive sliding mode controller is designed for the uncertain dynamic model in the presence of parametric uncertainties associated with the camera system. The proposed controller is robust not only to structured uncertainties such as mass variation but also to unstructured one such as disturbances. The asymptotic convergence of tracking errors to equilibrium point is rigorously proved by the Lyapunov method. Simulation results are provided to illustrate the performance of the control law.

Masked Autoencoders as Single Object Tracking Learners

Significant advancements have beenwitnessed in visual tracking applications leveragingViT in recent years,mainly due to the formidablemodeling capabilities of Vision Transformer(ViT).However,the strong performance of such trackers heavily relies on ViT models pretrained for long periods,limitingmore flexible model designs for tracking tasks.To address this issue,we propose an efficient unsupervised ViT pretraining method for the tracking task based on masked autoencoders,called TrackMAE.During pretraining,we employ two shared-parameter ViTs,serving as the appearance encoder and motion encoder,respectively.The appearance encoder encodes randomly masked image data,while the motion encoder encodes randomly masked pairs of video frames.Subsequently,an appearance decoder and a motion decoder separately reconstruct the original image data and video frame data at the pixel level.In this way,ViT learns to understand both the appearance of images and the motion between video frames simultaneously.Experimental results demonstrate that ViT-Base and ViT-Large models,pretrained with TrackMAE and combined with a simple tracking head,achieve state-of-the-art(SOTA)performance without additional design.Moreover,compared to the currently popular MAE pretraining methods,TrackMAE consumes only 1/5 of the training time,which will facilitate the customization of diverse models for tracking.For instance,we additionally customize a lightweight ViT-XS,which achieves SOTA efficient tracking performance.