【作者】 高杨；

【导师】 林景栋；

【作者基本信息】 重庆大学 ， 工程（控制工程）（专业学位）， 2022， 硕士

【摘要】 工业机器人在当今智能制造领域中扮演着至关重要的角色,机器人抓取更是制造业里极为普遍的环节,工业机器人对运动目标的抓取方法主要有两种:一种是延时补偿抓取,通过视觉传感器单次获取目标位置,并结合拍照与抓取时间差进行距离补偿来实现抓取任务,易受输送线振动和目标姿态变化的影响,抓取失败率高;另一种是跟踪抓取,通过视觉传感器实时获取目标位置,并引导机器人进行跟踪和抓取,能有效克服输送线振动和目标姿态变化的问题,抓取成功率高。因此,本文以机器人跟踪抓取展开研究工作。本文对工业机器人的目标跟踪抓取技术进行了分析,确定了机器人跟踪抓取的整体方法。在机器人部分,根据项目技术要求和成本要求,自主设计了一款工业4轴机器人,将JT3的驱动电机和减速器安装在JT1的正上方,减小了驱动扭矩,并基于MATLAB平台进行了仿真。在跟踪阶段,针对跟踪算法实时性和目标区域丢失的问题,基于Open CV进行了运动目标的跟踪算法研究对比,选取实时性较高的KCF算法并进行优化,结合输送线编码器数据进行目标区域重构,改善了目标区域丢失问题。在获取运动目标形心时,针对图像处理计算量过大的问题,通过跟踪算法输出ROI,基于ROI区域进行图像预处理、图像滤波、阈值分割和运动目标的形心计算,减少了图像处理的计算量。在机器人抓取时,针对抓取延迟的问题,对机器人轨迹规划及抓取策略进行了研究,通过3-4-5次多项式规划了典型的4轴机器人抓取运动路径,再通过最小二乘法预测运动目标形心在抓取时刻的位置,达到了准确抓取的效果。论文中搭建了包含自主设计工业4轴机器人、工业相机、其他外设、上位机、下位机在内的试验平台,并完成了系列测试,测试结果显示工业机器人可以通过本文优化后的算法对运动目标进行跟踪和精确抓取,设计的跟踪算法实时性比较好,形心获取算法计算量减小且有效,机器人抓取策略能够解决抓取延迟问题,说明了本文算法的有效性。更多还原

【Abstract】 Industrial robots play a vital role in today’s intelligent manufacturing field,and robot grasping is a very common link in the manufacturing industry.There are two main methods for industrial robots to grasp moving targets: one is delay compensation grasping,which obtains the target position through the visual sensor at a single time and realizes the grasping task by combining the distance compensation of the time difference between photographing and grasping,and it is easily affected by the vibration of the conveying line and the change of the target attitude,and the grasping failure rate is high;The other is tracking and grasping,which obtains the target position in real time through the visual sensor and guides the robot to track and grasp,which can effectively overcome the problems of conveyor line vibration and target attitude change,and the success rate of grasping is high.Therefore,this paper focuses on robot tracking and grasping.In this paper,the target tracking and grasping technology of industrial robot is analyzed,and the overall method of robot tracking and grasping is determined.In the robot part,according to the technical requirements and cost requirements of the project,an industrial 4-axis robot is independently designed,the driving motor and reducer of JT3 are installed directly above JT1 to reduce the driving torque,which the simulation is carried out based on MATLAB platform.In the tracking stage,aiming at the real-time performance of the tracking algorithm and the loss of the target area,the tracking algorithm of the moving target is studied and compared based on Open CV.The KCF algorithm with high real-time performance is selected and optimized,and the target area is reconstructed combined with the data of the conveyor line encoder to improve the problem of the loss of the target area.When acquiring the centroid of moving target,aiming at the problem that the amount of calculation in image processing is too large,ROI is output through tracking algorithm.Image preprocessing,image filtering,threshold segmentation and centroid calculation of moving target are carried out based on ROI region,which reduces the amount of calculation in image processing.In the process of robot grasping,aiming at the problem of grasping delay,the robot trajectory planning and grasping strategy are studied.The typical grasping motion path of 4-axis robot is planned through 3-4-5 degree polynomial,and then the position of the moving target centroid at the grasping time is predicted by the least square method,so as to achieve the effect of accurate grasping.In this paper,a test platform including self-designed industrial 4-axis robot,industrial camera,other peripherals,upper computer and lower computer is built,and a series of tests are completed.The test results show that the industrial robot can track and accurately grasp the moving target through the optimized algorithm in this paper.The designed tracking algorithm has good real-time performance,and the calculation of centroid acquisition algorithm is reduced and effected that the robot grasping strategy can solve the problem of grasping delay.It shows the effectiveness of the paper.更多还原