This paper deals with part sequencing and optimal robot moves sequence in 2-machine robotic cells according to Petri net graph. We have assumed that the robotic cell is capable of producing same and different parts. W...This paper deals with part sequencing and optimal robot moves sequence in 2-machine robotic cells according to Petri net graph. We have assumed that the robotic cell is capable of producing same and different parts. We have considered a new motion cycle for robot moves sequence which is the development of existing motion cycles in 2-machine robotic cells. The main goal of this study is to minimize the cycle time by determining the optimal part sequencing and robot moves sequence in the robotic cell. So, we have proposed a model based on Petri network.展开更多
【目的】针对机器人自主装配研究多聚焦静态目标,移动目标装配相关研究不足的问题,提出一种视觉阻抗控制器的变参数控制方法,以提升机器人在动态环境中目标跟踪与移动装配的适应性和柔顺性。【方法】首先,结合力反馈与视觉反馈信息,在...【目的】针对机器人自主装配研究多聚焦静态目标,移动目标装配相关研究不足的问题,提出一种视觉阻抗控制器的变参数控制方法,以提升机器人在动态环境中目标跟踪与移动装配的适应性和柔顺性。【方法】首先,结合力反馈与视觉反馈信息,在特征空间构建2阶阻抗控制器,为动态装配提供控制基础;其次,将蒙特卡洛随机失活(Monte Carlo Dropout,MCD)作为概率动力学模型,改进概率推理学习控制优化(Probabilistic Inference for Learning Control Optimization,PILCO)算法,平衡状态不确定性推理能力与计算效率;最后,以特征误差和机器人关节位置信息为观测空间,自适应调整控制器阻抗参数,优化动态跟踪与装配性能。【结果】仿真及试验台验证结果表明,相比原高斯过程模型,MCD模型在保留状态不确定性推理能力的同时,显著缩短训练时间(50轮训练时间从43.50 h降至12.82 h);装配成功率从94.5%提升至98.0%,平均装配用时从5.820 s缩短至3.253 s,超调量大幅减少,跟踪响应更及时。研究可为移动目标柔顺装配提供参考。展开更多
In conection with the complex working-surroundings of the wall-climbing Robot, this paper researched akind of alternatively moving mechanism with good obstacle-surmounting ability and high moving speed, making use oft...In conection with the complex working-surroundings of the wall-climbing Robot, this paper researched akind of alternatively moving mechanism with good obstacle-surmounting ability and high moving speed, making use ofthe thought of bionics. This paper designed a kind of self-adjusting multi-vacuum sucker. Furthermore, it employedthe theory of vacuum system to establish the work mathematics madel of control switch to are sucking disc and presented the design parameter of the control switch. In addition, this paper made use of the thought of bionics to design aobstacle-surmounting mechanism used in wall-climbing robot. Also it employed the theory Of robotics to analyze the kinematics and the dynamics movement of die robot.展开更多
栅格环境下A*算法规划出的移动机器人路径存在折线多、转折次数多、累计转折角度大等问题.为获得较优路径,提出平滑A*算法.在A*算法规划的路径基础上,遍历路径中的所有节点,当某一节点前后节点连线上无障碍物时,将延长线路的这一中间节...栅格环境下A*算法规划出的移动机器人路径存在折线多、转折次数多、累计转折角度大等问题.为获得较优路径,提出平滑A*算法.在A*算法规划的路径基础上,遍历路径中的所有节点,当某一节点前后节点连线上无障碍物时,将延长线路的这一中间节点删除,建立平滑A*模型.仿真结果表明,平滑A*算法优于Ant(蚁群),Anyti me D*算法.平滑A*算法路径长度降低约5%,累计转折次数降低约50%,累计转折角度减少30%～60%.平滑A*算法能处理不同栅格规模下、障碍物随机分布的复杂环境下移动机器人路径规划问题.展开更多
文摘This paper deals with part sequencing and optimal robot moves sequence in 2-machine robotic cells according to Petri net graph. We have assumed that the robotic cell is capable of producing same and different parts. We have considered a new motion cycle for robot moves sequence which is the development of existing motion cycles in 2-machine robotic cells. The main goal of this study is to minimize the cycle time by determining the optimal part sequencing and robot moves sequence in the robotic cell. So, we have proposed a model based on Petri network.
文摘【目的】针对机器人自主装配研究多聚焦静态目标,移动目标装配相关研究不足的问题,提出一种视觉阻抗控制器的变参数控制方法,以提升机器人在动态环境中目标跟踪与移动装配的适应性和柔顺性。【方法】首先,结合力反馈与视觉反馈信息,在特征空间构建2阶阻抗控制器,为动态装配提供控制基础;其次,将蒙特卡洛随机失活(Monte Carlo Dropout,MCD)作为概率动力学模型,改进概率推理学习控制优化(Probabilistic Inference for Learning Control Optimization,PILCO)算法,平衡状态不确定性推理能力与计算效率;最后,以特征误差和机器人关节位置信息为观测空间,自适应调整控制器阻抗参数,优化动态跟踪与装配性能。【结果】仿真及试验台验证结果表明,相比原高斯过程模型,MCD模型在保留状态不确定性推理能力的同时,显著缩短训练时间(50轮训练时间从43.50 h降至12.82 h);装配成功率从94.5%提升至98.0%,平均装配用时从5.820 s缩短至3.253 s,超调量大幅减少,跟踪响应更及时。研究可为移动目标柔顺装配提供参考。
文摘In conection with the complex working-surroundings of the wall-climbing Robot, this paper researched akind of alternatively moving mechanism with good obstacle-surmounting ability and high moving speed, making use ofthe thought of bionics. This paper designed a kind of self-adjusting multi-vacuum sucker. Furthermore, it employedthe theory of vacuum system to establish the work mathematics madel of control switch to are sucking disc and presented the design parameter of the control switch. In addition, this paper made use of the thought of bionics to design aobstacle-surmounting mechanism used in wall-climbing robot. Also it employed the theory Of robotics to analyze the kinematics and the dynamics movement of die robot.
文摘栅格环境下A*算法规划出的移动机器人路径存在折线多、转折次数多、累计转折角度大等问题.为获得较优路径,提出平滑A*算法.在A*算法规划的路径基础上,遍历路径中的所有节点,当某一节点前后节点连线上无障碍物时,将延长线路的这一中间节点删除,建立平滑A*模型.仿真结果表明,平滑A*算法优于Ant(蚁群),Anyti me D*算法.平滑A*算法路径长度降低约5%,累计转折次数降低约50%,累计转折角度减少30%～60%.平滑A*算法能处理不同栅格规模下、障碍物随机分布的复杂环境下移动机器人路径规划问题.
