QQ客服 官方微博 反馈留言

题名	机器人非确定性轨迹规划方法研究
作者	祁若龙
学位类别	博士
答辩日期	2017-05-27
授予单位	中国科学院沈阳自动化研究所
授予地点	沈阳
导师	周维佳
关键词	机器人 非确定性 轨迹规划 运动控制与仿真平台
其他题名	Research on robot trajectory planning method under uncertainties
学位专业	模式识别与智能系统
中文摘要	轨迹规划是新一代机器人运动决策的必要手段。存在非确定性信息的条件下，为了能够使机器人得到最优的运动轨迹，需要具有相应智能轨迹规划算法的支撑。本文以航天预研课题“空间科学手套箱系统研制”、国家科技重大专项（2010ZX04007-11）“大型薄壁高精度搅拌摩擦焊设备技术研究”、中国科学院青年促进会“机器人非确定性轨迹规划方法研究”、中国科学院沈阳自动化研究所“蒋新松基金”为依托。针对现有非确定性信息影响下轨迹规划理论和实践应用的不足，将机器人信息的非确定性分为三个方面：机器人加工对象几何信息非确定性、机器人运动动态特性的非确定性和机器人运动与传感误差的非确定性，并分别提出了相应的规划算法，最终得到非确定性信息影响下机器人的最优运动轨迹，提升机器人运动的精确性、安全性和稳定性。本文着重开展了如下几方面的研究： （1）在机器人加工对象几何信息非确定性轨迹规划方面：提出一种离线高精度估计轨迹规划方法，针对高精度机器人加工，讨论机器人轨迹规划过程中与被加工对象之间和被加工对象本身存在几何非确定性误差时，通过测量-加工一体化手段实现机器人高精度加工。提出了基于接触式稀疏点测量的高阶连续分段五次样条拟合方法、法向矢量估计方法，并以机器人化搅拌摩擦焊接为例，实现了针对非确定性几何信息的机器人测量加工一体化轨迹规划方法，完成了高精度加工。 （2）在机器人运动特性非确定性轨迹规划方面：提出一种运动特性多目标轨迹规划方法，针对已知机械臂目的位置情况下，机器人中间运动过程动态特性的非确定性，提出了一种在关节空间内的高次样条轨迹描述模型。提取了关键轨迹参数，通过遗传算法进行求解，在优化系统动力学性能的同时，最终得到了一条速度和加速度连续、关节扭矩不超过机器人关节扭矩极限、关节和末端运动行程较短、运动时间较短，并且能够使整个机械臂成功避开障碍的一条理想轨迹，完成了动态特性非确定性条件下机器人的多目标轨迹规划方法。 （3）在机器人运动与传感非确定性轨迹规划方面：提出了一种机器人受运动和传感误差影响下的轨迹规划和轨迹成功概率评估方法。这种轨迹规划方法将机器人系统运动与传感随机误差的非确定性通过建模、概率论及其几何化方法相结合，得到机器人在笛卡尔空间下的期望和方差，定性地判定机器人是否能和周围环境发生干涉碰撞、对机器人系统各位置在整个运动过程的误差概率分布进行估计并计算机器人到达指定位置区域的成功概率。在轨迹规划阶段考虑机器人操作的成功概率。 最后，通过面向对象的模块化编程方法搭建了机器人运动控制与图形仿真系统，其开放式体系结构能够兼容多自由度串联结构机器人的控制和仿真。依托于该机器人控制与图形仿真平台，对本文所提出的轨迹规划算法开展了系统仿真和实验验证，证明了本文所提出的轨迹规划方法的有效性和实用性。
英文摘要	Trajectory planning is a necessary mean for robot motion decision. Under the condition of uncertain information, corresponding trajectory planning algorithms should be used in order to get the optimal motion trajectory. This research was sponsored by the spaceflight advanced research project “Development of Space Scientific Glove Box System”, the National Science and Technology Major Project “Research on Large-scale and High-precision Friction Stir Welding Equipment for Thin-shell Structures” (2010ZX04007-11), Youth Innovation Promotion Association CAS “Research on Robot Trajectory Planning with Uncertainties” and “Jiang Xinsong Foundation” in SIA. In view of the deficiency of the existing method, the uncertainties of robot information is devided into three aspects: geometry information uncertainty of the object that a robot machines, dynamic property uncertainty in motion process, motion and sensor error uncertainty of the robot and their corresponding planning algorithms are proposed respectly. At last, the optimal motion trajectories under uncertain informations are got, and in the same time the precision, safty, and stability of the robot are promoted. The research process and detailed descriptions are as follows: (1) Trajectory planning method based on the geometry information uncertainty of the robot machining object is researched. Taking the FSW robot machining for large-scale complex surface thin-walled part that has high accuracy demanding as object of study, the high accurate robot manufacturing method realized by trajectory planning for the manufacturing object and the coordinate relationship between the robot and the manufacturing object have uncertain geometry error is discussed. Segmented quintic spine fitting and its normal vector estimation method based on sparse points contact measurement is proposed. The trajectory planning method based on uncertain geometry information is realized. (2) High order spline trajectory description in joint space is proposed to solve the dynamic uncertainty in the movement process of the robot when the aim point is known. The key trajectory parameters are extracted and solved by Gene Algorithm which optimize the dynamic performance of the system. At last, an ideal trajectory with continuous velocity and acceleration, small joint and end movement, short movement time and feasible joint torque within limitation is got. (3) A trajectory planning method that includes successful rate estimation based on robot motion and sensor uncertainties is proposed. This trajectory planning method gets deterministic result with the expectation and variance in Cartesian space by uncertainty modeling, probability theory and geometry method. Thus, the probability distribution of the robot position error can be estimated and the successful rate that the robot arriving at the specified location can be calculated. When the variance is expressed by geometry, it is possible to judge whether the robot collides with the surroundings and consider the successful rate of the robot operation in the process of trajectory planning stage. At last, a robot motion control and simulation platform is established based on modular programming method. Simulations and experiments for different robots are all based on the research of the robot motion control and simulation platform. System simulation and experiment are carried out to verify the algorithms proposed in this research and the results proves that the algorithms are efficient and practical.
语种	中文
产权排序	1
内容类型	学位论文
源URL	[http://ir.sia.cn/handle/173321/20531]
专题	沈阳自动化研究所_空间自动化技术研究室
推荐引用方式 GB/T 7714	祁若龙. 机器人非确定性轨迹规划方法研究[D]. 沈阳. 中国科学院沈阳自动化研究所. 2017.

个性服务

查看访问统计

相关权益政策

暂无数据

收藏/分享

除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。