摘要
为满足机构设计过程中目标变异的要求,提出面向目标变异的操作臂运动学优化设计方法.在D-H坐标系中,通过递推齐次变换获得操作臂末端执行器相对于基础坐标系的位姿,基于拉格朗日插值多项式的数值微分表达式,建立非数值集合的映射规则,在已有设计目标的基础上获得变异目标的显式表达式.将多维连续设计变量离散为网格,再将模拟退火机制引入蚁群算法,用在给定温度下服从波尔茨曼(Boltzmann)分布的随机热摄动表示蚁群在位置间的转换概率,求解多目标变异表达式,基于模糊集从帕雷托(Pareto)前沿选优,最终使末端执行器在特定轨迹上进一步满足运动学特性.开发了自主原型系统,实现了单环开链6R操作臂的运动学优化设计.结果表明,该方法有助于快速实现机构变异设计目标的表达、解析及选优,对于其他复杂机械产品的设计也具有参考意义.
To meet the requirements of object variation during design process for mechanism,an object variation oriented kinematics optimization design method for manipulator was presented.In Denavit-Hartenberg coordinate system,the position and orientation of the end effector of manipulator compared to the frame coordinate system was obtained using recurrence homogeneous transformations.The non-numerical set mapping rules were built to achieve explicit expressions of variation objects on the basis of existing design objects through numeric differential expression based on Lagrange interpolation polynomial.The multi-dimensional continuous design variables were discretized into discrete grids.The essence of simulated annealing algorithm was employed to the ant colony optimization.The random thermal perturbations at a given temperature conforming to a Boltzmann distribution was used to express the probability that the ant visiting between the nodes.The multi-expressions of variation objects were then resolved.The solutions were selected from Pareto optimal front based on fuzzy set.The end effector of manipulator was finally optimized to meet the kinematics characteristics in the specified trajectory.The independent prototype system has been developed and the design of single loop open chain 6R manipulator was realized.The test results proved that the proposed method helps to realize variation objects representation,resolution and solution-selection rapidly.It also has reference significance to the design of other complex products.
出处
《浙江大学学报（工学版）》
EI
CAS
CSCD
北大核心
2011年第2期209-216,共8页
Journal of Zhejiang University：Engineering Science
基金
国家科技数控机床重大专项资助项目(2009ZX04014-031)
国家青年科学基金资助项目(51005204)
中国博士后基金资助项目(20100471000)
教育部博士点基金资助项目(200803350031)
浙江省自然科学基金资助项目(Y1080358)
关键词
目标变异
拉格朗日插值多项式
退火蚁群算法
操作臂
运动学优化设计
object variation
Lagrange interpolation polynomial
simulated annealing ant colony algorithm
manipulator
kinematics optimization design
