冗余机械臂逆运动学分析与迷彩喷涂路径规划

Inverse Kinematics Analysis of Redundant Manipulator and Path Planning for Camouflage Painting

中文关键词: 冗余机械臂  三维数码迷彩  遗传算法  神经网络  路径规划

英文关键词: redundant manipulator  3D digital camouflage  genetic algorithm  neural network  path planning

中文摘要:

      文中基于改进智能算法开展冗余机械臂逆运动学分析与三维数码迷彩喷涂路径规划。首先，基于冗余机械臂的初始和极限角度获得测试数据集，同时在网络模型中添加嵌套循环，通过新的遗传算法–逆向传播（Genetic Algorithm - Back Propagation, GA-BP）神经网络模型得到各关节角的最大绝对值误差，完成逆运动学分析；其次，在上述逆运动学求解的基础上，开展迷彩喷涂路径规划，建立迷彩图案模型，并进行Harris角点检测，实现迷彩图案局部分块，利用四邻域搜索完成分块图案的局部路径规划；然后，搭建旅行商模型，改进遗传算法的种群初始化和变异方式，进而获得二维空间的最优路径；最后，结合快速搜索随机树（Rapidly-exploring Random Trees, RRT）算法，生成三维空间数码迷彩的喷涂路径。仿真和数码迷彩喷涂实验结果验证了该方法的可行性和有效性。

英文摘要:

      Inverse kinematics analysis of redundant manipulator and path planning of 3D digital camouflage painting are carried out based on improved intelligent algorithm in this paper. Firstly, the test data set is obtained based on the initial and limit angles of the redundant manipulator. And the genetic algorithm-back propagation (GA-BP) neural network model is improved to obtain the maximum error absolute of the angles of each joint and complete the inverse kinematics analysis by adding nested loops to the network model. Secondly, the path planning of camouflage painting is carried out on the basis of the above inverse kinematics solution. The camouflage pattern model and Harris corner detection are carried out to realize the local segmentation of the camouflage pattern. The local path planning of the segmentation pattern is completed by the neighborhood search. Thirdly, the genetic algorithm is applied to obtain the optimal path of two-dimensional camouflage spraying robot by building a traveling salesman model and improving the population initialization and mutation mode of the genetic algorithm. Finally, the planned path of digital camouflage painting in the 3D space is generated combined with the rapidly-exploring random trees (RRT）algorithm. The feasibility and effectiveness of the proposed method is verified by simulation and digital camouflage spraying experiments.