四自由度串联嫁接机器人架构约束模块识别可视化方法

Visualization method for architectural constraint module indentification of four-degree-of-freedom serial grafting robot

为实现育苗嫁接的智能化与高效化,一类四自由度串联嫁接机器人的新结构被提出;针对其机械结构、时序控制、联动运行等复杂性,发展该类机器人架构约束下的模块识别可视化方法及设计流程。首先,构建考虑特定架构约束下机器人的设计结构矩阵(DSM),并采用大型社区快速展开算法(LOUVAIN算法)实现对机器人架构模块的预划分;其次,采用改进遗传算法(AGA)对DSM进行聚类,获得多组高模块度的划分结果,由此形成成组可能性矩阵(GLM),并将其对角化得到对角GLM(DGLM);最后,提出针对DGLM的可视化策略并完成着色处理,实现对最优模块划分方案的辨识。结果显示:四自由度串联嫁接机器人可分解为较独立的7大模块,模块度可达0.782。

In order to achieve intelligent and efficient of seedling grafting process,a new mechanical structure scheme regarding the four-degree-of-freedom serial grafting robot is proposed.In view of the complexity of its structure,timing control and linkage operation,the module identification visualization method and design implementation with the architectural constraints is developed.Firstly,the design structure matrix(DSM)of the robot considering the specific architecture constraints is constructed,and a large community fast unfolding algorithm(LOUVAIN Algorithm)is used to realize the pre-division of the robot architecture blocks.Secondly,the DSM is clustered by using an advanced genetic algorithms(AGA),and multiple sets of high modularity divisions are obtained,which form a group possibility matrix(GLM).The diagonal GLM(DGLM)is obtained by diagonalizing GLM.Finally,the visualization strategy for DGLM is developed,and the colored optimal module division result is obtained.The optimal design result shows that the four-degree-of-freedom serial grafting robot should be divided into seven independent modules,and the modularity reaches 0.782.