犁体曲面装配孔精确定位的神经网络方法  

Research on 3-Dimension Positioning of Assemble Holes on Free Surface Based on Neural Network

作　　者：王建[1] 王京春[2] 高峰[1] 吴成武[3] 

机构地区：[1]北京航空航天大学汽车工程系 [2]吉林大学生物与农业工程学院 [3]吉林大学机械科学与工程学院

出　　处：《农业机械学报》2004年第4期44-46,共3页Transactions of the Chinese Society for Agricultural Machinery

摘　　要：犁体曲面装配孔的位置精度要求很高 ,稍有偏差就会影响耕作效果。采用神经网络方法对二维平面节点与三维空间节点之间的关系进行描述 ,建立了二者之间的网络映射矩阵。将平面点的二维坐标代入该网络 ,即可计算出精确的三维坐标。以犁体曲面的装配孔为例 ,利用自适应学习率的 BP网络进行了网络训练和仿真计算 ,结果表明 ,这种方法可以获得很高的定位精度 ,并已应用于 1L F4 35犁体曲面的设计中。The assemble hole on the plow surface need high precision. Any bias would affect the performance. The relationship between the coordinates on the 2-Dimension plane and the coordinates in the 3-Dimension space has been established by using of the Neuron Network in this paper. The mapping matrix of the 2-Dimension and 3-Dimension has been created. Providing the coordinates of a point on the 2-Dimension known, the coordinates of the point in the 3-Dimension can be computed accurately by the matrix. Take an assemble hole on the plow surface as example, the matrix has been trained and simulated by using of the BP ANN with self-adaptive learning ratio. The result indicated that high accurate position could be achieved. This ANN method has been used in the designing of the 1LF435 plow surface.

关 键 词：犁体曲面 装配孔 精确定位 神经网络法 网络映射矩阵 

分 类 号：S222.1[农业科学—农业机械化工程]