基于形状谱的平面四杆机构神经网络轨迹设计法  

Neural Network Synthesis of Four-bar Path Generating Linkages Based on Shape Spectrum

作　　者：翁桂荣[1] 钱志良[1] 

机构地区：[1]苏州大学,苏州215021

出　　处：《中国机械工程》2005年第3期215-218,共4页China Mechanical Engineering

基　　金：江苏省高校自然科学基金资助项目(Q2117032)

摘　　要：　　提出用图像形态学与神经网络相结合的方法进行平面四杆轨迹机构的优化设计。利用数学形态学结合图像处理,提取平面连杆曲线的形状谱特征参数,并构建曲线形态的识别、比较方法;再在曲线识别比较的基础上,提出了平面四杆轨迹机构的神经网络综合法,并用实例验证了方法的有效性。A synthesis method for four-bar path-generating linkages was presented by combining shape spectrum with neural network. Firstly, shape spectrum characteristics of coupler curves were extracted by picture analysis method based on mathematical morphology, and the method of distinguishing curves was developed. Secondly, on the basis of coupler curve shape spectrum, neural network method of synthesizing four-bar path-generating linkages was proposed. Some synthesis examples validate the effectiveness of the method.

关 键 词：机构综合 平面连杆机构 形状谱 神经网络 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]