基于BP网络的平面四杆可调机构轨迹综合  

Path Synthesis of Planar Four-Bar Adjustable Mechanisms Based on BP Neural Network

作　　者：雷经发[1] 扈静[2] 徐礼钜[3] 陈从升[1] 

机构地区：[1]安徽建筑工业学院机电系,合肥230601 [2]合肥工业大学机汽学院,合肥230009 [3]四川大学制造科学与工程学院,成都610065

出　　处：《机械工程师》2007年第10期67-69,共3页Mechanical Engineer

基　　金：国家自然科学基金资助项目(50375104);安徽建筑工业学院青年基金资助项目(04-038)

摘　　要：提出了一种基于BP网络的多轨迹平面四杆可调机构综合的方法。轨迹综合机构是指能使连杆上某点沿着某一给定轨迹运动的平面连杆机构综合。文中讨论了任务数目,未知变量和方程组数目之间的关系。该方法避免了繁琐的符号推导,减少了工作量。将理想结果和实际结果比较,表明利用该方法进行可调机构的轨迹综合具有较高的精度。A mechanism for performing path synthesis is designed to guide one poinl on a moving rigid body such that it passes through a specified sequence of points on a path. The relationship between the number of the tasks, unknown sariables and equations are discussed. A method for performing path synthesis of four-bar adjustable mechanisms for two alternate tasks is proposed based on BP neural network. A network is trained and all solutions are obtained. The method presented in this thesis can avoid tedious symbol derivation and reduee greatly computing work as compared with analytical approach. The errors between desired points and actual points are very small, and the path synthesis mechanism has higher accuracy.

关 键 词：轨迹综合 平面四杆可调机构 BP神经网络 

分 类 号：TH112[机械工程—机械设计及理论]