摘要
建立了基于BP神经网络的路径规划模型,首先介绍了非接触测量路径规划中的常用方法,包括多项式法和三次样条插值法等及其优缺点,然后构造了1×30×1的BP神经网络模型,说明了模型的原理,阐明了模型具体参数的设置和训练步骤的设计。最后用Matlab实现了神经网络,并用实例对网络模型进行了仿真,同时和三次样条插值的仿真结果对比,仿真结果表明,此网络模型很好的提高了路径拟合和预测的准确性和效率。
This paper established the BP neural network model about path planning, first, introduced common methods about path planning,including polynomial method and cubic spline interpola＇s tion method and their merits and flaws, then constructed a 1×30×1 BP neural network model, including the theory of this model and explaining the settings of specific parameters and the design of the training steps about this model. Finally, the model was achieved by Matlah and was simulated by an instance,the results _were compared with the results simulated by cubic spline interpolation. The results demonstrate that this model increases the measurement accuracy and measurement efficiency.
出处
《机械与电子》
2012年第9期35-39,共5页
Machinery & Electronics
关键词
非接触测量
神经网络
路径规划
三次样条插值
non - contact measurement
neural network
path planning
cubic spline interpolation
