基于RBF网络的虚拟仪表人机界面评价方法

提出了基于RBF网络的虚拟仪表人机界面主观评价方法和评价指标。利用RBF网络的自组织、自学习与自适应特性对网络进行训练,使网络学习隐含在训练数据中的人机界面主观评价指标的权重规律,自适应调整主观评价指标的权重,克服了主观赋权法的随机性因素影响。建立了虚拟光柱表人机界面,开发了基于RBF网络的虚拟光柱表人机界面主观评价模型;对训练样本数为50,75和100的三组虚拟仪表网络模型进行了误差分析。分析结果表明,采用75个训练样本可以得到满意的主观评价精度。

Evaluation Method of Human-Machine Interface of Virtual Meter Based on RBF Network

A novel subjective evaluation approach for evaluating the human-machine interface of virtual meters based on RBF neural network as well as the evaluation indexes were proposed. By using the self-organizing, self-learning and self-adapting properties of RBF neural networks, the regularity of subjective evaluation indexes weight concealed in the training data could be learned by means of RBF neural networks automatically adjusting indexes weight of subjective evaluation, and therefore the influence of randomicity could be overcome. In order to validate the proposed method, a human-machine interface of virtual bargraph meters was developed and the subjective evaluation model of virtual bargraph meter was established. Error analysis of the subjective evaluation model for three groups of virtual bargraph meters was performed by using 50, 75, and 100 training samples, respectively. Analysis results show that the subjective evaluation model by using 75 training samples is of satisfied accuracy.