基于自适应遗传粒子群优化模糊神经网络的疲劳驾驶预测模型

摘要

To improve the prediction accuracy of fatigue driving, a prediction model for fatigue driving based on fuzzy neural network optimized by subtractive clustering and genetic algorithm-particle swarm optimization (GA-PSO) is proposed. Subtractive clustering is used to determine the network structure and its initial parameters based on training samples. By means of evolution speed factor, the network parameters are optimized with adaptive GA-PSO. The proposed model is trained and tested by using the data obtained by real vehicle fatigue driving simulation experiment and the results are compared with that of traditional particle swarm optimization, genetic and back propagation three algorithms. The results of comparison demonstrate that the model proposed not only streamline the network structure and shorten the training time, but also reduce the global error and improve prediction accuracy.%为提高疲劳驾驶的预测精度,提出了基于减法聚类和遗传粒子群优化模糊神经网络的疲劳驾驶预测模型.根据训练样本,利用减法聚类确定网络结构和初始参数；借助于进化速度因子,采用自适应遗传粒子群算法优化网络参数.利用疲劳驾驶实车模拟实验获得的数据,对该模型进行了训练和测试,并将结果与传统的粒子群、遗传和反向传播算法进行对比.结果表明,该模型不仅精简了网络结构,缩短了训练时间,而且减小了全局误差,提高了预测精度.