基于改进模糊聚类与ANFIS的高速公路事件检测

摘要

为了准确并及时地发现高速公路上的交通事故隐患，减少事故引发的交通延迟，提高高速公路运行安全性，结合减法聚类与模糊C均值（FCM）聚类算法对输入样本数据进行聚类，建成初始模糊推理系统，然后通过神经网络的自学习机制，训练模糊系统参数，确定模糊推理规则，建立最终模糊模型。通过仿真实验结果对比，验证了基于改进模糊聚类与自适应神经模糊推理系统（ANFIS）建模方法的有效性。%In order to accurately and timely detect highway traffic accident, reduce traffic delay and improve highway safety, this paper combines subtractive clustering and Fuzzy C-Means(FCM) clustering method to cluster the input sample data to build the initial fuzzy inference system, then the hybrid algorithm is used to train the parameters of the fuzzy system, determine the fuzzy reasoning rules, and establish a final training fuzzy model. Compared with the simulation experimental results, the method obtains excellent performance on ROC(Receiver Operation Characteristic)curve, shows the validity of the modeling method based on the improved fuzzy clustering and Adaptive Neural Fuzzy Inference System(ANFIS).