改进主成分和K-均值聚类的行驶工况

为构建行驶工况,消除K-均值算法对初始聚类中心的敏感性及噪声点的干扰,提出一种改进主成分分析和基于密度的改进K-均值聚类组合方法.结合距离优化法和密度法,构建一种数据集密度度量方法.选取距离较大、密度较高的数据点作为初始聚类中心与候选集,优化聚类结果的同时剔除了孤立点,采用较大贡献因子的特征值进行工况合成,最后对行驶工况油耗进行分析.结果表明,所提方法构建行驶工况的速度-加速度联合分布差异值为1.17％,特征参数平均相对误差较小.可见,合成的行驶工况能够很好地反映某地实际交通道路特征,拟合度较高.

Research on Driving Conditions Based on Improved Principal Component and K-means Clustering

In order to construct driving conditions and eliminate the interference of K-means algorithm on initial clustering center sensitivity and noise points, this paper proposes an improved principal component analysis and an improved K-means clustering method based on density. Combining distance optimization method and density method, a data set density measurement method is con-structed. In this paper, data points with a larger distance and a higher density are selected as the initial clustering center and candidate set, the outliers are eliminated while optimizing the clustering results, and the eigenvalues of larger contribution factors are used to synthesize operating conditions, and finally the driving conditions Fuel consumption analysis. The results show that the difference value of the joint speed-acceleration distribution of the driving conditions constructed by the proposed method is 1.17%, and the average relative error of the characteristic parameters is small. It can be seen that the synthesized driving conditions can well reflect the characteristics of the actual traffic road in a certain place, and the degree of fit is high.