为提高城市交叉口交通状态综合评价结果的准确性和稳定性，选取交通量和时间占有率作为评价指标，将城市交叉口交通状态划分成畅通、轻微拥挤、拥挤、严重拥挤4 种类别。分别利用熵权法、变异系数法、层次分析法及组合赋权法确定交通状态评价指标的权重，然后，以K-均值聚类法确定的评价指标的聚类中心为基础，采用线性内插法确定评价指标的隶属度函数，对城市交叉口的交通状态进行综合评价。最后，通过实例分析对模型性能进行验证。研究结果表明：基于组合赋权法确定权重的交通状态评价与基于单一方法确定权重的交通状态评价相比，精确性和稳定性更高，在实验数据中表现出更好的适应性；并且基于组合赋权法1 （熵权法和层次分析法） 的城市交叉口交通状态评价结果精确性、稳定性及适应性最高。

In order to improve the accuracy and stability of comprehensive evaluation results about traffic state of urban intersections, traffic volume and time occupancy rate were selected as evaluation indexes to divide the traffic state of urban intersections into four kinds, including smooth, slight congestion, congestion and serious congestion. The entropy weight method, variation coefficient method, analytic hierarchy process and combination weighting approach were respectively used to determine the weight of traffic state evaluation indexes. Based on clustering center of evaluation indexes determined by K-means clustering method, via linear interpolation method, the membership function of evaluation indexes was determined to make a fuzzy comprehensive evaluation on traffic state of urban intersections. Finally, the model performance was verified through case analysis. Results show that the traffic state evaluation which weight is determined by combination weighting approach can evaluate traffic state more accurately and show better adaptability to the experiment data, compared with that by a single method. Moreover, the evaluation results of urban intersection traffic state through combination weighting method 1 (entropy weight method and analytic hierarchy process) have the highest accuracy, stability and adaptability.