基于视知觉的交通可变标志信息认度评价

交通可变信息标志(Variable Message Signs,VMS)作为交通信息动态发布设施,被广泛应用于城市道路及高速公路管理中。由于VMS发布的图符信息数量及大小等缺乏标准,无法衡量发布信息是否有效传递给了驾驶人。为了准确度量VMS发布信息的视认有效性,基于驾驶人视知觉理论和VMS识别过程的特征,建立VMS信息量的量化方法,提出VMS认度概念。根据影响静态VMS有效视认的信息量、驾驶人动视力、车速等因素,以认度率为评价指标,建立信息量-认度评价模型;对影响动态VMS有效视认的屏幕信息滚动周期、视认冗余距离等因素,以周期有效度为评价指标,构建周期-认度评价模型。然后,通过视觉信息获取试验,将1s时间内人眼可获取信息量与实际信息量的比值作为可获取信息系数:当信息量在50比特内,测试对象可获取信息系数最大,可获取信息系数趋于常数1;当信息量增加至56~105比特时,测试对象可获取信息系数呈指数骤减,随后趋于稳定(人眼可获取信息系数ρ=0.22±0.02)。最后,通过对实际道路中驾驶人对VMS信息获取情况的调查,采用认度评价模型与驾驶人实测结果进行对比分析,当静态VMS信息量大于90比特时,信息量轻微过载,驾驶人无法准确获取标志信息;当动态VMS标志周期有效度小于0.9时,驾驶人无法完整视认标志。该认度评价方法经过路测试验验证,可为VMS设计、信息发布管理提供参考。

Evaluation of Traffic Variable Message Sign Information Recognition Based on Visual Perception

As a dynamic traffic information release system, variable message signs (VMS) are widely used in the management of urban roads and highways. Because of a lack of standards on the number and sizes of VMS, determining whether the information is effectively transmitted to drivers is impossible. To accurately measure VMS information legibility, based on the perceptual theory and the characteristics of driver visual perception, we built a model to quantify the amount of VMS information and introduced the concept of recognition. Considering the amount of information that influences the valid recognition of static VMS as well as a driver's kinetic eyesight and the speed effect, we established an information-recognition evaluation model by using recognition rates as criteria. Considering the cycle of information and redundant distance as factors influencing a valid recognition of dynamic VMS, we used cycle validity as a judgment criterion to establish the cycle availability of the recognition evaluation model. Then, through a visual information acquisition experiment, the ratio of information available to the actual information in one second could be obtained as the information coefficient. Results show that as the amount of VMS information is 50 bits, the information coefficient is nearly constant at 1, and participants can acquire the most information. When the amount of information is increased to 56-105 bits, the information coefficient of participants is exponentially reduced and then tends to be stable. Finally, with a VMS on actual roads used as an example, combined with the use of a recognition evaluation model and an actual experiment to evaluate VMS, the results of a check test are as follows. When 90 bits is exceeded, the amount of static VMS is slightly overloaded. When the dynamic VMS cycle availability is less than 0.9, drivers cannot receive VMS information completely. The results show that the model can provide a reference for VMS design and information release management.