摘要
为研究黄昏时段不同颜色引起驾驶员空间距离判识差异的变化规律,选用32名驾驶员,在黄昏环境照度变化的实际道路中,对不同深度距离下红、绿色障碍物的空间绝对距离和相对距离进行判识。比较人在三维空间中对红、绿色障碍物距离判识结果,获得判识距离差异特征值。对所获得的试验结果,运用BP神经网络模拟距离判识差异随照度和深度距离不同而变化的规律。结果表明:在三维空间,红、绿色障碍物黄昏时段距离判识差异显著,绿色障碍物判识距离大于红色,BP网络可以很好地拟合距离判识差异变化规律;随着外界环境照度的下降,红、绿颜色引起的距离判识值差异均增大,其中绝对距离差异增加明显;随着空间深度方向距离增加,判识差异也增大,其中相对距离判识差异增加量较绝对距离小。
The present paper intends to report the results of its au- thor' s experimental study with the variation regularity of the drivers' distance distinction between red and green obstacles in a three dimen- sional space from our case study with 32 drivers participating in. The 32 driver-subjects were chosen randomly with the purpose of improv- ing the driving safety and finding better ways for improving their recognition capability of the distance information in the dusk. The ac- tual experiment was designed in the following way: in the front of a car, two obstacles and marks of perceptible positions were laid out. The driver-subjects are asked to identify the absolute distance in a static manner, that is, the distance between the subject and the first obstacle under the different environmental illumination conditions in depth distances. At the same time, the subjects are asked to tell the relative distance between the two obstacles. Similar programs are sup- posed to be carried out to distinguish the red from the green obsta- cles. And then, the results of distance recognition values between the red and green obstacles are compared so as to obtain difference values under the different illumination and depth distance conditions. The relation between the recognition distance and the environment illumi- nation can be simulated by BP neural network. The results of experi- ments show that there are significant differences between the red and green obstacles and the cognitive value from the green proves larger than the red. BP neural network can also well simulate the variation trends of difference and the augment with illumination. In addition, it is also obvious that the difference naturally increases with the depth distance increase and the augment from relative shorter.
出处
《安全与环境学报》
CAS
CSCD
北大核心
2010年第5期172-176,共5页
Journal of Safety and Environment
基金
国家自然科学基金项目(50778023)
关键词
交通运输安全工程
黄昏
驾驶员
环境照度
距离判识差异
BP神经网络
traffic and transportation safety engineering
dusk
drivers
environment illumination
distance cognitive difference
BP neural network