论文部分内容阅读
为研究昼夜动态环境中驾驶员对空间距离判识的规律,进行了实际道路试验。随机选取32名驾驶员分别在昼、夜环境中不同深度距离和速度下,判识红、绿色障碍物的空间距离,统计并检验驾驶员对红、绿色障碍物判识距离的差异,获得判识特征值;运用BP神经网络拟合距离判识结果,分析距离判识变化规律。结果表明:BP神经网络可以很好地拟合距离判识变化规律,精度优于现有模型;绝对距离和相对距离判识结果均随速度增加而减小,随深度距离增加而增大;夜间判识距离大于白天,驾驶员对相对距离判识准确性高。
In order to study the law of driver’s perception of spatial distance in day-night dynamic environment, the actual road test was conducted. 32 drivers were randomly chosen to identify the spatial distance between red and green obstacles at different depth distances and velocities in daytime and nighttime environment respectively. The differences between drivers’ recognition of red and green obstacles were statistically calculated and verified. Identify the eigenvalues; use BP neural network to fit the result of distance identification and analyze the change rule of distance identification. The results show that BP neural network can well fit the change rule of distance identification, and the accuracy is better than the existing model. The results of absolute distance and relative distance decrease with the increase of velocity and increase with the increase of depth; Judgment distance greater than during the day, the relative accuracy of the driver to identify high accuracy.