欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
特定交通环境下的驾驶员行为的变化是一个非线性的复杂系统，传统的驾驶决策模型和车辆行驶模型难以体现驾驶员的感知、判断、决策、动作等一系列心理、生理活动的不确定性和不一致性，而人工神经网络特别适合于因果间不易建立明确联系的问题。提出了基于神经网络集成的驾驶员行为学习算法DNNIA，该算法的有效性通过标准数据集和仿真实验得到了验证，学习到的诸如驾驶员踩踏踏板的习惯行为的仿真结果与采集的样本数据总体趋势较为一致，且实现了系统泛化性能的提高。
英文摘要：
The change of driving behavior under specific traffic environments is a nonlinear complex systems, and the traditional driving decision-making model and vehicle driving model are difficult to reflect the driver＇ s such a series of psychological, physiological activity of uncertainty and inconsistency as perception, judgment, decisions, actions. However, the artifi- cial neural network is particularly adapted to indefinite causal link between the problems. Based on neural network ensemble learning algorithm of the driving behavior DNNIA （dynamic neural network integrated algorithm）, the effectiveness of the algorithm through the standard data sets and simulation experiments had been verified. The simulation results of the driver learned habit behavior such as pedal acts and the general trend of the sample data are basically consistent and the generalization performance has been improved.
同期刊论文项目
同项目期刊论文
An improved approximation approach incorporating particle swarm optimization and a priori informatio
A new constrained learning algorithm for function approximation by encoding a priori information int
Modified constrained learning algorithms incorporating additional functional constraints into neural
期刊信息
位置：