摘要
为减少汽车碰撞事故的发生,提高智能车辆主动避撞能力,通过模拟驾驶平台对驾驶员的避撞行为特性进行了分析。采用皮尔逊相关系数法选取具有良好表征能力的特征参数,基于遗传算法优化的双隐含层BP神经网络建立一种符合驾驶员避撞行为特性的智能车辆换道避撞模型,该模型可以对制动减速度和方向盘转角进行较为准确的预测。利用未经训练样本数据,在Matlab平台对模型的可行性和有效性进行验证。结果表明:该模型的预测精度较高,能与驾驶员实际避撞行为较好匹配。
In order to reduce automobile collisions and improve the active collision avoidance ability of intelligent vehicles,the characteristics of driver′s collision avoidance behaviors were analyzed through a simulated driving platform.The Pearson correlation coefficient method is used to select characteristic parameters with good characterization ability,and the intelligent vehicle lane changing collision avoidance model is established based on a double hidden layer BP neural network optimized by genetic algorithm.The model can predict the braking deceleration speed and steering wheel angle more accurately.Using the untrained sample data,the feasibility and effectiveness of the model were verified on the Matlab platform.The results show that the model has a high accuracy prediction and matches the actual collision avoidance behavior of drivers better.
作者
王涛
关志伟
彭涛
李达
WANG Tao;GUAN Zhiwei;PENG Tao;LI Da(School of Automobile and Transportation,Tianjin University of Technology and Education,Tianjin 300222,China;School of Automobile and Rail Transportation,Tianjin Sino-German University of Applied Sciences,Tianjin 300350,China)
出处
《天津职业技术师范大学学报》
2021年第3期24-29,共6页
Journal of Tianjin University of Technology and Education
基金
天津市人工智能科技重大专项(17ZXRGGX00070)
天津市研究生科研创新项目人工智能专项(2020YJSZXS13)
天津市科技局重点研发计划——京津冀三地联合攻关项目(19YFSLQY00010)
天津市科技计划项目(20KPHDRC00030).
关键词
智能车辆
驾驶员避撞行为
遗传算法
BP神经网络
优化
intelligent vehicle
driver′s collision avoidance behavior
genetic algorithm
BP neural network
optimization
