汽车转向时驾驶员驾驶意图辨识与行为预测

提出了一种基于隐马尔可夫模型(HMM)和人工神经网络(ANN)相结合的模型,用于实现驾驶意图辨识和驾驶行为预测,从而达到辅助驾驶和提高主动安全性的目的。通过驾驶模拟器获得双移线和直线行驶的数据,通过结合HMM和ANN建立了驾驶员意图辨识和行为预测模型,对驾驶员意图和行为在紧急转向、正常转向和直线行驶三种工况下的操作分别进行了辨识和预测。实验结果表明,该模型在给定车速下能够准确辨识当前的驾驶意图和较为精确预测驾驶员的驾驶行为。

Driving intentions identification and behaviors prediction in car lane change

In order to realize the auxiliary driving and achieve the goal of improving the active safety,an integrated model based on the artificial neural network and the hidden Markov chain was proposed to identify and predict the driving intentions and behaviors.Such a model was established with the double-lane change and straight line data obtained from driving simulator.This model was used to identify and predict the driving intentions and behaviors under this three condition,such as emergency turning,normal turning,and straight line running.Experimental results show that the model can accurately identify the current driving intention and accurately predict the next driving behavior under a given speed.