面向模拟驾驶训练的驾驶意图识别方法

Recognition Method of Driving Intention for Driving Training with Driving Simulators

针对模拟驾驶训练情形,建立了一种能识别当前驾驶意图的复合隐Markov模型,对高速公路环境中的几种常见驾驶意图进行识别,并利用模拟驾驶器实验平台对识别模型进行了实验验证。结果表明:模型识别结果准确可靠,可及时将危险驾驶状态反馈给模拟训练中的驾驶人,为研究更可靠、高效的驾驶训练方法奠定了基础,对规范驾驶行为,提升驾驶人技术水平,减少现实道路交通事故的发生具有重要意义。

For driving training with driving simulators,a complex hidden Markov model has been es-tablished to recognize drivers’current driving intention in highway environment. Experimental verifi-cations have been made by using the driving simulator experiment platform,and the results show that the complex model is accurate and reliable. The dangerous driving status can be recognized and feed-back to the driver. The study has made a foundation for exploring a more reliable and efficient method of driving training. And it has realistic significance for standardizing driving behavior,improving driv-ers’technology level and reducing road traffic accidents.