基于AR-HMM的重型车辆侧翻预警模型与算法研究

项目名称： 基于AR-HMM的重型车辆侧翻预警模型与算法研究

项目编号： No.51205151

项目类型： 青年科学基金项目

立项/批准年度： 2013

项目学科： 机械工程学科

项目作者： 赵志国

作者单位： 淮阴工学院

项目金额： 24万元

中文摘要： 针对传统的侧翻预警方法在重型车辆载荷和重心变化较大时，预警可信度下降问题，在深入研究重型车辆侧翻失稳机理的基础上，全面分析重型车辆行驶稳定性的影响因素，构建基于AR-HMM的车辆侧翻预警模型，采用Baum-Welch算法进行预警模型参数优化；根据训练好的AR-HMM、车辆当前行驶状态和驾驶行为，结合自回归算法和Viterbi算法，探索有效的重型车辆侧翻预警算法，预测未来一段时间内车辆的运行状态，在危险工况下实现在线实时预警；基于多维GM-HMM，进行侧翻危险工况辨识算法研究；搭建重型车辆侧翻预警控制测试平台，验证理论分析和数值计算结果。本研究为侧翻预警和防侧翻控制等问题提供新思路和通用策略，对隐马尔可夫理论在碰撞、追尾及车道偏离预警系统中的应用具有重要借鉴意义。

中文关键词： 重型车辆；侧翻预警；隐马尔可夫模型；模型与算法；

英文摘要： When the load and center of gravity of heavy duty vehicle changes greatly, the credibility of the traditional rollover warning method decline badly .Aiming at this problem, on the basis of in depth research on the mechanism of instability of heavy vehicle, the factors influencing the stability of heavy duty vehicle comprehensively are analyzed, the model of rollover warning for heavy duty vehicle based on AR-HMM is built, and the parameters of the model by using the Baum-Welch algorithm are optimized .According to the trained AR-HMM, the current driving conditions and driving behavior of vehicle, combining with autoregressive algorithm and Viterbi algorithm, the algorithm of rollover warning for heavy duty vehicle effectively is explored, the vehicle operating status in the next period of time can be forecasted. Meanwhile, the online and real-time warning in the dangerous conditions could be achieved. In addition, the algorithm for identification of vehicle rollover in the dangerous conditions is studied, the test platform of the rollover warning control for heavy duty vehicle is built, and the theoretical analysis and numerical results is verified. This study provides new ideas and common method of the rollover warning and anti-roll control, which is of significance to the application of hidden Markov theory f

英文关键词： Heavy duty vehicle；Rollover warning；Hidden markov model；Model and algorithm；