不确定环境下移动对象自适应轨迹预测方法

发布时间：2018-12-16 19:58

【摘要】：已有的轨迹预测方法难以对移动对象运动轨迹进行准确地描述,尤其在复杂且不确定的车载自组织网络(vehicular ad hoc network)(也称车联网)环境中.为了解决这一问题,提出基于变分高斯混合模型(variational Gaussian mixture model,VGMM)的环境自适应轨迹预测方法 ESATP(environment self-adaptive prediction method based on VGMM).首先,在传统高斯混合模型的基础上使用变分贝叶斯推理近似方法处理混合高斯分布;其次设计变分贝叶斯期望最大化算法学习计算高斯混合模型参数,有效运用参数先验信息得到更高精度预测模型;最后,针对输入轨迹数据特征,使用参数自适应选择算法自动调节参数组合,灵活调整混合高斯分量的个数和轨迹段大小.实验结果表明:所提方法在实验中表现出较高的预测准确性,可应用于车辆移动定位产品中.
[Abstract]:The existing trajectory prediction methods are difficult to accurately describe the moving object trajectory, especially in the complex and uncertain vehicle-borne ad hoc network (vehicular ad hoc network) () environment. In order to solve this problem, an environmental adaptive trajectory prediction method, ESATP (environment self-adaptive prediction method based on VGMM)., based on variational Gao Si mixed model (variational Gaussian mixture model,VGMM) is proposed. Firstly, based on the traditional Gao Si mixed model, variational Bayesian inference approximation method is used to deal with the mixed Gao Si distribution. Secondly, the variational Bayesian expectation maximization algorithm is designed to calculate the parameters of Gao Si mixed model, and a more accurate prediction model is obtained by using the prior information of the parameters effectively. Finally, according to the characteristics of the input trajectory data, the parameter adaptive selection algorithm is used to adjust the combination of parameters automatically and flexibly adjust the number of the mixed Gao Si component and the size of the trajectory segment. The experimental results show that the proposed method has high prediction accuracy and can be applied to vehicle mobile positioning products.
【作者单位】：长沙理工大学综合交通运输大数据智能处理湖南省重点实验室;长沙理工大学计算机与通信工程学院;国防科学技术大学计算机学院;
【基金】：国家自然科学基金项目(61572514) 湖南省自然科学基金项目(14JJ7043) 湖南省交通厅科技进步与创新项目(201405)~~
【分类号】：O212;U495

【相似文献】