基于位置服务的轨迹预测方法

Trajectory Prediction Method for Location-based Service

现有移动对象在路网中位置预测的技术大部分是当前路段上短期预测.为了能够长期准确预测路径,提出一个路网分层模型,通过路网分层和减少路网中的路口节点来降低路网的复杂度,提高了轨迹预测算法的性能,同时也减少了数据库中的存储数据量,避免了不必要的通信开销.然后,基于路网模型提出了一个探测回溯算法.该算法按路网分层模型综合考虑路段信息选择概率最高的路段,提高了算法的精确度和效率.通过真实数据表明,这种方法比现有的预测方法更准确和高效.

Most of the existing location prediction techniques for moving objects on road network are mainly short-term prediction on the current road fragment. In order to accurately predict the long-term trajectory, this paper presents a hierarchical road network model, to lower the complexity by reducing intersection nodes on road networks. It improves the performance by reducing the amount of data storage and avoiding the unnecessary communication overhead. Based on the model, the paper proposes a detection backtracking algorithm,which comprehensively chooses the highest probability road fragment according to the hierarchical road network information. The experiments on real data show that this method is more accurate and efficient than other prediction algorithms.