摘要:路网匹配是基于位置服务中的关键预处理步骤，它将GPS轨迹点匹配到实际路网上.以此为基础对数据进行分析和挖掘，能够辅助解决城市计算中相关问题，例如建立智能交通系统、协助用户规划出行.对国内外****在该研究领域取得的成果进行了分类总结，发现这些匹配算法可以较好地解决高采样率的路网匹配问题.但是，随着城市交通的快速发展，获取和处理车辆位置信息的成本不断提高，低频采样点越来越多，现有算法匹配精确度大幅度下降.于是，近年来出现了基于隐马尔可夫模型（hidden Markov model，简称HMM）的路网匹配算法.隐马尔可夫模型可以较为平滑地将噪声数据和路径约束进行整合，从有许多可能状态的路径中选择一条最大似然路径.重点总结了基于隐马尔可夫模型的路网匹配算法，主要是从特点与实验结果的角度对其进行对比总结，有些实验结果的正确率在一定条件下最高可达90%，这说明了基于隐马尔可夫模型的路网匹配算法在低采样率下的有效性.最后，对未来的研究可能采取的方法进行了展望.



Abstract:Map matching is a key preprocessing step in the location-based service to match GPS points into a digital road network. Data analysis on the map matched trajectory data can be used to facilitate many real city computing applications such as intelligent transportation system and trip planning. This survey provides a systematic summary of existing research achievements of map matching. With the rapid development of urban traffic, the cost of acquiring and processing vehicle location information is increasing, low-sampling-rate GPS tracking data is growing, and the accuracy of existing algorithms is not adequate. In recent years, map matching algorithm based on hidden Markov model (HMM) has been widely studies. HMM can smoothly assimilate noisy data with path constraints by choosing a maximum likelihood path. The accuracy of HMM-based algorithms can reach 90% under certain conditions, which confirms the validity of map matching algorithm based on HMM at low sampling rate. A perspective of future work in this research area is also discussed.



PDF全文下载地址:

http://jos.org.cn/jos/article/pdf/5424