基于路径预测的不确定性推理组合地图匹配算法.pdfVIP

39 5 Vol. 39, No. 5 2010 10 Acta Geodaetica et Ca rtographica Sinica Oct. , 2010 :2010) 唐进君, 刘 芳 , 010018 A Driver Route Prediction Ba ed Mapmatching Algorithm Integrating Uncertain Rea oning TANG Jinj un, LIU Fang Depa rtment of Energy and Traffic Engineering, Inner Mongolia Agricultura l Univer ity, Hohhot 010018, China Ab tract : A novel integrated algorithm combining route prediction with uncerta in rea oning wa propo ed. The cloud model, a powerful tool to performuncertain converting between numerical quantita tive analy i to conceptual qua lita tive a na ly i , wa fir tly u ed to calcula te the credibility of po itioning point to candidate roa d by uncerta in reference with current locating data . Then, a hidden Markov model ( HMM) wa built to predict the driver route and de tinat ion . Through rede igning ob ervation function in HMM, an integrated ma p matching algorithm wa e tabli hed by combining route predicting informa tion with current matching approach. Additiona lly, the learning algorithm wa carried out to upport the algorithm and update the information. Finally, the experimental re ult demon tra te the fea ibility and effectivene of t he propo ed a lgorit hm: it ca n predict a driver route to improve the accura cy of ma p ma tching algorithm a ccording to the utility of prematching. Key word :mapmatching a lgorithm; vehicle naviga tion; cloud model; hidden Markov model; driver route prediction :提出一种新的基于预测的不确定性推理组合地图匹配算法该方法 先利用云模型这样一种处理定性与定量 间不确定转换的工具,通过对当前信息的不确定性推理,计算定位点相对于候选道路的匹配可信度之后, 利用隐马尔 科夫模型预测驾驶员出行路径及目的地,通过对模型中的观察值函数重新设定, 将路径预测信息与当前地图匹配算法结 合, 构建组合地图匹配算法同时, 采用学习算法更新模型中的信息最后, 通过仿真实例证明该组合算法的可行性与 有效性结果表明:该方法能够预测驾驶员行驶路径, 起到预匹配的作用, 提高地图匹配算法的精度 :地图匹配算法; 车辆导航;云模型; 隐马尔科夫模型;驾驶员路径预测 : P208 : A , 1 引 言 ,