摘要
在车辆导航中 ,地图匹配算法通过筛选正确道路来计算和显示车辆行驶的正确位置并校正车载定位系统的误差。基于神经网络的自适应确定性地图匹配算法便是其中的一种 ,但该方法运算量大 ,不能够适应实时性要求。对确定性地图匹配算法作了一系列的改进 :在各节点处选择适当固定的参数替代利用神经网络对参数进行自适应的调整 ,简化了车辆是否进入节点区域的判断条件 ,取消了在跟踪模式下的确定性值计算 ,从而使算法在不降低计算准确性的基础上大大减少了运算复杂度 ,提高了时间效率 。
In car navigation system,map-matching methods are required to calculate and show the exact position of car in the road and to correct the position errors of the car navigation system by electing the right road.Adaptive-fuzzy-network-based C-measure algorithm is one of them,however,its cost of computation is so high that it is difficult to meet the demands of real time systems.In this paper,the algorithm is improved in the following ways: using the fixed parameters to substitute training parameters by fuzzy-network at each node;simplifying the conditions on whether the car steers in the node area;abolishing computation of the certainty-measure value in tracking mode.So the improved C-measure algorithm greatly reduces the computation complexity and enhances the time efficiency and it meets the demands of real time system much more.
出处
《计算机应用研究》
CSCD
北大核心
2004年第6期117-119,共3页
Application Research of Computers
关键词
地图匹配
自适应确定性算法
节点
定位模式
跟踪模式
道路路况
Map-matching
Adapative-C-measure Algorithm
Node
Tracking Mode
Position Fixing Mode
Road States
