Developing of Self-learning Car Navigation System

1. 利用GPS天线（ETAK ET-GPS1）、地图和GPS软件（MapFan II）、个人电脑（NEC PC-9821Nr166）组成了GPS系统。提出了利用GPS行驶轨迹数据检测路网上的行驶路段并计算路段行驶次数的方法。形成了将行车时间数据作为知识库保存到计算机磁盘上的系统。构建了利用旅行时间数据库搜索出发地至目的地最优路线并在地图上显示的系统。5 .通过在冈山市区主要干线道路网络上行驶安装GPS系统的试验车辆，实际形成了行车时间数据库。通过计算机仿真，验证了导航系统在旅行时间数据库的基础上给出的最优路线的准确性。虽然结果因交通状况而异，但该导航系统的最优路线搜索准确率总体平均约为42%，最高约为75%。该导航系统平均可节省9%的出行时间，最高可节省20%的出行时间。通过研究表明，自主学习汽车导航系统是有效的，在技术上是可行的。

1. The GPS system that used the GPS antenna (ETAK ET-GPS1), map & GPS software (MapFan II), personal computer (NEC PC-9821Nr166) was composed.2. The method how to detect the traveling link on the road network and calculate link travel times from the GPS traveling locus data was formed.3. The system that preserves the travel time data as a knowledge database to the disk of the computer was formed.4. The system that searches the optimal route between origin and destination by using the travel time database and displays it on the map was formed.5. By driving the test vehicle that installed the GPS system on the major line road network of Okayama city area, the travel time database was formed actually.6. The accuracy of the optimal route, that is presented by Navigation System on the basis of the travel time database was examined by using computer simulation.7. Although the result differs by traffic condition, the accuracy of the optimal route search by this Navigation System is about 42% as the whole average and about 75% as the maximum.8. The travel time cut by this Navigation System is about 9% as the whole average and about 20% as the maximum.9. Through the study it was made clear that the self-learning car navigation system is effective and it is possible technically.