| dc.description.abstract | Traffic jams reduce transportation efficiency and increase transportation cost. Traffic jams also cause air pollution and rise global warming. Electric vehicle is the future trend in power saving and CO2 reducing; moreover, electric vehicles embedded with navigation system is an economical solution for reducing traffic jams. Navigation system integrates the global positioning system and electric map to guide vehicles to reach right positions. Currently, navigation systems incorporate traffic information to avoid congested roads. Navigation systems benefit from the predicable traffic and its accuracy. In general, an electric vehicle will drive into a predictability congested road or has battery depletion. Conventional navigation systems are unable to respond to the sudden conditions, because they did not take predicted traffic and battery power into account. Therefore, this dissertation proposes a green navigation system based on autonomic traffic predication in vehicular ad-hoc networks. In this architecture, road-side units and vehicles play a role of sensor or monitor, and they autonomically exchange date and aggregate information for traffic prediction. Navigation systems plan recommendation paths according to the predicted traffic and the estimated state-of-charging to improve the traffic efficiency and to avoid battery depletion. First of all, this dissertation studies the influence of weather factors including temperature, humidity and rainfall on traffic prediction, and improves the accuracy of prediction according to the speed of upstream road segments. Then, the proposed green navigation system is incorporating the predicted traffic information and the estimated state-of-charging. In real world, the internal combustion engine vehicles have long continues driving mileage because of full-oil can reach destination, but the electric vehicles have a restriction of short continues driving mileage because of battery capacity need go to a charging station before battery depletion. Therefore, this dissertation discusses when the internal combustion engine vehicles and the electric vehicles adopt the proposed green navigation system, individually. Real traffic measurements and weather data are used for the evaluation of the proposed prediction scheme. Civic Boulevard in Taipei City is selected as the prediction target. The prediction results show that the proposed traffic prediction improves accuracy by 57.4% when compared with a hybrid approach. The simulation of green navigation system has two types of vehicles: internal combustion engine vehicles and electric vehicles. The navigation results show that the proposed green navigation system improves average speed by 15.49% for internal combustion engine vehicles when compared with the distributed approach. The navigation results also show that the proposed green navigation system improves mileage by 9.52% for electric vehicles when compared with the distributed approach. However, in the peak petroleum price, this dissertation proposes a green navigation system to reduce the transportation cost that includes electric power or oil consumptions. | en_US |