چكيده به لاتين
Suggesting routs with the shortest travel time causes to prevent traveling on unnecessary routs and create long traffic along the way. To predict travel times and present it to Vehicles, there are many ways. Vehicular ad-hoc networks (VANET) considered as an efficient approach for traffic management applications. In the present study, to predict travel time of the paths, was used Vehicular ad-hoc networks, but in parts of the routes, that VANETs can not cover, to predict the travel time was used artificial intelligence methods, such as multi-layer Perceptron (MLP) and support vector regression models (SVR). Along the route, if the vehicle travels within range of the road side unit covering, could request the shortest path to the destination and to be answered to it with A* algorithm. this thesis investigates travel time prediction accuracy with two methods that employed and to implementation, the method with a lower prediction error is used. The results of the comparison of proposed method and the map that has been covered with road side units, have been investigated and The results shows that the proposed method can predict travel time with good accuracy and can lead vehicles on the shortest path.
To simulate vehicular ad-hoc networks, VEINS softwareis used that imports in OMNET ++ environment. and to simulate vehicular traffic, the SUMO software is used. to simulation The artificial intelligence algorithms, MATLAB software is used.
