Framework of experienced route planning based on taxis´ GPS data

Route planning is a core part of the traffic information system. Conventional route planning models rely heavily on an improved weighted shortest path algorithm, while this paper proposes a route planning methodology based on taxis´ experienced travels. Taking advantage of the data from floating cars, the framework explores drivers´ experience in route choices, extracts taxis´ operating behaviors to establish the database of experienced routes, and repairs incomplete and abnormal routes to update the database. Given the starting and ending segments, alternative experienced route choice set can be obtained by traversing the database. Taking the road network and taxis´ GPS data of Guangzhou city for instance, we make comparisons of the shortest paths and experienced ones, and focus on their differences in factors like length, turning frequency, frequency of signalized intersections, travel time, etc. Various numerical results are documented to show that experienced routes appear better considering all the factors than shortest-distance or shortest-time routes.