Title :
Location prediction for large scale urban vehicular mobility
Author :
Siyu Chen ; Yong Li ; Wenyu Ren ; Depeng Jin ; Pan Hui
Author_Institution :
Dept. of Electron. Eng., Tsinghua Univ., Beijing, China
Abstract :
Knowledge of where vehicles will be in near future helps users in daily planning, traffic monitors in vehicles scheduling, advertisers in fixed point advertising, and especially helps in communication network source provisioning. In this paper, we analyze the predictability of taxi mobility based on their locations and time period records and we present a prediction method of taxis for their next locations in 15 seconds using Markov predictor. The historical location trace of each taxi is used to train the transition probability matrix of next location for our predictor, and we use 3 different scenarios to predict. Based on records from over 2,000 taxis in Shanghai, and over 14,000 taxis in Beijing, we are able to predict the next vehicular location with an accuracy of 82%.
Keywords :
Markov processes; mobility management (mobile radio); probability; road traffic; telecommunication network planning; Beijing; Markov predictor; Shanghai; communication network source provisioning; large scale urban vehicular mobility; taxi mobility; traffic monitoring; transition probability matrix; vehicle scheduling; vehicular location prediction; Accuracy; Cities and towns; History; Markov processes; Measurement; Roads; Vehicles; Markov chain; mobility prediction; vehicular network;
Conference_Titel :
Wireless Communications and Mobile Computing Conference (IWCMC), 2013 9th International
Conference_Location :
Sardinia
Print_ISBN :
978-1-4673-2479-3
DOI :
10.1109/IWCMC.2013.6583818