Abstract
One of the keys in time-dependent routing is determining the weight for each road network link based on symmetrical and complete traffic information. To facilitate travel planning considering traffic situations based on historical global position system (GPS) trajectory data which uncover the whole road network, this paper proposes a fuzzy random forest-based road section data estimation method, which uses the third law of geography as the core idea. For different time periods, road grade, tidal lane, proximity to infrastructure (main places that affect traffic, such as schools, hospitals), and accident road sections were selected as indicators that influence the traffic. The random forest algorithm is used to build the mapping relationship between attribute data with average traffic which is obtained based on GPS data. Subsequently, the fuzzy reasoning method is used to obtain the weight of road links missing traffic information by calculating their similarities with typical road section samples. Using the road network of Suzhou City as an example, the proposed method was used to analyze estimate the average driving speeds of road sections with missing traffic information for different time periods. Experimental results show that this method can effectively avoid congested road sections and obtain high-speed travel routes.
Highlights
Traffic congestion has become a major problem in cities as the number of vehicles on city roads has increased
The traffic information is collected by sensors or Taxi global position system (GPS) trajectory data, the sensors are usually installed at the main roads and there are still a lot of roads lacking traffic data records, Taxi GPS trajectory data cannot cover the whole road
Quiroga proposed a localization system to estimate traffic conditions [26], using the precise information provided by GPS, the real-time traffic information of each section of the city is obtained through mathematical modeling, and the current traffic flow or congestion situation is judged
Summary
Longhao Wang 1 , Jing Wu 1 , Rui Li 1 , Yanjiao Song 1 , Jiayue Zhou 1 , Xiaoping Rui 2, *.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
