This paper presents a smart driving direction system leveraging the intelligence of experienced drivers. In this system,GPS-equipped taxis are employed as mobile sensors probing the traffic rhythm of a city and taxi drivers’ intelligence in choosing drivingdirections in the physical world. We propose a time-dependent landmark graph to model the dynamic traffic pattern as well as theintelligence of experienced drivers so as to provide a user with the practically fastest route to a given destination at a given departuretime. Then, a Variance-Entropy-Based Clustering approach is devised to estimate the distribution of travel time between twolandmarks in different time slots. Based on this graph, we design a two-stage routing algorithm to compute the practically fastest andcustomized route for end users. We build our system based on a real-world trajectory data set generated by over 33,000 taxis in aperiod of three months, and evaluate the system by conducting both synthetic experiments and in-the-field evaluations. As a result, 60-70 percent of the routes suggested by our method are faster than the competing methods, and 20 percent of the routes share the sameresults. On average, 50 percent of our routes are at least 20 percent faster than the competing approaches. T-Drive Enhancing Driving Directions with Taxi Driver’s Intelligence
The efficient driving directions has become a dailyactivity and been implemented as a key feature in manymap services like Google and Bing Maps. A fast drivingroute saves not only the time of a driver but also energyconsumption (as most gas is wasted in traffic jams). Therefore,this service is important for both end users andgovernments aiming to ease traffic problems and protectenvironment.Essentially, the time that a driver traverses a routedepends on the following three aspects: 1) the physicalfeature of a route, such as distance, capacity (lanes), and thenumber of traffic lights as well as direction turns; 2) thetime-dependent traffic flow on the route; and 3) a user’sdriving behavior. Given the same route, cautious driverswill likely drive relatively slower than those preferringdriving very fast and aggressively. Also, users’ drivingbehaviors usually vary in their progressing driving experiences.For example, traveling on an unfamiliar route, a userhas to pay attention to the road signs, hence drive relativelyslowly. Thus, a good routing service should consider thesethree aspects (routes, traffic, and drivers), which are farbeyond the scope of the shortest/fastest path computing.Usually, big cities have a large number of taxicabstraversing in urban areas.
The preprocessed taxi trajectories, we detect thetop-k frequently traversed road segments, which are termedas landmarks. The reason why we use “landmark” to modelthe taxi drivers’ intelligence is that: first, the sparseness andlow-sampling rate of the taxi trajectories do not support usto directly calculate the travel time for each road segmentwhile we can estimate the traveling time between twolandmarks (which have been frequently traversed by taxis).Second, the notion of landmarks follows the naturalthinking pattern of people.The threshold _ is used to eliminate the edges seldomtraversed by taxis, as the fewer taxis that pass two landmarks,the lower accuracy of the estimated travel time(between the two landmarks) could be. Additionally, we setthe tmax value to remove the landmark edges having a verylong travel time. Due to the low-sampling rate problem,sometimes, a taxi may consecutively traverse three landmarkswhile no point is recorded when passing the middle(second) one. This will result in that the travel time betweenthe first and third landmark is very long. Such kinds of edgeswould not only increase the space complexity of a landmarkgraph but also bring inaccuracy to the travel time estimation.