Study On Driving Behavior And Safety Of Urban Taxi

In many major cities of China, taxi, by providing flexible and convience service, becomes one of the most importation modes of passenger transport and makes contribution to the day-to-day operation of the business world and tourism. However, the increasing popular of taxi has brought several problems to urban traffic. Because of the complex structure, incomplete management, lack of training to taxi drivers in this industry, taxi suffers a high rate of violations and crashes. Other than a tendency to blame taxi drivers, the most often cited safety issue is acturally paying attention. The current study is to explore the characteristics and safety of taxi behavior on urban road.First, by analyzing the initial collected data including travel speed, following distance, lane changing frequency and distance, preference to different lanes, the taxi behavior is exporded. Meanwhile, some associated factors are discussed from the aspect of taxi drivers and travel environment. Factors that affect taxi drivers are the intentions to increase income, long hours driving on urban street, continuous "hunt for hails" while driving, abundant experience of the profession. It can be concluded that there are three main characteristics of taxi driving behavior on the road: close following, frequent lane change, and lane change in small space.On the evidence of the driving behavior of taxi, models which describe the taxi following and lane changing behavior are proposed by considering taxi driver's desired speed, desired follwing diatance, and risk-taking coefficient of changing lane. The lane changing behavior of taxi is simulated. Through extensive simulation analysis, the concept of risk-taking coefficient of changing lane is confirmed. The value of this coffient increases when the driver is more optimism biased. If the average value of the risk-taking coffient for the group of public drivers is equal to 1.0, the average value of this coefficient is 1.1 for the taxi drivers group. Simulation results also suggest that the actural speed and the desired speed make great contribution to the frequency of one driver to change lane.Furthermore, safety of taxi driveling behavior is examined quantificationally using several analysis models. In terms of the numerical examples, vehicle should take a lower speed when following close. It is much easier to be involved in rear end crashe once the lead vehicle takes a sudden brake. The frequent lane changing makes traffic flow to be unstable. It also increases the possibility exposed to crashes. The problem size of one changing lane vehicle to strike or be struck by the following vehicle in the objected lane is the largest. The vehicle that makes lane change is significantly affected by vehicles on the objected lane. In order to join in the new queue successfully, the vehicle makes lane change should be slower than the lead vehicle; it also needs to speed up to exceed the following vehicle. Any wrong doing leads to angle/sideswipe crashes in that course. All these results provide evidence that close following, frequent lane change, and unsafe lane change may reduce the security of taxi's driving. They are predictors of accidents.This study is helpful for a full-scale study of taxi industry and a primary concern in the taxi safety field.