1.1 Current Development
In recent years, a primary concern of road users is to realize the condition of the roads which they may take. Traffic Information Systems (TIS) collect raw data from a variety of sensors on the roads, and integrate the data into accurate and reliable
real-time traffic information, such average speed, travel time and traffic events. Using the information provided by TIS, road users can easily plan their travel route to avoid congested area and save time and fossil fuel.
In order to collect valuable traffic information from the sensors, a growing number of methods are now available to detect traffic congestions. These methods can be divided into two categories: static sensing and mobile sensing. The static sensing methods, uch as vehicle detectors (VD), radar devices and video image processor, installed on the roadways, observe the vehicles passing by and record the traffic speed and traffic flow. These stationary sensors require electrical power supply and a
communication link to send the traffic data to a Traffic Information Center (TIC).
When TIC receives traffic data from stationary sensors, it needs to check the
correctness of the data before generates reliable traffic information and announces the information to the road users. Restricted by the costly installation and maintenance fees, static sensors are mainly used for traffic surveillance on highways and freeways. On surface roads, the distance between the stationary sensors could be far part. As a result, raw data collected cannot generate accurate traffic conditions for the whole road network.. In addition, the failure rate of stationary sensors are usually high because of the exposure to the extreme temperature. Alternatively, traffic information can be
2
collected by mobile sensing devices. Using vehicles equipped with Global
Positioning System (GPS) receivers and wireless communication capability as probes, the TIC can collect real-time traffic information, such as vehicle speed, acceleration, heading direction and longitude and latitude coordinates. the GPS-equipped probe vehicles usually transmits these data to the TIC periodically or when they pass pre-determined locations. The information received by the TIC can represent the current road condition. Unlike the traditional stationary sensing technique described above, mobile sensing is more cost-effective and have a wide coverage of sensing area.
Without additional installation of sensing devices along roads, the TIC can collect traffic information from the GPS-equipped probe vehicles. Herrera, et al., found that with 2-3% penetration of probe cars in road networks, it is enough to provide accurate measurements of the network speed [1].
1.2 Motivation
In general, mobile sensing in traffic can be classified into two categories: floating cellular data and GPS-based probe cars. The principle of floating cellular data is to locate vehicles from the signals of cellular phones or mobile stations (MSs) constantly registering their locations to the cellular networks. Between two signals of an MS on a moving vehicle, we can use the difference of time to compute the speed of the vehicle.
This method is referred to as cellular floating vechicle data (CFVD). Compared to GPS-equipped probe vehicles, with the huge number of cellular phones, CFVD does not need additional hardware installed on vehicles. However, the low accuracy in positioning MSs of cellular network is the critical weakness, and this position
inaccuracy may result in inaccurate traffic information generated by CFVD, especially in urban areas.
3
Considering the low accuracy of CFVD and the costly installation and maintenance of stationary sensing, GPS-equipped probe vehicles solve the traffic information collection problem effectively. Relying on on-board GPS devices, we can figure out the condition of each moving vehicle. To obtain real-time traffic information, GPS-equipped probe car need to report the condition of cars, such as timestamp, speed, and direction to the TIC in a short time interval. The shorter period of report means the more real-time traffic information, but the TIC needs to deal with amount of messages sent from probe cars in a short time, which increase the burden on the TIC. Therefore, how to deal with these raw data to detect the unusual events on mobile devices and report to the TIC is becoming an important issue.
1.3 Objective
In this thesis, we proposed a novel traffic report system and method with event-triggered report policy to avoid periodical reports and maintain the
real-timeliness of the traffic information generated by the TIC. Our traffic information system has the following features:
(1) Provide real-time traffic information.
(2) Event-triggered report/broadcast policy.
(3) Economical computation on the TIC (4) Identify the traffic jam area on highways.
(5) A conditional report policy for probe cars.
Our contribution of this thesis is to propose a report policy that not only detects traffic jam areas on highways, but also provides the travel time of each traffic jam area.
4
1.4 Summary
The remaining part is organized as follows. Chapter 2 describes the current works in Floating Car Data related to our system. Chapter 3 describes our system design in details. Chapter 4 discusses the results in our system. Finally, we give our conclusions in Chapter 5.
5