As shown in Figure 1, finding skyline query results in such DCSs looks like a straightforward problem. If all unfortunately sensor nodes store sensor readings, we only need to scan boundary sensor nodes such as sensor node
The two main concerns with the increasing number of vehicles on the roads are congestion and safety. In the USA alone, congestion accounts for 115 billion dollars in fuel costs [1], with similar figures in other developed countries. Worldwide traffic casualties amount to 1.17 million per year [2]. In this context, Intelligent Transportation Systems (ITS) aim at enhancing transportation efficiency and safety through the use of advanced information processing, communications, control, as well as new electronic technologies.Sensing the environment is a major aspect of ITS, as well as of other novel applications in future vehicular scenarios.
Traditionally these systems have relied on different alternatives [3]. One group frequently employed to detect traffic flows comprises intrusive sensors, including sensors such as inductive loops, magnetometers, pneumatic road tubes and diverse kinds of weigh-in-motion sensors. However, the installation and maintenance of these sensors has important associated costs, since large sections of the road need to be torn up, disrupting traffic flow. Other non-intrusive sensors can also be used, such as video cameras, GSK-3 radars, acoustic arrays and ultrasonic sensors, which can be placed above ground. Their main drawbacks are that they are usually large-sized, power-hungry sensors and may be affected by different environmental conditions.
In addition, both intrusive and non-intrusive sensors are expensive and associated with difficult installation, classically requiring wired infrastructures and power lines for energy supply. This leads to the deployment customer review of those sensors only at critical locations, which work independently of each other. The information they produce must be transmitted to distant Traffic Management Centers (TMCs) for centralized data processing, which require the transmission of high amounts of data through expensive communication infrastructures. In general, this results in unacceptable data dissemination delays which limit the utilization of the system for vehicle safety applications requiring a quick response (even real time in most of the cases).An alternative to these highly centralized solutions is the use of a cooperative approach where processing is performed in-situ among distributed devices, enabling faster reaction times. In addition, if this is combined with wireless communications, some of the inconveniences derived from the emplacement of nodes may be alleviated. Vehicular Ad Hoc Networks (VANETs) are an example of such combinations [4].