Browsing by Author "Wilson, Michael"

Now showing 1 - 1 of 1
  • Thumbnail Image
    Item
    Congestion-Aware Routing (CAR): Vehicular Traffic Routing based on Real-Time Road Occupancy Estimates
    (MAY 2016. ) Wilson, Michael
    Burgeoning of data and road networks have been a major bolster for the global economy. These two sectors have been in competition for a while and whichever wins tends to render the other obsolete. I believed if there were transportation systems that would allow people to teleport from one continent to another in a split second at a lesser cost than the cost of sending an email, half of email users would switch from emailing to face-to-face message delivery. Traffic congestion, however, plaques the roads of most cities globally but especially in developing countries. Data networks have however, attained a state of better traffic management and transfer protocols, proper control algorithms, optimized routing and better traffic decongestion strategies as compared to road networks. This thesis work attempts to address the problem of routing vehicular traffic in order to ameliorate vehicular traffic congestion. Fair routing is effected using real-time data. Routing information is in the form of which route provides the fastest set of interlinked road segments between any departure-destination pair of nodes of road network. Routing traffic in such a way minimizes the overall traffic congestion on road networks. The thesis adopts a routing algorithm used in data networks and modifies it for application to vehicular traffic routing in road networks. The adopted algorithm is Dijkstra’s Shortest Path First (SPF) algorithm. To adapt SPF algorithm to the solution of the traffic congestion problem, the concept of classical Proportional Integral Derivative (PID) controllers and the Eddie model of traffic modeling are exploited. The PID controller concept is used to adjust road occupancy data acquired from a sensor network superimposed on the road network. Simulations have been performed to show the computations involved in deriving travel time from road occupancy data. Also, a hypothetical road network has been used to illustrate the vehicular traffic routing algorithm. Finally the method is implemented as a mobile application that draws data from a database server on the sensor network of a road sub-network in Accra, Ghana. A use case is presented. It has been shown that the method is more practicable and easier to use than a method in literature—Spatial and Traffic Aware Vehicular Routing (STAR).