Traffic Simulator and Embedded RealTime Warning System with RADARs
Our developing world have been bringing serious problems that needs to be addressed delicately. Perhaps the most important problem that arise after the improvement in industries can be elaborated as traffic. Especially road traffic, if not managed efficiently, could cause congestions that would bring environment, energy, health, and most transparently economical problems which degrade the quality of our lives. Under this motivation, this work addresses the issue of traffic congestion.
In this work, literature research on how congestions behave and discrete traffic modeling is given. With this information, a graphical MATLAB-based simulation tool is further developed that is able to simulate up to three-lanes of highway traffic, given parameters such as simulation, vehicle and road characteristics. This tool is also used to simulate the traffic congestion, effects of construction areas, and effects of trucks on roads. Furthermore, a warning algorithm given the data obtained from a Doppler-RADAR is presented.
Basic idea about the simulator is given in the following video:
The results concerning real-life tests in highways in Germany is analyzed and how real-life limitations affect warning algorithm is discussed. The results and system limitations are further discussed in order to create a real-word real-time traffic warning algorithm.
Real-world warning system that is developed is tested on Germany highways under the permission of a highway installation company. The real-world real-time congestion algorithm is developed so that it gets the refined version of the warning algorithm from the offline tests and it has to involve real-world interfaces inside. In the project, the real-world algorithm gets the RADAR-Signal Processing Unit data from an FTP server every 60 seconds and produces a warning output. The results obtained from the embedded warning system is tested using a Linux computer and the results prove the essential diagrams of the traffic stream. The obtained congestion levels showed the backward movement of congestion as well as the flow-density relationships using a real system.
Further information and results are stated in the thesis and might soon be shared given proper permissions.
Technology overview: Traffic flow analysis, Nagel-Schreckenberg Model, RNSL Model, DM Model, Falcon Plus 3 RADAR, MATLAB, Beaglebone, Python.
The following pictures involve our system setup, graphical user infernace, a glance at the algorithm and a log...