Using Cooperative Intelligent Transport Systems for Real-Time Determination of Dangerous Locations in Traffic
Master thesis
Permanent lenke
http://hdl.handle.net/11250/2560151Utgivelsesdato
2018Metadata
Vis full innførselSamlinger
Sammendrag
A prototype solution for continuously determining the current level of danger in traffic at different locations has been developed, based on event detection on a stream of Cooperative intelligent transport systems (C-ITS) data. Detection of abrupt braking events from Cooperative awareness messages (CAMs) was chosen in order to provide a proof-of-concept of the solution, and an algorithm for the detection of such events has been constructed. Datasets were generated in real-life traffic in order to tune and evaluate the algorithm. This shows potential, but the algorithm suffers from a lack of larger volumes of data that is needed in order to train it for higher accuracy.
Abrupt braking events are combined with pre-detected Decentralized environmental notification message (DENM) events through a system of relative weighting based on a measure of event severity. The spatial and temporal characteristics of traffic events have been modeled mathematically, providing a way to express their effect dynamically on a map in real-time.
The full solution has been implemented in Python, as a series of modules that effectively divide the full problem into intuitive subtasks that can be developed and tested independently. The solution regularly outputs a list of dangerous locations and their calculated level of danger, which can be used as an input to other systems. Additionally, a way to visualize the incoming data, important intermediary results and the final output has been implemented, which, among other things, allows a human operator a clear view of the current level of danger in traffic at all times.