This thesis investigates the feasibility of using AIS technology as a support for determining the risk picture in naval transport. Traffic data is gathered from the AIS interface and compared with accident data from NMD (Norwegian Marine Directorate) to form a statistical foundation for a risk model.
A risk model based on conditional probablities is then developed based on both data from the Norwegian waters as a whole, as well as data from the Bergen seaward approach. The model is then tested on the inner Oslo fjord, providing promising, if not exact, results.
A number of known accidents are also investigated trying to find either common traffic patterns when accidents occur or common traits that “define” an accident. Traffic patterns will help develop a completely dynamic risk model and traits that define accidents will help reveal unreported accidents. Both of these areas are considered to be major benefits of the AIS system, though unfortunately not much work have been published on these topics.
The main outcomes of this project are:
- An overview over published reports and articles regarding AIS and risk analysis
- A working, though not exact, risk model with suggestions for improvements
- An overview over accident traits and suggestions for how to reveal these from larger data sets
- Suggestions for improvements in the online AIS interface
It is concluded that AIS technology greatly improves both the statistical foundation for risk models, as well as the possibility to apply the models to a correct traffic pattern. With some minor improvements in the AIS interface, revealing unreported accidents will also become possible. More work in topics regarding practical use of AIS data is highly recommended.
2011. , 95 p.