Decision-making based on measurements of behavioural safety - An explorative study into using Situation Awareness measurements for a specific problem as input to system-level decisions.

ProRail is trying to increase the capacity of the rail infrastructure since the need for capacity is still growing. One of the ways to do this is by implementing DDDR (Dynamic double non-interrupted braking), which is part of the DSSU project. By implementing DDDR the expectations of the status of the signs of the train drivers might change. Therefore ProRail wants to investigate if implementing DDDR will have a negatively impact on safety. In this thesis is investigated how situation awareness (SA) can be measured and if it is a useful safety indicator to take safety decisions. SA measurements can be used for testing if the perception, comprehension, and expectations of the signs in the environment are influenced by DDDR. Decisions are made based on the SA; an informed, conscious decision cannot be made without a correct SA. SA is a requirement, not a guarantee, for making informed, conscious decisions. Three factors are identified that influence the acceptability of the SA measurements. First, the safety culture has an impact on the acceptability. It is more likely that negative results will be neglected when companies have a bad safety culture. Second, strategic decision-making has an impact on the acceptability. In all phases of decision making; goal formulation, alternative generation, and evaluation and selection, should safety be incorporated. Other interest, such as economic interest, might get the upper hand when safety is not incorporated in all three steps. Third, the use of the Common Safety Method (CSM) is a formal requirement for the acceptability of the results. The CSM has three requirements for measuring safety: the safety of a new design should be comparable with the current situation, since the safety should be at least as good as the current situation, the method should correctly reflect the system, the used parameters and results should be accurate enough to serve as a robust decision-support. The best method for comparing designs is a controlled before and after study. For the best comparison between designs, the researcher can use a simulator in order that he can exactly determine which elements are changed in the situation. Since people can show different behaviour when they are being tested, they should not be informed when they are being tested. Arbitrary measurement moments can be chosen to enhance that participants do not show different behaviour when they are studied; the participant cannot prepare for the measurements. To make the outcome of the study more robust, a mixed method design can be used; the conclusion of a study is more robust if all the methods have the same outcome. Situation awareness global assessment technique (SAGAT) is the best-known direct and objective measurement method for SA; this is the recommended method for determining SA for train drivers. By combining this method with the Mission awareness rating scale (MARS), a more robust answer can be found. In the DSSU study, SAGAT was capable to identify differences in SA between scenarios. MARS was not capable to show significant differences between scenarios, because of the low amount of participants. The SA study was combined with a workload study via eye-tracking and biometric measurements, and a qualitative time-to Signal Passed at Danger (TT-STS) study. These studies combined concluded that the probability of a Signal Passed at Danger (SPAD) decreases in DSSU compared to the current situation. Generalizability should be taken into account before a study is executed if in the future new SA measurements will be held. In the DSSU project the outcome of the study only applied to DSSU. By using the identified elements in chapter 7, scenarios can be made which are expected to have a bad influence on the SA. The worst-case scenario for SA can be used for measuring SA in the project. If this scenario is not applicable for the researched project, management has to decide whether they want results that are only applicable for the project or generic results, which are applicable on a system-level. The SA measurements showed to be a useful indicator for safety when it is combined with TT-STS measurements. The combination of the measurements gave a detailed insight in changed behaviour and awareness in the new situation. SA cannot be used as a single safety indicator, since SA is a requirement, and not a guarantee, for making informed, conscious, and safe decisions.Engineering And Policy AnalysisPolicy, Organisational Law & GamingTechnology, Policy and Managemen