This paper presents the results of a simulator experiment conducted within the Traffic Management Center Simulator at the Georgia Tech Research Institute. In this experiment, design issues for automated incident detection systems were of interest. The objective was to explore results from an initial simulator investigation of incident detection systems. Initial results indicated that a system operating under relatively high false alarm and hit rates, as well as a short detection latency, led to better operator performance. In the follow-on experiment described here, a system was simulated operating under higher false alarm and hit rates than those specified in the initial investigation. Analysis of the most recent human performance data reinforced the initial results. Despite its increased false alarm and hit rates, the new support system led to superior operator performance. Follow-on data offered no empirical evidence to suggest that the increased false alarm rate either degraded operator performance or increased workload levels.
Abstract