Simulation-based development of a cognitive assistance system for Navy ships

Cognitive technology can support increasingly complex missions, but with increasingly complex and even cognitive technology, the inner and outer compatibility with the human becomes more and more important. In combat information centres (CIC) of Navy ships operational data about objects in the environment are processed in a complex command and control cycle. Due to different task domains and an increasing performance of both modern sensors and data transfer, the command and control processes are characterized by high complexity that might affect the operators' workload and consequently the error rate in the workflow. One approach to improve this situation may be the use of assistance systems. In addition, the lack of competency and skills or the absence of the qualified operators caused by demographic changes will make assistance systems indispensable in the future. In a simulation study a cognitive assistance system (COGAS) is being developed for supporting the crew, and especially the decision maker, of a CIC during air target identification. COGAS contains two supporting modules called cognitive units, which are combined to fulfill its support functions by means of flexible automation. Both cognitive units are based on Rasmussen's Decision Ladder which describes the behavior of well-trained and motivated operators controlling complex dynamic systems. This article outlines the functional structure of COGAS and its cognitive units. Both units contain a memory of a-priori knowledge, a memory of present situation knowledge, and activities for processing data. The structure of supporting cognitive units, which comprise modules of the relevant system environment and system goals, as well as possible system tasks and actions, is described in some detail. In order to simulate CIC working processes with or without COGAS, we utilized the Integrated Performance Modeling Environment (IPME). IPME provides model components which represent the activities of the system, COGAS- and the operator in the form of task networks. IPME also helps to analyze human performance in highly complex systems and provides models for determining operator workload. This paper describes the development of COGAS components and their IPME implementation.