ISSUE 182 : Jul/Aug - 2010 - Australian Defence Force Journal

FREEDOM. They conclude that the systems issues to be addressed are increased emphasison visual CID, more reliable target identification decision aids and a greater emphasis on SA,in combination with the use of target indication aids. The failings identified in the Patriotsystem were attributed to undisciplined automation during the development of the systemand misuse of the automation by crews.In this latter case, the problems identified refer to an implicit assumption that the operatorwill take care of whatever ‘foibles’ the system may have. This is a classic developmentdilemma, given an engineer’s enthusiasm to capitalise on the potential performance of acombination of new sensors and software integration opportunities. In order to cope withunacknowledged—and perhaps even unknown—system fallibilities, operators need to havea very intimate knowledge of the hardware and embedded software and a detailed contextrelevantunderstanding of how the system responds to inputs. The traditional coalitionmanning and training regimes aim for this type of knowledge and skill. But it is extremelydifficult to attain without prolonged and costly exercises and a big investment in generatinguncompromised performance feedback, while testing the complete system under extreme andcompetitive conditions.The section on SA has just two papers. The first by Bolstad, Endsley and Cuevas (Chapter9), focuses on a framework for analysing information requirements for individuals andteams by using a ‘goal-directed task analysis’ method. This method of analysis, conductedby knowledgeable practitioners, reveals systems insights that generate requirements forinformation, systems design and team training.The second paper reports on a demonstration at the Center for Network Centric Cognitionand Information Fusion at Pennsylvania State University (Hall and Aungst, Chapter 10). Theirdemonstration exploits the fusion of information from human observers, called ‘soft sensors’,who provide direct observations by text, voice reporting, images and full motion video. TheCenter is carrying out applied research on how one might fuse this data with other internetdata sources to analyse the design implications of using the human as an active observer toprovide the context that machines, with their powerful pattern recognition techniques, havedifficulty in doing. The relevance to the military is the idea that every warfighter is a ‘softsensor’ and that we might be able to exploit information technologies to provide greater spanof control in counter-insurgency operations.A paper by Mistry et al (Chapter 14) takes the classic ‘task analytic’ approach to close airsupport (CAS) system requirements. Their analysis produced an activity sequence with anemphasis on cognitive processes, which provided the framework for making observationsin field exercises and laboratory-based simulation experiments. While the paper reports thetask activity sequence and indicates the importance of team coordination between the pilotand the controller, it indicates that further insights will be generated by building an agentbasedmodel of the cognitive and system attributes of the manned CAS system. Through thismodel-based systems analysis approach, they should be able to illustrate some variables in CIDperformance and make the system engineers and training practitioners aware of critical issuesin preparing these warfighting systems for combat.98