Design of the real time aspects of associate systems
John M. Hammer · 2002
A pilot's associate system contains artificial intelligence modules that assess the situation, recommend responses, manage displayed information, monitor for errors, and adaptively execute tasks on behalf of the pilot. Artificial intelligence modules represent a risk to real time performance because their run time performance is typically both slow and unpredictable. A methodology is described for using scheduling theory to allocate run time resources to modules to address this problem. Once the communication and module execution sequences have been characterized, the methodology is simple to use and understand, and it also provides an assessment of real time risk.>