| Autor: |
Paul, C. J., Acharya, Anurac, Black, Bryan, Strosnider, Jay K. |
| Předmět: |
|
| Zdroj: |
Communications of the ACM; Aug1991, Vol. 34 Issue 8, p80-93, 14p |
| Abstrakt: |
Real-time systems are playing an increasingly vital role in today's society. Such systems include manufacturing, control, transportation, aerospace, robotics and military systems. No longer are Real-time systems limited to low-level control functions. They are now being asked to monitor and control complex, hierarchical systems in dynamic, sometimes hazardous environments. Furthermore, some real-time systems such as the Mars Rover are being asked to operate with little to no human interaction. Other large real time systems are required to operate in environments that are not fully characterized. The lack of information and uncertainty of environment requires the use of problem-solving techniques. To make things more difficult, real-time tends to be critical in nature where the impacts of failure have serious consequences. The authors begin the article by discussing the sources of execution time variance and methods to deal with them. The authors then develop the requirements of real-time problem solving architecture, providing not only the mechanics to tackle the variance, but also the functionality required for integration into real-time environments. Later, the authors examine these issues in the context of CROPS5. An aircraft collision avoidance system is used as an example of how problem-solving tasks can co-exist with conventional real-time tasks on a common computing platform while maintaining guaranteed response time performance for the conventional real-time tasks. |
| Databáze: |
Complementary Index |
| Externí odkaz: |
|
