Real-Time Systems. Flight-Critical Systems

The computers on aircraft are required to perform their functions within short times. Flight control systems must make fine adjustments quickly in order to maintain stable flight. Sensor suites must detect and analyze potential threats before it is too late. Cabin pressure and oxygen must be regulated as altitude changes. All these activities, plus many others on aircraft, must happen in real time.

Nielsen (25) defines a real-time system as a controlled (by software or firmware) system that performs all of its process functions within specified time constraints. A real-time system usually includes a set of independent hardware devices that operate at widely differing speeds.

These devices must be controlled so that the system as a whole is not dependent on the speed of the slowest device. Hatley and Pirbhai (26) describe timing as one of the most critical aspects of modern real-time systems. Often, the system’s response must occur within milliseconds of a given input event, and every second it must respond to many such events in many different ways.

Flight-Critical Systems. Flight-critical systems are those activities of an aircraft that must be completed without error in order to maintain life and flight. The aircraft flight controls, engines, landing gear, and cabin environment are examples offlight-critical systems. Failures in any of these systems can have catastrophic results. Flight-critical systems are held to tight levels of performance expectations, and often have redundant backups in case of failure.

Federated Systems. Federated systems are loosely coupled distributed systems frequently used in aircraft system architectures to tie multiple processors in multiple subsystems together. The loose coupling allows the multiple subsystems to operate somewhat autonomously, but have the advantage of the shared resources of the other subsystems. A typical aircraft federated system might include its central computer, its INS, its radar system, and its air-vehicle management system.

The INS provides the radar with the aircraft’s present position, which is reported to the pilot through displays put forth by the central computer. The pilot adjusts his course through the air-vehicle management system, which is updated by the INS, and the cycle is repeated. These subsystems perform their individual functionality while providing services to each other.

Cyclic Executive. A cyclic executive on an aircraft computer provides a means to schedule and prioritize all the functions of the computer. The executive routine assigns the functions and operations to be performed by the computer. These assignments are given a specific amount of clock time to be performed. If the assignment does not complete its task in its allocated time, it is held in a wait state until its next clock period.

From the beginning of the clock period to its end is one clock cycle. High-priority functions are assigned faster clock cycles, whereas low-priority functions are assigned slower cycles. For example, the high-priority executive function might be assigned a speed of 100 cycles per second, whereas some lower-priority function might have 5 cycles per second to complete its tasks.

Sometimes, the latter might take several clock cycles to perform a task. An additional feature of cyclic executives is that they are equipped with interrupts, which allow higher-priority systems to break into the executive assignments for system-level assigned tasking.

There are several types of scheduling methodologies that provide performance improvements in cyclic executives. One of the more prominent is rate monotonic analysis (RMA), which determines the time requirement for each function and the spare time slots, and then makes time assignments.