Autopilot

The aerodynamics, propulsion, and mass properties, activated by Newton’s and Euler’s equations, model the vehicle dynamics. They define the plant to be con­trolled. An aircraft needs rate damping for stability augmentation, roll control for coordinated turns and flight-path-angle tracking. For long duration flights, altitude and heading hold autopilots are indispensable for a weary crew. Missiles, flying in the atmosphere, are primarily controlled by rate and acceleration feedback loops. A roll controller maintains their orientation. For longer duration flights the trajec­tory can be shaped by a flight-path tracker or altitude hold autopilot. Outside the atmosphere missile attitude is controlled by angular feedback loops.

The feedback signals come from sensors that are located throughout the vehicle. The INS, an indispensable component in any modern aerospace vehicle, provides rate and acceleration feedback for all three axes, directly from its accelerometers and gyros, and attitude angles from its navigation computer. If wind is not a factor, it can also calculate vertical and horizontal flight-path angles. The air data computer provides Mach number and dynamic pressure. Using incidence angles as feedback signals is problematic because it is difficult to accurately measure or compute the angle of attack or sideslip angle. The alpha vanes that you see on the outside of airplanes are only used as warning devices and not as autopilot sensors.

The coverage in this section picks up from Sec. 9.2.3 where I discussed autopi­lots of pseudo-five-DoF simulations. As you may recall, in five-DoF simulations autopilots do not issue control commands to the control surfaces but model closed-loop response, delivering angle of attack to the trimmed aerodynamic ta­bles. In the real, nonpseudo world the autopilot, or the human operator, controls the aileron, elevator, and rudder directly, and only through the airframe and the sensors are the dynamic loops closed.

Most of the autopilot designs of Sec. 9.2.3 do not apply here, with the exception of the altitude hold autopilot. Its outer altitude position and rate loops can be maintained as long as the inner acceleration autopilot is replaced by its six-DoF equivalent. So let us get a new start and build control systems for aircraft and missiles that are suitable for our six-DoF simulations.

Before we begin, let me caution you, however, that my autopilot designs are simplified versions of what you actually find onboard these vehicles. Aircraft flight control systems can be very sophisticated and are beyond the scope of this discussion. If you are tasked to model an existing vehicle, you should get the autopilot specification and replicate it as faithfully as possible. Only then will the customer trust the results of your analysis. However, if your job is to study a generic system or to develop a new concept, you will find the following designs useful.

To make the autopilots serve a variety of vehicles and flight conditions, I employ the pole placement technique. You can find this method discussed by Stevens and Lewis.9 Yet if you have some familiarity with classic and modern linear control, you should be able to follow this self-contained presentation.