| dc.description |
System identification techniques are routinely used in experimental stability and control
studies throughout the aerospace industry. Over the years, various researchers at the
College of Aeronautics have contributed to this field; most recently some of the latest
methods have been employed to estimate the stability and control derivatives of a variety
of aircraft types. Although the more recent investigations provide a useful insight into
the capabilities and characteristics of several up-to-date methods, they have not resulted
in tools which may be used on a routine basis.
Consequently, the purpose of this report is to describe a set of procedures which are
straightforward to apply, and produce reasonable solutions to the type of linear parameter
identification problems which are often found in aerospace work. Recordings of the short
period and phugoid modes from Handley-Page Jetstream G-NFLC are used throughout as
examples.
Firstly, those characteristics of the aircraft’s instrumentation system which influence
the quality of the signals - sample rate, antialiasing filters, time delays - are considered.
This information is used in conjunction with standard signal processing techniques to
ensure that the data is of sufficient quality to be used in the parameter estimation process.
Next, a basic Fourier analysis and a least squares algorithm are employed to produce non-
parametric and parametric models respectively. The results thus obtained are comparable
to those generated using more sophisticated techniques.
In conclusion, standard signal processing methods combined with relatively simple
estimation theory offer an adequate solution to the linear parameter estimation problem. |
|