Power spectrum optimization in the design of multisine manoeuvre for identification purposes

In this paper, two sets of multisine signals are designed for system identification purposes.
The first one is obtained without any information about system dynamics. In the second case,
the a priori information is given in terms of dimensional stability and control derivatives.
Magnitude Bode plots are obtained to design the multisine power spectrum that is optimized
afterwards. A genetic algorithm with linear ranking, uniform crossover and mutation
operator has been employed for that purpose. Both designed manoeuvres are used to excite
the aircraft model, and then system identification is performed. The estimated parameters
are obtained by applying two methods: Equation Error and Output Error. The comparison
of both investigated cases in terms of accuracy and manoeuvre time is presented afterwards.