Improved crest factor minimization of multisine excitation signals using nonlinear optimization

Optimization of the Crest Factor for Complex-Valued Multisine Signals

“…The main advantage of the method is the improved reduction of the CF compared to time-frequency swapping algorithms; the disadvantage is that as in the case of any gradient-based method, it is not guaranteed that the global minimum is found. The algorithm presented in [4] enhanced Guillaume's method by subsequently solving another optimization problem, where the elements of a slack variable vector are minimized, while the absolute value of the time-domain samples of the signal are required to be less than the slack variable via an inequality constraint.…”

“…The main differences in comparison with the algorithm presented for real-valued multisines in [4] are the following:…”

“…In the past decades, a number of approaches have been presented in order to minimize the crest factor, most of them dealing with the real-valued case [4,[15][16][17][18][19] -which is also relevant in the case of complex multisine signals as the I and Q channels are handled by separate D/A converters and mixers.…”

“…It is also hard to approximate, and it might not be the best choice in the case of bandlimited systems. Depending on the limitation posed by the system on the input signal, other signals can also be applied: in case of binary input systems pseudorandom binary sequence [2] or in case of bandlimited systems, chirp [3] or multisine [4] signals can be used.…”

“…An application of the Artificial Bee Colony (ABC) algorithm was presented by [20]. Recently, a nonlinear optimization method [4] was proposed, which can achieve the best performance among the compared methods. This method is based on the algorithm presented in [19] using an interior-point solver, then improves the result by utilizing a slack variable.…”