Chirp Channel Estimation and OFDM Transmission using Discrete Prolate Spheroidal Sequences

S. Senay, L.F. Chaparro (USA), and A. Akan (Turkey)


Orthogonal Frequency Division Multiplexing (OFDM); Prolate Spheroidal Wave Functions; Slepian se- quences; Chirp channel estimation; Time-frequency disper sive channels.


In conventional Orthogonal Frequency Division Multiplex ing (OFDM) systems, the Fast Fourier Transform process ing corresponds to rectangular pulse shaping. Such a pulse has spectral leakage causing the modulated subcarriers to have considerable frequency dispersion during transmis sion and thus losing orthogonality at the receiver. In this pa per, we exploit the frequency concentration of the Slepian or discrete prolate spheroidal sequences (dpss) to generate basis for OFDM with maximum energy concentration in each of the subcarrier bands. As such, minimum interfer ence from other subcarriers occurs. Since the performance of the OFDM system depends on accurate estimation of the parameters of the transmission channel, we use a simple but effective estimation method based on chirps. This estima tion method permits accurately estimation of the channel parameters: the number of paths, the time-shifts and the Doppler frequency-shifts associated with each path. The estimation problem becomes one of finding the frequency of complex exponentials embedded in noise, which can be done using a periodogram for which the accuracy of the parameters depends on the resolution of the Fast Fourier Transform used. In this paper we propose an improve ment to the two-chirp channel estimation, previously pro posed, by sending a frequency modulated cosine as pilot sequence instead of two chirps of different rates. In the simulations, the bit error rate (BER) results show that the proposed method is able to considerably increase the spec tral efficiency of OFDM systems compared to bases using the conventional rectangular pulse, and that it performs as well as bases that use an optimized combination of Slepian pulses.

Important Links:

Go Back