Error floor in OFDM systems

The time dispersion of the mobile radio channel causes large intersymbol interference (ISI) in the high-data-rate- systems that are used today in mobile radio applications, and also digital audio broadcasting and digital TV. One way to combat these problems is to use OFDM (orthogonal frequency division multiplexing). In an OMM system, the transmitted bit stream is divided into many substreams, each of which is modulated onto a different carrier. These substreams, which have much lower data rates, then suffer from much less time dispersion. Still, some residual ISI remains, lowering the transmission quality and the spectral efficiency by an (often unacceptable) amount. The purpose of this research is to find ways to combat also this ISI in an efficient way. The proposed research will run along three main lines: (i) Adaptive determination of the optimum sampling time. This method has been recently proposed by us for single-carrier modulation, and in that case gives a strong decrease of the error floor. The basic idea is that depending on the distortions introduced by the instantaneous impulse response of the channel, the optimum sampling instant may vary over more than one bitlength. We propose to use this scheme also for OFDM, and to choose also the sampling frequencies in an adaptive way. (ii) We suggest that the basis pulses for the transmission of the data should not be chosen as rectangular pulses, but as some smoother waveform that exhibits a stronger decay in the frequency domain. Especially, we propose the use of nonorthogonal basis pulses - contrary to common belief, perfect information transmission is also possible with these pulses, and they have properties that are advantageous in dispersive environments. (iii) We will analyse adaptive modulation. This concept, which has been suggested in the last few years, means that we do not transmit the same data rate with the same modulation format over every subchannel, but adjust the data rate to the SNR of the subchannels. We suggest that in dispersive channels, the data rate should not be decided just by the SNR but also by the channel distortions. The reduction of the error floor will result in better transmission quality and, since we need less coding, a better spectral efficiency for high data rate systems.

The frequency spectrum is scarse and expensive! This can easily be shown by the horrible license cost for UMTS. Hence, mobile operators try to transmit the users` data as fast as possible, as reliable as possible, and in a frequency band that is as small as possible. This goal can be reached with modulation formats with so-called high spectral efficiency. GSM (2nd generation) and UMTS (3rd generation) use transmission systems with a quite small spectral efficiency. The international communication community is therefore investigating systems that transmit a multiple of data of todays systems in a bandwidth like the 5MHz of, e.g., UMTS. The OFDM system (Orthogonal Frequency Division Multiplex) that was investigated in this research project theoretically reaches up to 16 times faster transmission than the 2MBit/s of UMTS. Up to now, state-of-the-art OFDM had a hard upper bound in spectral efficiency and thus in the transmission speed. We showed that even this "hard limit" can be crossed. It is possible by using exact information of the mobile radio channel and the utilization of mathematical and communications-theoretical methods. Such theoretical results are finally applied and used by computer programs running in small microprocessors inside the handies and base stations of future mobile radio systems. Prof. Ernst Bonek and his team of the mobile communications group at the Institute of Communications Engineering of the Vienna University of Technology have recognized years ago that a thorough knowledge of the mobile radio channel is the basis for all further investigations. Thus, they had a lot of research projects investigating radio channel properties in the last years. His group is known as one of the leading teams in international mobile radio research. The results of our project bring us one big step towards the 4th-generation mobile communications system.