A New Low-Cost Approach to Wireless Communications over Severe Multipath Fading Channels


Frank Althaus


PhD thesis, Universit├Ąt des Saarlandes, Logos Verlag Berlin, May 2003.

[BibTeX, LaTeX, and HTML Reference]


Hard limiting receivers are a proven low-cost technology for narrowband wireless communication systems. To date, modern broadband mobile communication implies linear receiver concepts: the high data rate introduces severe intersymbol interference and known equalizers perform poorly with hard limiting receivers. Space-time processing promises a major increase in spectral efficiency. Popular Multiple Input/Multiple Output (MIMO) and beamforming techniques also mandate linear receivers, in particular because the knowledge of the (multidimensional) channel impulse response is required to train the antenna array. In this thesis two important steps towards a low-cost paradigm for broadband wireless communication systems and space-time processing are reported: a) A new equalizer for hard limiting receivers, that maintains a good performance even in severe intersymbol interference. This equalizer is based on phase samples and performs a Maximum-Likelihood sequence estimation. In contrast to state of the art receivers, it uses the knowledge of the complex channel impulse response. Important system parameters, as sampling frequency, sampling instant and bandpass bandwidth are examined for this new decoder. Furthermore, a complexity reduction of the metric is considered and simpler approximated metrics are proposed. b) A new data aided method to estimate the channel impulse response (CIR). With help of a comparable short training sequence, the complex CIR can be estimated very well. This two level estimation technique is also based on phase samples. The first step estimates the desired phase and amplitude of the training sequence. The amplitude estimate is based on phase variances. The second step performs a subspace approach that provides the CIR. Performance results are supported by simulations for model channels that represent typical and worst case propagation environments.

This publication is currently not available for download.