Date of Award

Fall 2006

Document Type


Degree Name

Doctor of Philosophy in Computer Engineering - (Ph.D.)


Electrical and Computer Engineering

First Advisor

Ali Abdi

Second Advisor

Yeheskel Bar-Ness

Third Advisor

Alexander Haimovich

Fourth Advisor

Hongya Ge

Fifth Advisor

Dimitry Chizhik


The Doppler spread, or equivalently, the mobile speed, is a measure of the spectral dispersion of a mobile fading channel. Accurate estimation of the mobile speed is important in wireless mobile applications which require such as knowledge of the rate of channel variations. In this dissertation, first the performance of classical crossing- and covariance-based speed estimators is studied. Next, the problem of mobile speed estimation using diversity combining is investigated. Then, a nonparametric estimation technique is proposed that is robust to different channel variations. Finally, cyclostationarity-based speed estimators which can be applied either blindly or with the aid of pilot data, are developed.

A unified framework for the performance analysis of well-known crossing and covariance based speed estimation techniques is presented. This allows a fair analytical comparison among all the methods. Interestingly, it is proved that all these methods are asymptotically equivalent, i.e., for large observation intervals. The extensive performance analysis, supported by Monte Carlo simulations, has revealed that depending on the channel condition and the observation interval, one needs to use a crossing or a covariance based technique to achieve the desired estimation accuracy over a large range of mobile speeds.

Two common diversity schemes, selection combining (SC) and maximal ratio combining (MRC), are considered for Doppler spread estimation. Four new estimators are derived which rely on the inphase zero crossing rate, inphase rate of maxima, phase zero crossing rate, and the instantaneous frequency zero crossing rate of the output of SC. Two estimators, which work based on the level crossing rates of the envelopes at the output of SC and MRC, are also proposed. The performances of all these estimators are investigated in realistic noisy environments with different kinds of scatterings and different numbers of diversity branches.

Then a novel speed estimation technique is proposed that is applicable to both mobile and base stations, based on the characteristics in the power spectrum of mobile fading channels. The analytic performance analysis, verified by Monte Carlo simulations, shows that this low-complexity estimator is not only robust to both Gaussian and non-Gaussian noises, but also insensitive to nonisotropic scattering observed at the mobile. The estimator performs very well in both two- and three-dimensional propagation environments. By taking advantage of resolvable paths in wideband fading channels, the robustness against both nonisotropic scattering and line of sight can be further increased, due to the differences among the Doppler spectra observed at different paths. This technique is also extended to base stations with antenna arrays. By exploiting the spatial information, the proposed space-time estimator exhibits excellent performance over a wide range of noise power, nonisotropic scattering, and the line-of-sight component. This is all verified by simulation. The utility of the new method is further demonstrated by applying it to the measured data.

Finally, to design robust blind and data-aided mobile speed estimators, a proposal is made to exploit the inherent cyclostationarity of linearly modulated signals transmitted through fading channels. Two categories of cyclic-correlation- and cyclic-spectrum-based methods are developed. Extension to space-time speed estimation at the base station in macrocells is also provided. In comparison with the existing methods, the new estimators can be used without any need for pilot tones and are robust to additive stationary noise or interference of any color or distribution. Unlike the conventional multi-antenna based method, the proposed space-time speed estimator does not assume the receiver noise to be spatially white. A suboptimal training sequence is also devised for pilot-symbol assisted methods, to reduce the estimation error. The performance of the proposed estimators are illustrated via extensive Monte Carlo simulations.