The current options for wireless signal prediction in indoor scenarios generally either lack precision or require immense computational resources. Thus, a new method is proposed that attempts to consolidate the desired accuracy with an easy to implement and time efficient scheme. This work identifies and takes advantage of dominant physical qualities of indoor environments to solve indoor channel problems using a hybrid of numerical and analytical approaches. Specifically, the guiding effect of hallway structures is investigated as they allow electromagnetic fields to propagate with relatively low attenuation. Combining waveguide mode analysis and rigorous numerical techniques, the proposed prediction model computes the hallway fields in a large building floorplan both quickly and with good accuracy in comparison to full finite-difference simulations. Signal measurement data will also be used to verify the applicability of the model.

The current options for wireless signal prediction in indoor scenarios generally either lack precision or require immense computational resources. Thus, a new method is proposed that attempts to consolidate the desired accuracy with an easy to implement and time efficient scheme. This work identifies and takes advantage of dominant physical qualities of indoor environments to solve indoor channel problems using a hybrid of numerical and analytical approaches. Specifically, the guiding effect of hallway structures is investigated as they allow electromagnetic fields to propagate with relatively low attenuation. Combining waveguide mode analysis and rigorous numerical techniques, the proposed prediction model computes the hallway fields in a large building floorplan both quickly and with good accuracy in comparison to full finite-difference simulations. Signal measurement data will also be used to verify the applicability of the model.

This thesis describes the use of adaptive equalization techniques to compensate for the intersymbol interference (ISI) that results when digital data is transmitted over a multipath radio channel. The equalization structures covered in this work are the linear transversal equalizer (LTE), the fractionally spaced equalizer (FSE), the decision-feedback equalizer (DFE), and the maximum-likelihood sequence estimation (MLSE) equalizer. This work also covers adaptive algorithms for equalization including both the least mean squares (LMS) and the recursive least squares (RLS) algorithm. All these equalizer structures and algorithms will be modeled using various simulation modules. Equalization for both stationary and mobile radio channels is considered. Stationary channels are modeled with a simple exponentially decaying profile. The mobile radio channel is represented using a two-ray Rayleigh fading model for an outdoor environment. The SIRCIM channel modeling tool is used to create channel profiles for an indoor mobile radio channel. Adaptive arrays and their similarities to linear equalizers are also studied in this thesis. The properties and performance of simple adaptive array systems using the LMS and RLS algorithms are examined through simulation. This thesis concludes with an in-depth study of the use of adaptive equalization for high-speed data systems operating in an indoor environment. Both stationary and slowly varying radio channels are examined. Simulations of DFE and MLSE equalizers operating in such a system show that both equalizer structures provide better BER performance over a system with no equalization. These simulation results also show that the MLSE equalizer provides better performance than the DFE in almost all cases, but requires a great deal more computations. / Master of Science

Machine-to-machine (M2M) communication is becoming an important aspect of warehouse management, remote control, robotics, traffic control, supply chain management, fleet management and telemedicine. M2M is expected to become a significant portion of the Industrial Internet and, more broadly, the Internet of Things (IoT). The environments in which M2M systems are expected to operate may be challenging in terms of radio wave propagation due to their cluttered, multipath nature, which can cause deep signal fades and signal depolarization. Polarization diversity in two dimensions is a well-known technique to mitigate such fades. But in the presence of reflectors and retarders where multipath components arrive from any direction, we find the detrimental effects to be three-dimensional and thus consider herein mitigation approaches that are also 3D. The objectives of this dissertation are three. First, to provide a theoretical framework for depolarization in three dimensions. Second, to prepare a tripolar antenna design that meets cost, power consumption, and simplicity requirements of M2M applications and that can mitigate the expected channel effects. Finally, to develop new channel models in three dimensional space for wireless systems. Accordingly, this dissertation presents a complete description of 3D electromagnetic fields, in terms of their polarization characteristics and confirms the advantage of employing tripolar antennas in multipath conditions. Furthermore, the experimental results illustrate that highly variable depolarization occurs across all three spatial dimensions and is dependent on small changes in frequency and space. Motivated by these empirical results, we worked with a collaborating institution to develop a three-dimensional tripolar antenna that can be integrated with a commercially available wireless sensor. This dissertation presents the testing results that show that this design significantly improves channels over traditional 2D approaches. The implications of tripolar antenna integration on M2M systems include reduction in energy use, longer wireless communication link distances, and/or greater link reliability. Similar results are shown for a planar antenna design that enables four different polarization configurations. Finally, the work presents a novel three-dimensional geometry-based stochastic channel model that builds the channel as a sum of shell-like sub-regions, where each sub-region consists of groups of multipath components. The model is validated with empirical data to show the approach may be used for system analyses in indoor environments.