Indoor radio propagation studies in recent years have been motivated by wireless telecommunications applications such as PCS, Wireless LANs, wireless PBX, etc. and their focus has been towards characterizing the signal coverage and RMS multipath delay spread that influence system infrastructure and transceiver design. Emerging position location (geolocation) applications, (civilian, such as E-911 services, and military) require a significantly different characterization of the radio channel. The parameters of importance in this case are the probability of detecting the direct line-of-sight path and the distribution of range error when it is not detected.
Measurements, simulation, analysis, and modeling of the indoor radio channel are considered with geolocation applications in view. Qualitative and quantitative frameworks are developed incorporating the parameters of importance for geolocation applications. The drawbacks of existing wideband channel models (for e.g. the JTC models for indoor PCS applications) are illustrated. New radio channel models are proposed for geolocation applications and techniques for a generalized model that is suitable for both telecommunications and geolocation applications are suggested.

Site-specific statistical modeling for indoor radio channel using the results of the ray-tracing (RT) simulation is presented in this thesis. The motivation of this study are to yield (1) a more accurate statistical propagation model than the ad-hoc standard statistical models such as JTC [JTC94], (2) a less expensive statistical model than the measurement based site-specific statistical models such as those developed in [Gan91][How91], (3) a less computationally complex model than the brute force RT [Yan94]. The time domain statistics of arrival times and path amplitudes of the third floor of the Atwater Kent laboratories at WPI obtained from results of RT are compared with the statistical results obtained from empirical data. A fifth order autoregressive (AR) model is adopted to characterize the channel response in the frequency domain. The poles obtained from the RT channel response are compared to the poles obtained from measurements. A new statistical model for site-specific modeling of indoor radio propagation is proposed to replace the use of the brute forec RT method. This model relates power prediction directly to the floor plan and the material used in the construction of the building. The predictions using the new model in a typical office environment are checked against the results of RT prediction and measurements.

This dissertation provides analytical frameworks for evaluating the performance of a WLAN for different patterns of traffic. First, the performance of the WLAN with ALOHA and CSMA access methods for voice and data traffic is analyzed. In this system, TCP is used for the data packets to guarantee accuracy of information and UDP for the voice packets to handle the delay requirement. Then the capacity of a WLAN access point is related to the type of applications such as web access, email processing, Telnet, FTP, and data base access. The behavior of the two way traffic for each application is modeled with a statistical ratio of the up-link to down-link traffic. Using empirical data from HP Open View, WLRL benchmark tool, and MS Internet Explorer the statistics of the network parameters are measured. The statistical information from the empirical data is used in the analytical framework to obtain the number of users per AP for different applications. The analytical framework introduces the effects of natural hidden terminal in the overall throughput of a WLAN and also includes the effects of capture caused by the near-far problem.

This dissertation uses the results of ray tracing in a typical indoor environment to compare the performance of five radio transmission techniques that are under consideration as the air interface by standards bodies. The performance of direct sequence and frequency hopping spread spectrum are compared with the performance of multi-carrier modems, as well as modems using decision feedback equalization and sectored antenna systems. Validity of using ray tracing for performance evaluation is examined by comparing the results with the results of performance evaluation obtained from the empirical wideband measurement of the channel characteristics. Using the definition of the JTC air interface standard for indoor office environments, it is shown that the test area qualifies as a typical office environment with low to moderate delay spread. Based on the maximum achievable data rate and minimum power requirement, operation of all modems in bandlimited applications are discussed and compared.

he design and layout of PCS or WLLANs requires accurate characterization of radio channels. A 3D universal ray tracing model for indoor and outdoor microcellular radio propagation is presented. The model is developed for the multipath propagation based on ray optics and the Uniform Geometrical Theory of Diffraction (UTD). Propagation mechanisms for specular reflection, specular transmission, diffraction and diffuse scattering are included in the model. The accuracy of predictions is examined with measurement results obtained in an indoor and an urban radio environment. These results include propagation loss, statistics of RMS delay spread, effects of antenna pattern and polarization. The accuracy and suitability of the model allows us to use it to characterize a typical indoor environment for performance evaluation of various high speed wireless data systems. This thesis presents a comprehensive investigation of various performance enhancing techniques to increase maximum data rates of systems.
Using the ray tracing model, the thesis first examines the outage probabilities of BPSK and BPSK/DFE radio modems with omnidirectional and six-sector antennas. The effects of room size, data rate, transmitted power and sector antenna patterns on the systems are analyzed. Comparative performance evaluation of DFEs and sector antennas provides a comprehensive understanding of the behavior of sector antennas. The thesis also evaluates the performance of multicarrier BPSK and QPSK radio modems with and without Reed Solomon coding schemes in the same typical office environment. The effects of key parameters such as number of carriers, bandwidth efficiency, roll-off factor of filters and adjacent channel interference are analyzed.

 

Multipath propagation measurements performed in various indoor environments are analyzed to develop a time domain statistical model for the wideband characteristics of the indoor radio channel. The measurements are broadly classified as local, spatial, and partitioned. The local measurements are performed to analyze the effects of movement of personnel near the transmitting or receiving antennas and/or local movements of the terminals around a given location on the channel characteristics. The spatial measurements are analyzed to determine and compare the channel characteristics in two different environments: manufacturing floors and college office areas. Five manufacturing floor areas and three college office areas are used as measurement sites. The partitioned measurements involve dividing the cite into three partitions on the basis of building architecture and number of walls between transmitter and receiver. Channel characteristics are studied and compared for the partitions. The spatial and partitioned measuremens are used to develop a statistical model for the impulse response, suitable for computer simulation of multipath profiles in each environment. To evaluate the performance of this simulation model, the distribution of the rms delay spreads from the simulated profiles is compared with that obtained from the measured profiles.

A new statistical autoregressive model for the frequency selective behavior of the indoor radio channel is presented. The model was derived for frequency domain measurements that were performed as paret of this research. The measurements were made in the 0.9-1.1 GHz band. It is the first known use of a frequency domain system, which is relatively simpler than other measurement systems, to characterize and model the indoor radio channel. A two pole autoregressive model is shown to regenerate channel frequency responses that have the same characteristics as those observed in the actualy measurements. Compared to existing models used for computer simulation of the indoor radio channel, the model uses fewere parameters (random variables). To examine the accuracy of the model, the peroformance f a BPSK modem (with and without decision feedback equalization) is presented as derived from actual time domain measurements of the indoor radio channel, a time domain simulation of the channel, the frequency domain measurements of the channel, and the proposed frequency domain model of the channel. The closeness of the results from the frequency domain measurements and the proposed model reveals the accuracy of the model. All the results are consistent and show similar performance.

This thesis discusses channel access protocols and voice/data integration methods used for digital wireless local area networks. The network is assumed to have a star configuration witha fixed central station. The performance of demand assignment and contention based protocols are related to the propagation characteristics of the radio channel. A mathematical model for the calculation of the probability of capture in contention based access protocol for fading channels is introduced. This model relates the performance to the modulation and coding of the system, distance power relation of the indoor radio channel, length of the packets and the distribution of the terminals. The performance of slotted ALOHA and Carrier Sense Multiple Access (CSMA) protocols are analyzed using this capture model. Then, a high speed integrated voice/data system using framed polling and movable boundary scheme is proposed and analyzed. This system uses multiple antenna diversity and double rate transmission to achieve a channel througput in the order of megabits per second to comply with the current data rates used in the wired LANs.

Modeling of the indoor radio channel for the purpose of predicting data rate limitations is discussed. The channel is considered to be frequency selective and fading. The performance of equalized and conventional modems is found using continuous and discrete channel models. A channel model having a finite number of channel paths at Poisson distributed delays is found to have the same form, in the limit, as a model continuous in dleay. The importance of path population as the distinguishing feature between the models is analyzed. The sensitivity of the performance to the shape of the delay power spectrum, the shape of the envelope delay power spectrum, and the arrival rate of the paths is discussed. The probability of outage for a decision feedback equalizer is derived and evaluated. It is determined that, for both conventional and equalized modems, the shape of the delay power spectrum or delay power spectrum envelope is irrelevant in determining peformance for useful data rates. This is verified for the equalized modem by comparing performance results based on exhaustive channel simulation and results based on the probability density of the delay spread only. For continuous channel modeling, the delay spread of the delay power spectrum determines performance. For discrete channel modeling, knowledge of two parameters is required: the delay spread of the delay power spectrum envelope as well as the mean arrival rate of the paths.