ITR: Collaborative Research: 'Free' Bits: The Challenge of the Wireless Internet

Second and third generation wireless systems have been designed primarily for voice, so they are con-nection oriented, delay sensitive, and provide fixed bit rates. Additionally, since service is desired 'any-time/anywhere,' these systems must provide ubiquitous coverage. This coverage is achieved through rela-tively uniform grids of cell sites, which are placed to control interference and minimize outage rather than to maximize throughout. As a result, such systems deliver low bit rates and are relatively expensive when used for large amounts of information. At the same time, wireline connection to the Internet has encouraged uses (and users) that depend upon bits being virtually 'free.' We contend that this 'economic' mismatch between wired and wireless access is the primary obstacle to the dramatic growth of a wireless Internet. The solution may lie in designing systems specifically for wireless data, recognizing that data services are often connectionless, delay insensitive and have no specific bit rate requirements. These differences

suggest that ubiquitous (anytime/anywhere) coverage is not a strict requirement for wireless data networks and makes possible systems in which small, separated coverage areas facilitate transfers of megabytes of data in fractions of a second, and for a fraction of the cost associated with conventional ubiquitous coverage. Communication theory and simple link budget calculations tell us that it is possible to build such sys-tems, but the signal processing challenges are numerous and distinct from the historical challenges offered by connection-oriented wireless services. When a mobile user passes an Infostation, there will be a window of opportunity, perhaps as short as a fraction of a second, in which the user will have access to a high-rate communication channel. A key task is to identify that window and transmit at an appropriate rate. The mobile must make these decisions based on measurements of a wideband radio channel in which there is frequency selectivity and time variation in the fading as well as in the interference. In the specific context

of an Infostations system, we plan to divide our research into four components:

Radio Channel Modeling: The characterization of typical Infostation radio channels.

Transceiver Design: The analysis and performance evaluation of transmitters and receivers for both single carrier and multicarrier systems.

Radio Resource Management Transmitter power and rate adaptation policies derived from receiver measurements.

Algorithm Development Testbed A platform employing DSP and FPGA technology for the practical evaluation of transmitter and receiver algorithms.

The activities of this project will encompass three research institutions in New Jersey (New Jersey In-stitute of Technology, Princeton University, Rutgers University) under the auspices of the N.J. Center for Wireless Telecommunications (NJCWT). The NJCWT is an inter-institutional research and educational or-ganization sponsored and funded by the N.J. Commission on Science and Technology. The focus of the center is a multi-year effort in Digital Radio Technology for Computing, Communications and Information Systems. This effort is supportive of and will enhance the present proposed project in wireless networks.