With the advent of software radios, a fundamental paradigm shift has occurred in radio design. Traditional treatments of radio engineering have emphasized analog circuit design using discrete components. While analog Radio Frequency (RF) design remains an important component of radio engineering, the radio engineers of the future will need to understand a broader set of issues, including signal processing methods, power management, the interface between the analog and digital portions of the radio, and the role of the radio as a node within the larger network. Indeed, the design of software radios calls for a body of knowledge that is largely absent from current radio engineering texts. Specific topics needed are described below.

• Digital generation of signals - Many signals in the transmitter and receiver of the radio will now be generated digitally. In many cases, the direct digital synthesis methods used to generate these signals will be more than just digital realizations of analog oscillators and will afford the designer greater freedom in design signal waveforms.

• Analog to digital conversion - Great care is needed in constructing the boundary between the analog and digital portions of the radio receiver. The analog to digital converter samples, quantizes, and, in some cases, downconverts the received signal. A rigorous understanding of these operations is needed to choose tradeoffs between the resolution, sample rate, and dynamic range of the resulting system.

• Digital signal processing techniques for demodulation and synchronization - While these are familiar problems to radio engineers, the software radio will implement these operations completely or partially through signal processing. In many cases, the best DSP solutions are more than just a digital approximation to the corresponding analog circuit.

• Advanced processing techniques for range extension and interference rejection - The availability of high speed signal processing capabilities facilitates the use of advanced techniques such as adaptive equalization, adaptive error correction, interference rejection, and smart antennas, which may previously have been too complex to implement in commercial radios. These algorithms offer the system designer new tradeoffs between performance and complexity.

• New power management strategies - Traditionally, the transmitter power has been the predominant determinant of power consumption within a radio, and great efforts have been expended to optimize the link budget. However, for systems that transmit over short distances, other components of the radio such as DSP chips, analog to digital converters, and digital to analog converters become major power considerations. New power management techniques such intelligent sleep modes and reconfigurable computing must be considered. The performance improvement obtained by additional signal processing techniques must be weighed against the increased power requirements.

• Need for understanding of network considerations - In traditional radio design, the radio link is merely a conduit for data that can be largely separated from the design of the larger radio network. However, a software radio is more than just a mere pipe for information; it is an active node within a larger network. As a result, it must react to changing conditions on the network and may even rely on the network itself for instructions to change between modes of operation. Therefore, the software radio designer must be aware of the relationship between the radio and the network.

The resulting body of material required for an understanding of software radio design is quite different from any existing radio engineering textbook. Therefore, I believe that the emergence of DSP-based radios necessitates the development of a next generation textbook on the subject.