Radio Hardware and Software
What is a radio? A radio is some kind of wireless network between a receiver and transmitter. One may think that this technology has been around only for a short time, but in reality, Guglielmo Marconi made the first radio transmitter in 1895. Using a spark transmitter, (which utilized a large potential across an air gap to transmit the spark signal) Marconi was able to transmit a signal across the Atlantic in 1901 [1]. Since then many advances have been made in the area of wireless communication. Even though many things have changed since the spark transmitter, many concepts remain vital to communication theory that must be understood to effectively manipulate signals across the air.
In general, when the conversation of wireless communication is started, the term ?RF? is used often. ?RF? stands for radio frequencies. RF is a very general term for a large range of frequencies in wireless transmission. Essentially the frequencies from 0 to 300GHz (which is basically voice all the way up to microwaves) are considered Radio Frequencies. For more information on how the spectrum has been allocated see FCC.GOV for more information.
Software Defined Radios
The term ?software radio? was first coined by Joe Mitola in 1991 as a ?class of reprogrammable or reconfigurable radios [2].? According to Dr. Jeffrey Reed, a professor at Virginia Tech and author of ?Software Radio: A Modern Approach to Radio Engineering? (2002), a software radio is a flexible radio that can be able to accommodate formats and protocols yet to be determined.
Hardware Vs. Software
First, let's look at a classic "Hardware" radio, that everyone is somewhat familiar. Below is a very simplified block diagram of a superheterodyne transceiver, which can be used to make an AM or FM radio and most other types of radios.
Figure 1: Superheterodyne Transceiver Block Diagram.Figure 1 shows a very abstract way in which we can represent the components in a communication device, but in reality the device is much more complicated, filled with many passive (resistors, capacitors, inductors, etc.) and even active components as well (transistors, diodes, etc.). Although this ?Hardware? radio design has been studied years for many and proven reliable in many applications, it is still limited by a number of factors.
With the rise of computers and hand-held devices, the world demands gadgets at our fingertips. Not only do we want them to be small, but we want them to be very powerful. Able to connect to the Internet, text and call our friends, and even download information to our home computers. Unfortunately, even with transistor sizes reducing at the rate of Moore's law [3], an all hardware device isn't as common as it used to be. FGPA and DSP chips are developed to hold millions of lines of code to replace hundreds of components in cell phones, laptops's, PDA's and many other electronic devices. Clearly, the need for a more software minded approach in RF design has already been developed in technology today, so now what is the next step and why?
Ideally, a completely flexible ?software radio? would connect an antenna directly to the digital to analog converter. But unfortunately that is not as easy as it seems. In light of the technology and research being done today, any software-defined radio is NOT a pure ?software? radio. It still needs some sort of hardware front-end.
So, if the SDR still needs hardware, then how does it differ from the superheterodyne design in Figure 1? In the diagram below (Figure 2), the main difference is in the left most two stages. Due to the lack of computational power of the A/D converters available most SDR's still must down convert the RF signal to a lower IF (Intermediate Frequency) before converting it into baseband digital data.
Figure 2: Software Defined Radio Block Diagram.An Example: The Mysterious Digital Signal Processing Block
The software defined radio block diagram shown in Figure 2 is a very abstract model of how an SDR is implemented, but in order to really understand what the SDR is doing, we must have a deeper look into the mysterious "Digital Signal Processing" block shown in Figure 2. The mysterious block is essentially a set of objects defined in software. The types of objects used in each SDR will be varied infinitely depending on the application and software used in the SDR. In order to illustrate this concept an example will be used to clarify the purpose of the "Digital Signal Processing" block.
Figure 3: AM Transmitter Objects.In Figure 3, some of stages of the AM transmitter are shown to illustrate that each of these stages can be defined in software and each of these stages would be considered software "objects." These objects can be connected to multiple other objects or combined to make one larger single object.
Figure 4: AM Transmitter Block Diagram.Figure 4 is the combination of the objects from Figure 3. For the application of the AM transmitter Figure 4 becomes the mysterious DSP block from Figure 2. The arrows are also objects that are defined in software that "connect" the inputs and outputs of each block object in their defined order. Even the connection to the "Hardware Interface" is a function that sinks the signal generated to a Digital to Analog Converter (DAC) and eventually to the outside world.
With a better idea of how SDR difference from their conventional ("hardware") counterpart, a brief look at the advantages and limitations of software radio will be explored.
Advantages of SDR
Flexibility
The idea of a more flexible radio has already been mentioned throughout this discussion, but it becomes the most obvious and important advantage over classic radio systems. The ability to upgrade your system not only for better performance, but also for new performance means endless possibilities for software radios as they are developed.
Cheaper RF Front-End Design
One problem with classic RF design is the complexity and the labor in developing a reliable design. With the design of a reliable SDR, the quality and performance of the SDR can be enhanced by the digital hardware (Digital Signal Processor (DSP), Field Programmable Gate-Array (FPGA), or General Purpose Processor (GPP)) in order to reduce the complexity (and therefore the cost) of the RF front-end.
Smaller List of Components
Another related benefit to SDR is that by making a less complicated RF front-end, the total parts needed (theoretically) become much more simplified. With digital components like the DSP and FPGA taking the place of many passive and active components the list is smaller and hopefully cheaper in the long run.
Faster Time To Market
With the ease of porting software to the SDR in order to upgrade the system less time can be spent on manufacturing which will eventually bring the product to market faster than classic designs.
Limitations of SDR
The limitations of SDR still relatively few, but there are some very foreseeable obstacles that must be overcome in order to create a functional SDR mainstream.
Software Reliability
If the SDR is sending a signal at 2 Watts in a particular band when FCC regulations say its max output can only be 1 Watt, then there is a serious issue. Although these are certainly obstacles they are not deterrents from using SDR. Software development has been available for years for communication systems in more specific applications (cell phones, WiFi, etc.). The problem is controlling and monitoring multiple waveforms, regulations, and rules that has yet to be fully developed.
Security
One great benefit of SDR?s is the flexibility to reconfigure new waveforms, but what if you are not allowed access to certain waveforms? Some precautions have to be made to allow different levels of access to waveforms and spectrum as determined by local and federal regulations.
Cost
One limitation, at present, that will always be improving is cost. As of today?s available technology, the SDR technology is either too monstrous (for example the DMR covers a huge spectrum but weighs 210lbs [4] which is certainly an indication that it is not cheap) or on the other end, some SDR?s are still limited in its capabilities (for example, the USRP is around $500 [5] but is limited to 128MS/s transmit and 64MS/s receiving). So as technology develops more and more options will be opening to the radio community.
Power Consumption
One ever improving and constant battle in communications devices is power. One problem with SDR?s at present is the computationally heavy applications are not able to run at high speed unless they are using large quantities of power. As battery and chip technology improves, the consumer market will converge on reasonable power restrained SDR for everyday use
References
[1] www.ieee.org/organizations/history_center/milestones_photos/transatlantic_radio.html
[2] J. H. Reed, Software Radio: A Modern Approach to Radio Engineering, 1st Edition, New Jersey: Prentice-Hall, 2002.
[3] http://www.intel.com/technology/mooreslaw/index.htm
[4] http://www.gdc4s.com/documents/D-DMR-5-0704.pdf
[5] http://www.ettus.com/custom.html
[6] M. J. Roberts, Signals and Systems: Analysis Using Transform Methods and MATLAB, First Edition, Boston: McGraw-Hill, 2004
[7] Leon W. Couch II, Digital and Analog Communication Systems, Sixth Edition, New Jersey: Prentice-Hall, 2001.
[8] http://www.fcc.gov/aboutus.html