Technical paper on Software radio

The pace of technological advances is so fast that the communication devices become obsolete within just few years of their production. To keep up with this pace, communications systems must be designed to maximize the transparent insertion of new technologies at virtually every phase of their lifecycles. When these new technologies are inserted, the upgraded devices should still be able to communicate with each other.

The term Software radio (SR) was coined in 1990s to overcome these problems. A software radio is a communications device whose functionality is defined in software. Defining the radio behavior in software removes the need for hardware alterations during a technology upgrade. By moving the radio functionality from hardware to software, promises to change the economics of deploying and operating wireless networks. In order to maintain interoperability, the systems should be built on a well defined, standardized, open architecture.

Wireless networks are heavy spectrum consumers. Spectrum being very scarce quantity there is a problem in allocating spectrum to new standards and services because it could hinder the future developments in this industry. At this stage the developments in the software defined radio is a boost to the wireless industry as a whole.

In this paper we deal with the following points:

  1. Drawbacks of traditional radios.
  2. Digitalization in traditional and software radio.
  3. Software Radio architecture.
  4. Advantages and disadvantages of SR.


Introduction:

The Problem: The wireless industry has emerged as one of the fastest growing industries in the world. So as the industry is increasing rapidly there is a lot of competition and hence there are a lot of improved services. But all the equipment will not be able to support the new technologies. It requires a hardware upgrade and hence it is very costly. New wireless systems take up large spectrum areas. Spectrum is a scarce resource and therefore allocating spectrum to new standards and systems would hinder the development of future systems.

One of the main problems in the wireless systems is the interoperability. To communicate between two separate standards it is not possible with the traditional radio.

To overcome all these problems the concept of Software Radio has evolved. SR proposes the new way of developing wireless systems. The traditional radios are hardware based but SR is a software-oriented application.

The term “Software Radio” refers to reconfigurable or reprogrammable radios that can show different functionality with the same hardware. Because the functionality is defined in software, a new technology can easily be implemented in a software radio with a software upgrade. Physical radios can function over different services providing seamless operation. More efficient use of spectrum is possible because same equipment is able to operate in different standards.

Traditional Versus Software Radio:

Before dealing with the software radio let us look at how traditional radio works and the drawbacks of it. The traditional radios are based on hardware-oriented approach. The hardware-oriented approach is the main cause of the high development costs, long times, low flexibility. Let us first review the wireless signals.

Wireless signals are radio waves, usually in the MHz and more recently GHz bands, in which information has been inserted. Receivers extract the information from the radio waves and present it in a suitable form like audio or video to the final user. Transmitters perform the inverse function. This process requires multiple steps that are carried out in a chain of hardware pieces. Figure 1 exhibits a simplified model of the hardware chain for a traditional radio receiver. The antenna collects the radio waves in the MHz or GHz bands, called radio frequency (RF) signals. In the case of a GSM base station, the antenna receives 124 channels of 200 KHz each situated in 890-915MHz band. The antenna presents the RF signal to the receiver.

Extracting information directly from an RF signal is difficult and expensive since a mix of channels is received at the antenna. First, a RF filter selects the desired channel. The RF filter must be tunable, i.e. it must be able to select different channels if the communication changes between different channels. Manufacturing accurate tunable filters is expensive. Cheap filters are usually placed at the RF stage. In consequence, the output of the filter is not of high quality. To eliminate the adjacent bands, the signal is first down converted to a lower and fixed frequency called intermediate frequency (IF). The channel would be down converted, At this point, the signal is filtered again by an IF filter to eliminate adjacent components.

The hardware-oriented approach of traditional radios imposes a set of limitations. First, traditional radios have low flexibility to adapt to new services and standards. Each hardware element of the radio chain performs a radio function. These components are designed to operate in a particular frequency band (RF) and standard. When the frequency or any of the parameters of the standard changes, traditional radios cannot correctly extract the information. Before being able to operate under the new conditions, the system must be redesigned and hardware modules have to be replaced. Redesigning, manufacturing and replacing hardware components require high times and costs. Traditional radios present long times and high costs for the development and manufacturing of new products.

Traditional radios are also limited in the number of services they can provide. When two or more services are to be integrated in one device then one hardware is required for each service. For example if you need to integrate two services like GSM* and IS-95# in one cell phone you need to arrange totally two hardwares in one box.

*GSM ------ Global System For Mobile Communications.

#IS-95 ------ CDMA (Code Multiple Division Access)

Software Radio: A Software Approach

The traditional radio follows a hardware-oriented approach. Hence dedicated hardware is required for every application. Contrary to the traditional approach, the SR follows a software-based approach. It is Software that plays the major role in extracting the information but not the hardware. In SR receivers, analog-to-digital converters (A/D) digitalize the analog RF signals.

Signal processing techniques extract the information from the digitalized samples. As in traditional radios, the information is presented with the aid of digital-to-analog converters (D/A) in a suitable form like audio or video to the final user. In software radios, general-purpose processors that run special software, together with A/Ds and D/As, replace the chain of hardware components of traditional radios. SR software carries out not only usual radio functions, but also advanced features like channel selection and error correction.

The use of general-purpose processors and signal-processing software increases the flexibility to adapt to new services and standards. New software is installed and hardware pieces do not need to be replaced. Software development and production require lower times and costs than the development of hardware modules.

A/D converters digitalize the user information and provide the software running over general-purpose processors with the digital samples. These samples are treated and D/A converters generate the signal to be transmitted by the antenna.

Digitalization:

Digitalization converts the analog signals received at the antenna into digital samples. Signal processing techniques treat the samples to extract the information. Digitalization right after the antenna, i.e. before the RF filter, is the most flexible approach since it allows treating the signal fully in software. However, this kind of digitalization is currently impossible to implement due to the state of the art of analog digital converters (A/D) and the limitations on computational capacity of present processors. Digitalization may take place at other points of the traditional radio chain: after the IF filter or after the demodulator at the base band stage. Traditional radios use no digitalization or base band digitalization. IF digitalization is the solution currently implemented in software radios. This section explains each configuration and discusses their advantages, disadvantages and limitations in the frame of software radios.

RF Digitalization:

In RF digitalization, an analog-digital converter (A/D) digitalizes the radio waves collected at the antenna. Signal processing software running over general-purpose processors extracts the information from the digital samples.

IF Digitalization:

To surmount the present problems of RF digitalization, SR designers place A/D converters after the IF stage. This design requires an RF front-end, which consists of an RF filter; an RF/IF converter and an IF filter. The RF front-end selects and converts the signal to IF, as do traditional radios. Before demodulation, an A/D converter digitalizes the signal. Signal processing running over general-purpose processors extracts the information.

There are two advantages in this method. First, current A/D converters can achieve enough speed and resolution at IF frequencies. Second, this design requires less computational resources because the tunable RF filter of the front end limits the number of received channels reducing the burden of software channel selection.

Base band Digitalization:

Digitalization at baseband level is common in traditional transceivers. Information is analogically extracted and baseband sampling is used in subsequent stages to profit from signal processing techniques such as music equalization. This is a common practice in widely used devices such as car radios. Because none of the radio functions for information extraction is carried out in software, radios using baseband digitalization are not considered software radios but traditional equipment.

Evolution Of Digitalization point from traditional radio towards software radio

Processors:

The traditional radios use dedicated hardware. They used ASIC’s (Application Specific Integrated Circuits), DSP’s (Digital Signal Processors), FPGA’s (Field Programmable Gate Arrays). They had low flexibility. They were unable to support new services and standards. But these specific task oriented processors can attain high efficiency because they are designed to perform particular task only. The manufacturing costs and times are very high for these processors.

Now in the SR, the programmable ASIC’s or general-purpose processors are used. The new processors have high flexibility and high speeds but less performance because they are not specific task oriented.

Upgradation Of Radio from Present Standard to a New Standard:

(Traditional Vs Software Radio)

Traditional Radio:

Let us consider a traditional cell phone with dual channel mode (i.e. GSM/IS-95) and a software radio with dual channel mode (i.e. GSM/IS-95). We need to upgrade both the phones to a new standard known as DCS. If we consider the traditional one the phone consists of hardware for GSM and IS-95.So to upgrade to DCS standard new hardware must be installed on the phone. The software control also has to be up graded. This up gradation is a very costly process and also increases the size of the cell phone.

Software Radio:

To upgrade a SR to a new standard we just need to download the new software (DCS standard) and the rest of the hardware need not be changed. This up gradation is quite a cheap process and doesn’t increase the size of the cell phone.

Software Radio Architecture

To maintain good operability between various standards the software radio must be built on a well-defined, standardized, open architecture. The key subjects to put SR to use in a handheld device are Application Program, Radio Function Program, Software specification language, Digital radio processor, Broadband RF stage.

Application program

The AP is written in some software specification language that describes the radio architecture and performance using a radio function library. Application Programs are prepared for a specific radio standard, for example, one for GSM, one for CDMA1, an AP for IMT-2000, and so on.

The performance of the handheld SR is easily reconfigured by rewriting the AP to change the use of the library. AP is downloaded over the air from a central or base station to the SR terminal.

Radio function library

This is a software set that expresses basic radio function. The examples of this library are control programs for hardware such as analog-to-digital converter (ADC) and digital-to-analog converter (DAC). Software programs for modem function embedded in DSP are another example. In this case, the changeable input parameters include resolution and sampling rate corresponding to modulation scheme, bandwidth, and multiplexing.

Software specification language

This will be a high-level language simply to express radio characteristics and specifications for each mobile communication standard. Using this language, the AP programmer will write the physical-layer specification such as transmission power, data rate, modulation scheme, channel codec, and speech codec

Digital radio processor

The DRP is a general purpose hardware functional module in LSI configuration, which can be used for several standards. The DRP typically includes ADC, DAC, and FPGA and DSP hardware. FPGAs and/or DSPs express digital down/up converters, digital filters, and modem functions. Modulation scheme, bandwidth, and transmission rate can be controlled by changing DRP parameters. In general, two architectures will be available where one is intermediate frequency input configuration and the other baseband input. The configuration should be selectable according to the modulation scheme.

The receiver specifications should be changed by changing parameters. It is preferable to be able to change ADC resolution for each standard, for example between GSM and CDMA.

Software Radio Physical Layer

A dynamic stack using software radio as the physical layer. Each layer can be replaced or reconfigured at runtime.

Using the dynamic stack structure it is possible to create a flexible physical layer featuring software radio. The software radio physical layer encapsulates another set of dynamic layers that form the software radio system. Software radio can take advantage of dynamic layer loading and reconfiguration as these features are built into the structure of the stack. This means that the software radio implementation becomes truly flexible. For example, new modulation schemes and coding algorithms can be swapped in and out as required. Figure shows a dynamic stack using software radio as the physical layer. All layers can be replaced or reconfigured at runtime.

Applications:

1. Automatic Modulation Detection and Loading:

Modulation schemes are changed on a per-packet basis. In this case the modulation scheme is determined by a packet header that identifies the modulation scheme. It is possible to determine the modulation scheme of a signal without prior knowledge of the modulation scheme used at the transmitter. It is implemented using the Dynamic Stack Structure. In this scenario, the stack detects the modulation scheme of the incoming signal and reconfigures the stack accordingly. This powerful feature could be used to provide flexibility for roaming mobile devices. Devices could dynamically load new modulation schemes and become part of new networks without changing hardware.

A layer in the communication stack acts as a modulation detector and loads the appropriate layers to demodulate the incoming signal. The modulation detection layer performs an analysis of the incoming signal to determine the modulation scheme used. This information is added to the attributes of the message block being passed up the stack. The modulation scheme loader makes sure that the appropriate layer is available to demodulate the incoming signal. If a suitable layer is not present, the loader will load the appropriate modulation scheme and reconfigure the stack so that demodulation can continue.

2.Peer-to-Peer Component Sharing:

Using the concept of Over-The-Air Reconfiguration (OTAR), a remote device can be reprogrammed by the transfer of new software into the device. Up to now hardware restrictions has meant that OTAR has only been used for once-off projects such as satellites. We believe that advances in hardware will allow OTAR to be used in ad hoc networks.

An ad hoc network is a loosely connected set of mobile wireless nodes without any centralized or hierarchical structure. In this scenario, it is feasible that such large numbers of nodes will be using many different protocols, modulation schemes and location-specific parameters for RF communication. The ability of nodes to share information about network conditions is of key importance to ensuring reliable communication. We have developed this concept into ‘peer-peer component sharing’, which goes beyond information sharing. In this scheme nodes can offer each other actual layer components as downloads. These components are the layers making up the structure of the dynamic stack. This approach allows a node to reconfigure its communication stack without contacting a central source.

For example, a node entering a new location could contact the nearest node using a common modulation scheme to download the most popular communication component used in that area. These settings would include information such as available bandwidth, frequency allocations and noise conditions. In addition, nodes can share location specific data that can be used to optimize communications. Flexibility of this nature is not available in any current technology.

Technical Barriers Of Software Radio:

Not only A/D converters and processors limit SR development. Other technical barriers slow down the development of SR technology. Batteries are an important problem for SR handsets. A/D converters have high consumption of power. Signal processing requires lots of computation that also imposes high power consumption. Power supply is not usually a problem in network equipment but it is in handsets, where autonomous operation could be limited to one or two hours with current batteries. The second barrier is amplification. The RF filters not only limit in frequency the signal received at the antenna but also amplify it to compensate the attenuation due to the propagation over the air. Quality and bandwidth are amplification trade-offs. If the bandwidth of the signal is large, amplifiers may cause distortion on the edges of the bandwidth.

This problem is particularly important in software radios using an RF front end, where the filters have to amplify signals of larger bandwidth than in traditional radios. Important research efforts are being dedicated to obtain high quality amplification over extensive bandwidths.

Finally, cost is a discouraging component for SR handsets. Nowadays, for up to three standards, traditional implementations are less expensive than SR. For four and more standards, SR handsets are cheaper.


Conclusion:

Software defined radios enable us to build reconfigurable and interoperable radios that can be upgraded for future technologies. Defining an open architecture and implementing radios compatible with this architecture further enhances interoperability. In future the handheld terminals will become “Future Proof” by having the capacity to download new air interfaces and operate new communication standards, using the reconfiguration technology of Software Radio. The future of wireless communications is Software Radio.

References:

  • Joe Mitola, “The software Radio architecture”, IEEE Communications Magazine, 1995, vol.5, May, 26-38.
  • Philip Mackenzie, Linda Doyle, Donal O’Mahony, Keith Nolan, “Software Radio on General-Purpose Processors”, Networks and Telecommunications Research Group. Trinity College, Dublin 2.
  • Sabri Murat Bicer, “A Software Communications Architecture Compliant Software Defined Radio Implementation”, PhD Thesis, Northeastern University, Boston, Massachusetts, 2002.
  • P K Saxena and Jagdish Prasad, “Emerging Trends in Mobile communication”,IETE Technical Review vol.20, No.4, July-August 2003, pp 297-307.
  • Hiroshi Tsurumi and Yasuo Suzuki, “Broadband RF stage Architecture for Software-Defined Radio in Handheld Terminal Applications”. IEEE Communications Magazine, 1999, vol.2, Feb, 90-95.
Technical paper on Software radio
4/ 5
Oleh