As we all know the next century demand for wireless technology is expected to be the commercial implementation of Software Defined Radio (SDR) paradigm to improve base station efficiency and reduce receiver complexities to provide seamless high quality service to the end user to meet ever increasing customer demands. This will require gradual transition of traditional hardware intensive architectures being replaced by reconfigurable hardware which can run multiple software based radio solutions. The numerous benefits provided by SDR have created widespread interest in modulation recognition problem which is one of the major tasks of an intelligent receiver in a radio communication system in SDR.
In this paper we study modulation recognition problem in some detail. A modulation recognition module in a receiver should be able to identify the modulation format present in the received signal with minimum prior knowledge. Based on this information and user requirement an intelligent receiver will take the best decision regarding the software to be run on the reconfigurable hardware to give best quality of service to the user. All such schemes critically depend on Signal to Noise ratio (SNR) of the received signal. In this paper, we presented two methods for SNR estimation
and discuss a few papers on modulation classification.
Keywords - Modulation classification schemes, Modulation recognition, SNR estimation, Software Defined Radio (SDR).
1. Introduction
Since past few decades, telecommunication systems are continuously evolving from 2G to 3G to 4G etc. This has resulted in multiple technologies and many types of receivers which don't interoperate. As a result seamless roaming is very difficult. Due to constant evolution of link-layer protocol standards (2.5G, 3G and 4G), incompatible radios having specific standards, problem of global roaming and also new standards coming into existence, wireless communication industry is undergoing many hurdles.
Software Defined Radio (SDR) technology tries to address these problems by implementing the functions of a radio in software having a common hardware platform. The software modules required to implement new services/features can be downloaded over-the-air onto the handsets. Thus SDR offers this flexibility without any problem of different standards used [1]. The implementation of software into the radio systems introduced the concept of software radio. Software radios have brought a revolution in the radio engineering. We can implement different radio functions using suitable software running on common hardware platform. Such radios have been referred to as SDR. Software radios are expected to emerge as platforms for multiband multimode personal communications systems [2].
SDR is a rapidly evolving technology implementing radio functions in software on a common hardware platform. Thus, it provides flexibility in reconfiguring the radio and allowing the wireless devices to quickly adapt to the environment. It offers flexibility in implementing radio functionality such as signal generation, coding, and modulation, in software. Thus, the radio can quickly adapt to the environment and it supports multiple standards without requiring different hardware for every standard [3]. Our primary focus is on modulation recognition techniques which is explained in detail in section 2.
1.1 Problem Considered for Study
In simple words, given a few samples of a received signal, the problem is to recognize the modulation scheme that is present using least apriori information. Once the modulation technique has been identified, demodulate the transmitted bits if required. The parameters of interest will be the speed of recognition (time taken by the algorithm to recognize the modulation scheme) and length of the signal needed to optimally recognize the modulation scheme.
For example, the received signal in AWGN channel is given as
Where is transmitted signal, is AWGN noise introduced in channel and is received signal.
For ASK, the transmitted signal is defined as
for bit 1
for bit 0
Wheredenotes amplitude, is carrier frequency and , is bit duration
A FSK transmitted signal is given by
for bit 1
for bit 0
Where and are two carrier frequencies
A PSK signal is defined as
for bit 1
for bit 0
The general form of a QPSK signal is
Without prior information of the incoming signal, identification of modulation scheme is a difficult problem. Thus, preprocessing of the signal is done to extract parameters of the signal such as SNR, symbol rate, bit rate. After extracting signal parameters, a suitable modulation recognition algorithm is applied to identify the modulation technique used in the signal. For example, assume that we know that the received signal contains one of the two schemes BPSK or QPSK in the presence of noise. The problem is to say which is present.
The paper is organized as follows.
Section 2 discusses some more applications of solutions for modulation recognition. Section 3 presents two methods for SNR estimation techniques and some of the very interesting work done in the area of modulation recognition. Conclusions are given in Section 4. Some references are given at the end.
2. Need for Modulation Recognition
There are so many communication signals all available with different types of modulation and different frequencies. It is necessary to identify and closely monitor these signals for few applications. Surveillance and military operations require classification of modulation whether analog or digital modulation technique because it is necessary to know the type of incoming signal. The recognizers help to distinguish the signal in presence of AWGN along with presence of other signals. The process of recognition is the first step in demodulating the original signal. [4]
Without any knowledge about incoming signal and its parameters, blind identification of modulation technique is difficult. Proper demodulation of the incoming signal takes place with the help of automatic modulation recognition without knowing the modulation scheme beforehand [5].
Furthermore, blind recognition of the modulation format of the received signal is an important problem in SDR [5]. SDR is widely used in military and commercial communication applications. Meanwhile, the requirement of identifying the modulation format of the incoming signal is necessary for designing intelligent wireless systems that can be used in different applications [6]. Software defined radio finds applications in following areas:
Military services: With the use of SDR technology, the military can reduce radio development costs by providing a common platform to which their wide variety of wireless systems could be integrated. The other advantage of SDR being used in military forces is to tap the information of the enemies i.e. the soldier gets to know the whereabouts of the enemies and thus prevent any operations carried out by enemies. SDR offers more flexibility to military in having joint operations with other nations.
Cellular telecommunication sector: As of now, every cell phone transmits and receives information through a speciï¬c standard. If we have to change the service provider then we need to change the cell phone also. SDR reprograms the cell phones to operate on different radio standards by simply uploading different software running on a common hardware platform. It can switch from one radio standard to another radio standard. For ex- the software can be upgraded easily for switch over from 2G to 3G or 4G, CDMA to GSM.
Space communication: Software defined radio can save cost of missions while constructing multiple mode, multiple band radio systems that can be upgraded using software. A common hardware is used for performing multiple functions through up gradation of software thus reducing the number of radios needed for a mission. If requirements change prior to the launch of an orbit, the software can be upgraded faster and easier than hardware. If an orbit fails, the software can be reconfigured. Thus, with minimal use of hardware, the total weight and cost of space hardware is reduced [7].
Public safety: During emergency situation such as earthquake, tsunami, terrorist attack, public safety officials from different agencies (in some cases, different counties and States) should able to communicate with each other effectively. If they are not able to share the information soon, then millions of lives will be lost. Public safety officials such as police officers, firefighters and doctors cannot always depend on wireless radio communication as their radios are often incompatible. SDR thus helps public safety officials to exchange information among themselves and thus millions of lives can be saved without wasting time [8].
3. Literature Review
In this section, first two methods to estimate SNR are reviewed and a few methods for modulation classification schemes are also studied after SNR estimation techniques. In most wireless communication systems, signals are usually corrupted by noise. SNR defined as the ratio of signal power to noise power, is commonly used as an essential metric in determining related system parameters, such as BER and SER. Moreover, various algorithms and system components require knowledge of the SNR for optimal performance. Thus, SNR is an important measure of channel quality in many modern wireless communication systems. [9] proposes an SNR estimation algorithm for digital modulated signals in AWGN without any prior information about the signal parameters such as carrier frequency, bit rate and modulation scheme of the signals. First, the autocorrelation matrix is obtained by calculating the autocorrelation of the received signal. Eigenvalues are calculated according to eigenvalue decomposition of the above matrix. Then the signal subspace dimension is calculated from minimum description length (MDL) criteria and then finally SNR is estimated. Simulation results are carried out for MPSK signals (M=2, 4, 8), MFSK signals (M=2, 4) and MQAM signals (M=16, 64, 128, 256). The results show that the corresponding STD is below 0.55 when the true SNR varies from -5dB to 25dB. [10] presents different approach to the problem of the estimation of signal bandwidth and SNR estimation in this paper, which is based on the corner of power distribution function. This method does not need to know the exact parameters such as the carrier frequency, symbol rate and the modulation pattern of the received signals and it makes no requirements on training sequence and the synchronization information. Simulation results shows that the algorithm has better estimation accuracy for wireless communication signals with various modulation schemes under the condition of AWGN, especially applicable to the cases with unknown modulation parameters and types. We now present essential features of a few interesting papers on modulation recognition. In [11], Blind Modulation Recognition (BMR) algorithm using Hilbert transform was used to identify modulation types. Computer simulation shows that performance of recognition algorithm increases as SNR increases. For low SNR, performance degrades. This technique is able to recognize only higher order QAM signals and fails to identify other modulation schemes. A modulation classification algorithm using wavelet Transform and histogram calculation [12] was used to identify QPSK and QAM with GMSK and M-ary FSK modulation type. The simulated results show that the correct modulation identification is possible to a lower bound of 5 dB and 12 dB for GMSK and QPSK respectively. When SNR is above 5 dB, the throughput of the recognition algorithm is more than 97.8%. The above technique identifies only constant envelope modulation schemes. It does not identify non-constant envelop modulation schemes. In [13], decision theoretic approach was proposed to identify digitally modulated schemes based on few key features and the threshold values determined for these key features. Simulations show that as SNR decreases, the percentage of correct classification with 100 % success rate also decreases. Instantaneous features such as amplitude, phase and frequency and stochastic features such as amplitude mean, amplitude mean-square, phase mean [14] were used to distinguish modulated signals for varying Signal to Noise Ratio (SNR). This method achieves high success of classification having low SNR for non-constant envelop signals whereas for constant envelope signals, there is high success of classification having high SNR. Three digital modulation classifiers for the application in blind modulation detection stage of adaptive OFDM modulation are presented and investigated in [15]. These digital modulation classifiers are namely Maximum Likelihood Modulation Classification (MLMC), higher order statistics using fourth order cumulants and higher order statistics using sixth order cumulants. From simulation results, it is seen that performance of 4th order cumulants shows confusion between 16 QAM and 64 QAM whereas there is a huge gap between 16 QAM and 64 QAM in performance of 6th order cumulants.
4. Conclusions
In this paper, we examined how software defined radios evolved and how the integration of software alongwith minimum hardware components made software defined radios cost effective. The SNR estimation plays an important role in many modern wireless communication systems. It can help us adopt adaptive demodulation algorithms to enhance the performance and provide the channel quality information required by the handoff, power control, and channel assignment algorithms. We studied few papers related to the methods of SNR estimation. We also studied various methods presented in literature survey to identify the modulation type of the signal for use in software radio systems. These methods required minimum prior knowledge of various signal parameters such as carrier frequency, symbol rate or class of modulation scheme etc. When modulation type is identified, an appropriate demodulator can demodulate the signal to recover the information that has been transmitted. Therefore, modulation recognition is an essential step to retrieve the exact transmitted signals.
