Informatics Publishing Limited

Shripad P. Mohani ¹, Prashant B. Patel ², Mukul S. Sutaone ¹

Affiliations
1 Electronics and Telecommunication Engineering Department of College of Engineering, Pune, IN
2 Rolta India Ltd., IN

Abstract

Orthogonal Frequency Division Multiplexing (OFDM) is a multicarrier system supporting high data rate transmission and generally requires no equalization at the receiver, making it simple and efficient. One of the challenging issues for OFDM system is its high Peak-to-Average Power Ratio (PAPR). The high PAPR in OFDM leads to the use of high power amplifiers (HPAs) of wide dynamic range, leading to high cost and low efficiency of these amplifiers. There are several techniques proposed in the literature to tackle this problem of high PAPR in OFDM. One of the commonly used techniques is Partial-Transmit-Sequence (PTS). Artificial Bee Colony (ABC) algorithm used in basic PTS has also been suggested in the literature. In this paper, a new modified ABC-PTS algorithm using Companding and Reed-Solomon coding is proposed, which offers more reduction in PAPR compared to ABC-PTS algorithm. The increased Bit Error Rate (BER) caused due to companding is reduced by using Reed-Solomon code.

Keywords

ABC-PTS, Companding, OFDM, PAPR, Reed-Solomon Coding.

Full Text

     

Shripad P. Mohani ¹, Shipra Sarkar ¹, Mukul S. Sutaone ¹

Affiliations
1 Electronics and Telecommunication Department, College of Engineering, Pune-411005, Maharashtra, IN

Abstract

In tone reservation (TR) based OFDM systems, the peak to average power ratio (PAPR) reduction performance mainly depends on the selection of the peak reduction tone (PRT) set and the optimal target clipping level. Determining the optimal PRT set requires an exhaustive search of all combinations of possible PRT sets and this search is impractical for the number of tones used in practical systems. There are other selection methods too but with high computational complexity.
In this paper, an efficient scheme based on genetic algorithm (GA), is proposed for searching a nearly optimal extended PRT set. TR-based clipping is simple and attractive for practical implementation but determination of the optimal target clipping level is difficult. To overcome this problem, an adaptive clipping control algorithm is used. Simulation results show that the proposed algorithm efficiently obtains a nearly optimal extended PRT set and significant PAPR reduction.

Keywords

Genetic Algorithm, OFDM, PAPR, Tone Reservation.

Full Text

     

Madan B. Mali ¹, Mukul S. Sutaone ²

Affiliations
1 Department of Electronics and Telecomm, Sinhgad College of Engineering, Pune, IN
2 Government College of Engineering, Pune, IN

Abstract

Due to advantages of vector quantization (VQ) over scalar quantization, VQ became most favorable scheme for systems where extreme fast decoders are needed. An algorithm using constrained codebook, based on full search VQ and Hierarchical VQ is proposed here. It performs appropriate cluster search depending on the two inequalities defined as positive polarity and negative polarity. This reduces the searching time of the codebook by (N-(NCc×(NC-1))) codevectors. Where N, NC and NCc are total number of reference code vectors, number of clusters and number of code vectors in a cluster respectively. For developed algorithm a processor architecture supporting its computational requirements is also proposed. Processor makes use of pipelining, parallelism, and data control path. The processor performs fast nearest matching codevector search compared to FSVQ by (NCc+NC/N) amount and compared to tree search by (NCc+(NC/log2N)). It also outperforms hierarchical VQ with k-dimensional input and n-bit Nh codevectors codebook as it requires k×Nh search complexity against the complexity of (1×(NCc+NC)) proposed algorithm. The dedicated optimized chip has been designed and laid out using CMOS technology.

Keywords

Full Search VQ, Hierarchical VQ, Tree Search VQ, Vector Quantization.

Full Text

     

Prashant P. Bartakke ¹, Sameer D. Mali ¹, Mukul S. Sutaone ¹

Affiliations
1 Department of Electronics and Telecomm, College of Engineering, Pune, IN

Abstract

Texture, being surface property of every object, plays important role in human visual system for object identification and recognition. Texture classification and segmentation are the important operations towards recognition. Simultaneous Autoregressive (SAR) models had been successfully used in texture classification and segmentation but it has difficulty in selecting the appropriate neighborhood and window size used to estimate the model parameters. The Rotation-Invariant (RI) and Multi-resolution (MR) SAR models were found to produce acceptable classification and segmentation results. The suitability of these two models is verified in the work with comprehensive and variety of texture banks namely stochastic, periodic and mixed. The size of textures used to train the classifier also affects the classification accuracy to great extent. The comparison between Euclidean and Mahalanobis classifier is also provided in the paper. MR-RISAR with two or more resolution levels and model order of two or more gives acceptable classification results. The MR-SAR model parameters are used here to segment a multi-textured image. Preprocessing and feature weighting improves the segmentation quality except at the texture boundaries. The % ERROR parameter, defined as ratio of number of miss-labeled pixels to total number of pixels, is used to quantify the segmentation quality.

Keywords

Multi-Resolution, Rotation-Invariant, Simultaneous Auto-Regressive Models, Stochastic and Structural Textures.

Full Text

     

Shrinivas P. Mahajan ¹, Jyotsana Sahu ², Mukul S. Sutaone ³, V. K. Kokate ²

Affiliations
1 Department of Electronics and Telecomm, College of Engineering, Pune, IN
2 College of Engineering, Pune, IN
3 Electronics and Telecommunication Department, IN

Abstract

Audio classification has found widespread use in many emerging applications. It involves extraction of vital temporal, spectral and statistical features, and using these in creating an efficient classifier. Most of the audio classification work has been done on binary class classification. In our work we suggest best suited features for classification of different audio classes. Here, we present an algorithm for audio classification that is capable of segmenting and classifying an audio stream into speech male, speech female, music, noise and silence. The speech clips are further segment into voiced and unvoiced frames. A number of timbre features have been discussed, which distinguish the different audio formats. For pre classification, Probability Density Function (PDF), which is a threshold-based method, is performed over each audio clip. For further classification, K-Nearest Neighbor (KNN) and Support Vector Machine (SVM) Classifiers are proposed. Experiments have been performed to determine the best features of each binary class. Utilization of these features in multiclass classification yielded accuracy 96.34% in audio discrimination.

Keywords

Audio Feature Extraction, Bayesian Classification, K-Nearest Neighbor, Support Vector Machine.

Full Text

     

Shrinivas P. Mahajan ¹, Abhijit A. Pawar ², Mukul S. Sutaone ³, V. K. Kokate ²

Affiliations
1 Department of Electronics and Telecom, College of Engineering, Pune, IN
2 College of Engineering, Pune, IN
3 Electronics and Telecom Department, College of Engineering, Pune, IN

Abstract

Automatically extracting music information is gaining importance as a way to structure and organize the increasingly large number of music files digitally available and has become an important part of multimedia research. This research becomes more interesting as well as challenging when the music analysis is done according to the instruments used in the making of music. One such very popular percussion instrument called Tabla widely used in accompanying Indian Classical music recitals is analyzed based on the collection of digital recordings. The database consists of popular taals commonly used in Indian classical music. Bols are the basic notes of Tabla. A taal is a predefined sequence of Tabla bols. Different such bol arrangements give rise to various taals. Based on the speed of repetition of bols, the taals are broadly classified into low (Vilambit), medium (Madhya) and fast (Drut) tempo. Thus a tempo represents the rhythmic information of a taal.
In this paper, an automatic system for identifying and transcribing Tabla bols of different tempos is explored. The transcription process is based on three main steps: firstly tempo of the audio clip is identified using autocorrelation technique. Secondly, the recorded clip is segmented where each segment represents a bol. In the third step, Mel-Frequency Cepstral Coefficients (MFCC) features are extracted from the separated bols to form templates for pattern classification. Two pattern classification techniques namely Dynamic Time Warping (DTW) and Vector Quantization (VQ) are analyzed and compared for evaluating the performance of bol identification. Overall bol identification accuracy of the system for tempo independent classification is 96.18%.

Keywords

Autocorrelation, Bol Segmentation, Dynamic Time Warping, Feature Extraction, Onset Detection, Vector Quantization.

Full Text