The image classification is a classical problem of image processing, computer vision and machine learning fields. In this paper we study the image classification using deep learning. We use AlexNet architecture with convolutional neural networks for this purpose. Four test images are selected from the ImageNet database for the classification purpose. We cropped the images for various portion areas and conducted experiments. The results show the effectiveness of deep learning based image classification using AlexNet.

Ionospheric time delay variations are highly variable and random in the low latitude Indian region. It is a key responsible parameter for the range error in Global Positioning System (GPS) ranging signals which is proportional to the total electron content (TEC). Dual frequency GPS receiver at KL University, Vaddeswaram (16.31° N, 80.37° E), India (falls under the transition zone of the Equatorial Ionization Anomaly in low latitude) is considered in the analysis. An Auto Regressive Moving Average (ARMA) model is implemented for short term forecasting of the VTEC (Vertical TEC) values. Three Geomagnetic storms occurred in the current solar maximum period (2013-2014) are considered to test the performance of ARMA model. The forecasted VTEC values are compared with original VTEC the values and IRI models. The forecasted results indicate that ARMA model would be useful to set up an early warning system of ionospheric disturbances.

This paper focuses on reducing the processing time of software GPS receiver using sub-sampling techniques. As the GPS signals are wide band signals, sampling frequency is very high. Sub-sampling enables to reduce time for processing. From the simulation results, it is observed that the sampling frequency can be reduced up to 2.5 MHz without loss of tracked signal. The processing time is reduced for software GPS receiver after sub-sampling. The overall reduction in processing time is from 3.456647 sec to 2.15946 sec respectively for sampling frequency 5 MHz , 2.5 MHz. Thus time saved is 37% of the original one.

The modern signal and image processing deals with large data such as images and this data deals with complex statistics and high dimensionality. Sparsity is one powerful tool used signal and image processing applications. The mainly used applications are compression and denoising. A dictionary contains information of the signals in the form of coefficients. Recently dictionary learning has emerged for efficient representation of signals. In this paper we study the image compression using both analytical and learned dictionaries. The results show that the effectiveness of learned dictionaries in the application of image compression.

Software GPS receiver (SGR) are becoming prevalent because of their flexibility, re-configurability and computational efficiency. Acquisition, pull in frequency and tracking are the main building blocks of SGR design. In this paper, sub sampled version of signal has been processed through these three blocks which effectively reduce processing time. Sub-sampling mainly violets Nyquist criteria. Wide bandwidth of GPS signal set up the sampling frequency very high. The implementation of sub sampling technique followed by thresholding the tracked signal to 1  reduce the associated higher chipping rate and henceforth the processing time. The proposed technique decreases sampling frequency (F= 5 MHz) below center frequency (Fcar= 1.25 MHz) of GPS signal. Without loss of original information, signals are recovered. Experiments carried out estimate the computation from 4.2894 sec to 3.0853 sec, 2.178024 sec and 3.468482 sec respectively for sampling frequency F/2, F/4, F/6. This Simulation and real data carried out ensures the fast acquisition and tracking and power consumption.