Dr V Umapathi Reddy's Profile

Wireless Communications:
Recent work was on low-complexity near ML decoding algorithms for MIMO architectures. Most recent work was focused on multicarrier code division multiple access (MC-CDMA) based physical layer. OFDM (orthogonal frequency division multiplexing) modulation format is very popular and many of the current wireless communication standards adopted this format. In OFDM based system, uniform loading of the carriers is normally resorted to. If some of the carriers are in deep fade, the symbols loaded on these carriers will experience poor received SNR thereby degrading the overall performance. One way to mitigate this degradation is to spread the symbols across the carriers. In order to avoid loss in data rate, we need to spread the symbols using orthogonal codes. This method has been analyzed and shown that MC-CDMA yields much improved performance compared to the conventional OFDM.

Blind Modulation Classification:
Blind modulation classification is the intermediate operation between signal detection and demodulation. It is important in both commercial and defense applications. In the presence of various practical conditions such as carrier frequency and timing offsets, and fading multipath environment, blind identification of the transmitted modulation format is a challenging task. The problem becomes more complex if the class size, i.e., number of modulations which the transmitted signal can take, is large. Also, if the modulations are a combination of those with memory and without memory, the problem becomes much more complex.

Recent work was focused on the 9-class problem, consisting of digital modulations without memory, in the presence of timing offset and Rayleigh fading multipath propagation. The method developed for this problem is a two-step approach. In the first step, the multipath channel is estimated using a subspace based approach which exploits the property that when the input to the channel is a second-order cyclo-stationary process, the oversampled output is a second-order vector stationary process. In the second step, the channel estimate is used to compute the required features for classification. A hierarchical algorithm is then proposed which performs very well for moderate values of SNR.

MIMO Radar:
MIMO radar is a special case of multi-static radar system which employs multiple transmit waveforms, jointly designed, and has the ability to jointly process the signals received at multiple receivers. The ambiguity function determined by the transmitted waveforms specifies resolution in range, Doppler and angle, and hence, design of desired waveforms is an important problem. Frequency-hopping codes have been used in pulse compression radars because of their highly desirable ambiguity properties. Recent work was focused on low-complexity design of orthogonal frequency-hopping codes for arbitrary Doppler case. The method developed is based on the hit-matrix concept which extends the hit-array formalism extensively used in the context of frequency-hopping waveforms for phased-array radar.

Current work is focused on synthesis of transmit waveforms from the zero-lag correlation matrix of the waveforms (obtained from other considerations such as desired spatial beamforming pattern) subject to certain practical constraints. Also, optimal detector, for a specified MIMO configuration in terms of the corresponding channel covariance matrix and waveform correlation matrix, is under study.