Indian Institute of Science

Srinidhi Nagaraja

Balaji Sundar Rajan

Ananthanarayanan Chockalingam

Saif Khan Mohammed

Tanumay Datta

Techniques for enabling an estimate of a transmittal signal vector, given a received signal vector and a channel matrix to be computed, are provided. An initial solution signal vector for the estimate is calculated, and based on the initial estimate, a pool of possible solutions is generated. Methods described herein may be iterative in nature, and may cycle through possible estimates of the transmitted signal vector so as to continually improve the estimates and the pool of possible estimates. The methods may terminate once stopping criteria are reached. In some example methods, solutions may be marked at taboo and a taboo period can be established that indicates a number of subsequent iterations for which the current solution signal vector cannot be considered.

Techniques for decoding transmitted signals using reactive taboo searches (RTS)

A receiver of a multiple-input multiple-output (MIMO) system performs QR decomposition of the channel matrix to enable detection of a transmitted vector in a layered manner. In each layer, a sub-vector of the transmitted vector is estimated. A reactive tabu search is performed if an estimated symbol differs from a nearest symbol in the alphabet by a predetermined value. The receiver may order the entries of the channel matrix prior to QR decomposition to enable estimation in an optimum order. In another embodiment, a receiver performs multiple reactive tabu searches to estimate a transmitted vector. The receiver employs a fixed threshold or a variable threshold for a cost function used in the multiple reactive tabu searches depending on whether the MIMO system is under-determined or not. The techniques enable low bit-error rate (BER) performance in MIMO systems with large number of antennas and when higher-order modulation techniques are used.

Techniques for detection of signals in multiple-input multiple-output communication systems

Techniques for enabling an estimate of a transmitted signal vector, given a received signal vector and a channel matrix to be computed, are provided. An initial solution signal vector for the estimate is calculated, and based on the initial estimate, a pool of possible solutions is generated. Methods described herein may be iterative in nature, and may cycle through possible estimates of the transmitted signal vector so as to continually improve the estimates and the pool of possible estimates. The methods may terminate once stopping criteria are reached. In some example methods, solutions may be marked at taboo and a taboo period can be established that indicates a number of subsequent iterations for which the current solution signal vector cannot be considered.

Growing community of inventors

Bengaluru, India

Srinidhi Nagaraja