Abstract

This paper proposes a high-resolution adaptive range-dimensional beamforming approach to improve the performance of the range estimation of active sonars.In our approach,three types of weight matrices are proposed.

The first approach utilizes a diagonal matrix,where the first element on the main diagonal is the residual sum of squares,while all the rest is set to zero.The second approach uses an identity matrix.In the third approach,we use a diagonal matrix,where the main diagonal is filled with the residual sum of squares used in the first one.

Furthermore,we also propose a high-order linear prediction model with different references of sub-bands by using the three weight matrices to construct the output of the power spectrum.One advantage of our proposed approach avoids the low-order linear prediction used by the range-dimensional beamforming method.

We compare our approach with states-of-theart range-dimensional beamforming methods.The experimental results show that our proposed approach achieves a higher range resolution.