We consider a the problem of coherently combining signals having several sinusoidal components as measured via multiple sensors, in which each sensor has its own transfer function. It is desirable to combine the sensor outputs to improve the signal-to-noise ratio, but simply co-adding the signals could result in signal cancellation, due to phase changes in the sensors. We propose here a combining architecture which blindly adjusts the phases of the signals to maximize signal output. Different combining filters are considered: an allpass filter, and FIR filters designed according to a maximum SNR and minimum mean-squared error constraint. The allpass filters are trained both via steepest ascent and simplex optimization. The allpass combining filters provide excellent SNR improvement, while preserving all the frequency components.
[1]
Robert H. Halstead,et al.
Matrix Computations
,
2011,
Encyclopedia of Parallel Computing.
[2]
William H. Press,et al.
Numerical recipes in C
,
2002
.
[3]
IEEE Signal Processing and Signal Processing Education Workshop, SP/SPE 2015, Salt Lake City, UT, USA, August 9-12, 2015
,
2015,
SP/SPE.
[4]
Robert Bregovic,et al.
Multirate Systems and Filter Banks
,
2002
.
[5]
Stephen P. Boyd,et al.
Convex Optimization
,
2004,
Algorithms and Theory of Computation Handbook.
[6]
T. Moon,et al.
Mathematical Methods and Algorithms for Signal Processing
,
1999
.