The tone capture properties of CMA-based interference suppressors

• Published in: IEEE transactions on acoustics, speech, and signal processing Vol. 33; no. 4; pp. 946 - 958
• Main Authors: Treichler, J., Larimore, M.
• Format: Journal Article
• Language: English
• Published: New York, NY IEEE 01-08-1985; Institute of electrical and electronics engineers
• Subjects: Adaptive algorithm; Adaptive filters; Additives; Applied sciences; Degradation; Digital filters; Exact sciences and technology; Filtering algorithms; Finite impulse response filter; Information, signal and communications theory; Interference suppression; Quadrature phase shift keying; Signal processing; Telecommunications and information theory; Filtering; Electromagnetic interference; Algorithm; Narrow band; Suppression; Capture
• ISSN: 0096-3518
• DOI: 10.1109/TASSP.1985.1164640

An earlier paper introduced the constant modulus algorithm (CMA), an adaptive filtering technique for correcting multipath and interference-induced degradations in constant envelope waveforms such as FM and QPSK signals. This algorithm exploits the fact that both multipath propagation and additive interference disrupt the constant envelope property of the received signal. By sensing the received envelope variations, the adaptive algorithm can reset the coefficients of an FIR digital filter so as to remove the variations and, in the process, suppress the various interference components from the desired signal. This paper examines a problem that arises when using CMA to suppress narrow-band interference. If both the interferer and the signal have constant envelope and are spectrally nonoverlapping, then it is possible to find two different filter solutions, one which suppresses the interferer and another which "captures" the interferer and suppresses the desired signal. This paper examines how "capture, can occur and how it may be prevented. This problem is studied by characterizing the algorithm's behavior to an input consisting of only two sinusoids. Assuming slow adaptation, the N-dimensional adaptive weight recursion is shown to compress into a two-by-two recursion in the tone output amplitudes. This simplified recursion is then analyzed to determine what combinations of input amplitudes (signal-to-interference ratios) and filter initial conditions lead to "lock" and which lead to the capture of the interferer. The results are then broadened to include multiple input tones and signals with nonzero bandwidth.