On the Effect of a Bias Error in the Adaptive LMS Algorithm.
Abstract
A basic assumption in the derivation of the adaptive beamformer LMS algorithm is that the input sample vector is of zero mean. In the actual hardware implementation of this algorithm, bias errors can be encountered due to system electronics. For example, if the beamformer is operated for a considerable length of time, then a small bias in a nominally even a/d quantizer can seemingly cause erroneous algorithm convergence with respect to the expected zero mean system input. This memorandum considers some of the elemental aspects of such biasing errors.
Document Details
- Document Type
- Technical Report
- Publication Date
- Apr 07, 1970
- Accession Number
- ADA060236
Entities
People
- Norman L. Owsley
Organizations
- Navy Underwater Sound Laboratory