This preview shows pages 1–3. Sign up to view the full content.
This preview has intentionally blurred sections. Sign up to view the full version.View Full Document
Unformatted text preview: Leaky LMS Algorithm Leaky LMS Algorithm Â¡ Convergence of tap-weight error modes dependent on conditioning of input : Ï‡ (R). Â¡ For ill-conditioned inputs tap-weight error modes are undriven and undamped. Â¡ Stability and convergence time issues of concern for ill- conditioned inputs. Leaky LMS Algorithm Leaky LMS Algorithm Â¡ Cost function: Â¡ Tap-weight update: Â¡ Tap-weight converges in mean to a biased solution: Â¡ Tap-weight error modes stabilized if : Sign Sign---- based Algorithms based Algorithms Â¡ Sign-error algorithm: Â¡ Sign-data algorithm: Â¡ Sign-sign algorithm: Sign Sign-based Algorithms based Algorithms Â¡ Incorporate 2-bit quantization of estimation error e[n] and/or tap-input u[n]. Â¡ Tap-weight update not guided by gradient. Â¡ Tap-weights do not converge to optimal Wiener solution. Â¡ May need regularization to stabilize the tap-weight update. Affine Affine Projection Algorithm Projection Algorithm Â¡ Tap-weight adjustment of NLMS: w[n+1] â€“ w[n] in the direction of u[n]. the direction of u[n]....
View Full Document
This note was uploaded on 12/02/2011 for the course AR 107 taught by Professor Gracegraham during the Fall '11 term at Montgomery College.
- Fall '11