Compensating for inter-symbol interference (ISI)
Zero-Forcing (ZF): Completely inverts channel: W(f) = 1/H(f). Eliminates ISI but amplifies noise at frequency nulls.
MMSE: Minimizes mean squared error: W(f) = H*(f)/(|H(f)|² + N₀). Balances ISI removal with noise.
Decision Feedback (DFE): Uses past decisions to subtract post-cursor ISI. Better for severe channels.
Adaptive: LMS/RLS algorithms adjust taps in real-time using training sequence or decisions.