The investigation of new adaptive filtering algorithms for telecommunications echo cancellation implemented in TMS32010 fixed point assembly code

Current state-of-the art echo cancellers for telecommunications typically employ the least mean square (LMS) adaptive algorithm developed in the late 1950's and early 1960's. The canceller is usually implemented as a nonrecursive, or finite impulse response (FIR), filter structure. More recently, interest has grown in recursive, or infinite impulse response (IIR), adaptive filter structures for the echo cancellation application. Floating-point computer simulations have shown that adaptive IIR filters require less computation to match certain transfer relations than do adaptive FIR filters. However, it is a practical question as to how well these algorithms perform when implemented in a fixed-point environment, which is typically the case in practical situations. This paper examines the performance of several IIR adaptive filtering algorithms as implemented on a Texas Instruments TMS32010 digital signal processor for echo cancellation in telecommunication environments.