In direct-sequence (DS) spread-spectrum communications, it is often necessary to obtain pseudo-noise (PN) code synchronization in the presence of both narrowband interference and frequency-selective multipath. However, conventional delay-lock code tracking loops are not optimized for such applications, nor are they well-suited to digital implementations. A code tracking algorithm based on the extended Kalman filter (EKF) is described that provides both code synchronization and joint estimates of interferer and channel parameters. It is first assumed that the interference can be modeled by an N-th order autoregressive (AR) process, and the multipath by a finite impulse response filter. A composite channel, equivalent to the convolution of the prewhitening filter and multipath coefficients is then constructed. The received waveform is shown to be a linear function of the composite channel parameters, that can therefore be directly estimated by an extended Kalman filter. The code delay /spl tau/ is viewed as a nonlinear parameter, that can likewise by estimated, after an appropriate linearization, using the EKF. The performance of the algorithm is first evaluated by computing the average bit-error rate (BER) of a digital RAKE receiver, that employs the joint delay, channel and interferer estimates derived from the EKF. In addition, a nonlinear analysis of the EKF is performed in which the code delay error and error covariance are modeled as a two-dimensional Markov process. The joint probability density function of these parameters, and nonlinear delay error variance, are obtained via numerical integration of a two-dimensional Chapman-Kolmogorov equation.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">></ETX>