Kalman filtering has been proposed in the literature for wireless channel estimation, however, it is not sufficiently robust to uncertainties in the channel auto-correlation model as well as to multiple access interference (MAI). This paper presents a receiver structure for direct-sequence code-division multiple-access (DS-CDMA) systems by using robust Kalman estimation and post-correlation (i.e., symbol rate) processing for channel estimation. The proposed structure is also generalized to incorporate multiple-antenna combining and interference cancellation techniques. The resulting receiver outperforms earlier structures in the presence of channel modeling uncertainties, MAI, and low-received signal-to-noise ratio. The enhancement in performance is achieved at the same order of complexity as a standard Kalman-based receiver.