Robust channel estimation using the hybrid particle filter

Mobile wireless channels are subject to multipath with potentially large Doppler spreads. Previous joint channel and delay estimation algorithms, based for example on the extended Kalman filter (EKF), are subject to divergence. We present a hybrid particle filter (HPF), also called the particle extended Kalman filter (PEKF) for delay/channel estimation in direct-sequence CDMA applications. The PEKF uses a sampling technique to update the nonlinear delay state variables, combined with Kalman filters to track the linear channel coefficients. Initial simulation results indicate the ability of the PEKF to track delays and channels with large Doppler spreads on the order of 1/spl times/10/sup -3//T, where 1/T is the symbol rate.