Blind channel estimation for long code multiuser CDMA systems

In CDMA systems with long codes, the users' signatures change in each bit period, impeding the estimation of the time-invariant multipath parameters. In this paper, correlation-matching methods are employed to blindly estimate those multipath parameters. For given code sequences, the output correlation matrix (parameterized by the unknown channel coefficients) is compared with its instantaneous approximation. By minimizing the Frobenius norm of the resulting error matrix, the channel parameters can be estimated up to a complex scalar ambiguity. Both batch and adaptive algorithms are derived. Under the assumption of i.i.d. code sequences, identifiability up to a complex scalar ambiguity for each channel is guaranteed, and the asymptotic convergence of the proposed algorithm is established. Furthermore, step-size selection for the adaptive version is investigated. When only the code sequence of the user of interest is available, a single user receiver is also derived. Simulation results verify those claims and provide comparisons with other methods.