A novel blind multiuser detection based on adaptive quasi-Newton algorithm for eigenspace estimation

A new adaptive quasi-Newton eigen-estimation method is applied to blind multiuser detection. Computer simulation results are provided to compare its performance with the PASTd algorithm. The results show that the proposed approach performs identically to the PASTd algorithm and is significantly superior to it. Especially, the proposed approach has a relatively fast convergence rate, which may pose a problem for a time varying system, so it is a good adaptive multiuser detection algorithm.