Implementation of transmit antenna weight adaptation through stochastic gradient feedback

Feedback assisted transmission antenna weight adaptation has been proposed for Third Generation cellular systems to provide increased capacity on the forward link. This approach provides enhancement concurrently from fading mitigation and beam steering. One such algorithm maximizes the usable power to the mobile by perturbation extracted gradient adaptation of the transmit weight vector. With perfect channel estimation and other idealized assumptions this method has been shown to be capable of providing weights approaching that of the optimal multi-element transmitter. In this paper a method for common perturbation transmission is introduced and compared to dedicated perturbation transmission in a CDMA system. In addition, a subspace weighting approach is introduced to enhance the algorithm performance in correlated fading. Simulations of the algorithm operating in a cdma2000 system incorporate feedback bit errors, latency and channel estimation errors with a dedicated pilot.