Neural network handoff in shadow-Rayleigh fading

A handoff algorithm in a cellular system under shadow-Rayleigh fading propagation is presented. The algorithm applies the timing advance for mobile positioning. The inputs of the proposed radial-basis function network consist of the mobile direction obtained by the MUSIC method on an array antenna, the received signal strengths (RSSs) of the mobile from the serving cell and from nearby cells and traffic intensities. Consequently, we obtain a significantly improved performance in comparison to the hysteresis rule. The evaluations are done in measurements of (1) handoff rate, blocking rate, dropping rate versus the traffic intensities and mean arrival times and (2) the difference between the mobile signal strength and the required minimum strength value before the system drop the calls versus the number of handoffs.