Yadollah Salahshuori - Dept. of Electrical Engineering, Razi University, Kermanshah, Iran
Ghasem Azemi

Traditional handoff algorithms generally cannot keep both the average number of unnecessary handoffs and handoff decision delays low. In order to achieve this goal, this paper presents a novel pattern recognition based handoff algorithm which uses a neural network as a classifier. It is based on the fact that in micro-cellular environments in cities, users often move on predetermined paths and as the built environment is not changed often, this regularity can be exploited in a pattern recognition based handoff method. Simulation results show that unlike traditional handoff algorithms, the proposed algorithm has the ability to keep the average number of handoffs and average number of call drops low for a negligible handoff decision delay. The proposed handoff algorithm is simpler than the existing pattern recognition based ones and only uses one training vector for each path in a given environment

Handover, handoff, pattern recognition, neural networks