A New Normal Walk Model for Cellular Networks

Abstract

In most two-dimensional random walk models, the directions of a mobile station or subscribe moving out of a cell are assumed to be independent and identical distributed. The mobility patterns obtained from such a model may be less sufficient and less effective to measure the performance of cellular networks. This paper presents a new normal walk model to provide a more realistic and more versatile mobility patterns. This model mainly utilizes an equivalent drift angle with normal distribution to determine the next moving-out direction in a hexagonal cellular network. The experimental results demonstrate that if let the standard deviation of drift angles, σ, be 71°, a normal walk could almost behave as a random walk. With three different σ, 15°, 30°, and 60°, the results show that the errors between the analysis values and the simulation values are all within ±0.75%, even ±0.5% for most test cases. Our normal walk model can be effectively used for evaluating the performance of cellular networks, based on those more realistic and more objective mobility patterns.