A neural network-based mobile positioning with hierarchical structure

A hierarchical structure for mobile positioning in cellular communication systems is proposed in this paper based on multi-layer perceptron neural networks. The location of the mobile phone is estimated based on combining angle of arrival (AOA) and time of arrival (TOA) location determination methods implemented by a neural network. The coverage area of each base station is divided to different cells at each level of hierarchical structure. The estimation of mobile location is improved at each level by using a new neural network that has been trained for a smaller cell. Simulation results for an urban area show that the proposed method achieves a good performance by increasing the number of levels in the hierarchical neural networks.